id
stringlengths 9
16
| submitter
stringlengths 1
64
⌀ | authors
stringlengths 3
62.8k
| title
stringlengths 1
382
| comments
stringlengths 1
1.15k
⌀ | journal-ref
stringlengths 1
557
⌀ | doi
stringlengths 7
153
⌀ | report-no
stringlengths 1
509
⌀ | categories
stringlengths 5
125
| license
stringclasses 9
values | abstract
stringlengths 6
6.09k
| versions
listlengths 1
187
| update_date
timestamp[s]date 2007-05-23 00:00:00
2025-04-25 00:00:00
| authors_parsed
sequencelengths 1
2.83k
|
|---|---|---|---|---|---|---|---|---|---|---|---|---|---|
2503.08021 | Daowei Lu PhD | Daowei Lu, Dingguo Wang | Constructions of Rota-Baxter operators by L-R smash products | null | null | null | null | math.RA | http://arxiv.org/licenses/nonexclusive-distrib/1.0/ | Let $A$ and $H$ be two cocommutative Hopf algebra such that $A$ is an
$H$-bimodule Hopf algebra. Suppose that $R:A\rightarrow A$ is a linear map and
$B$ is a Rota-Baxter operator of $H$. In this paper we will characterize the
Rota-Baxter operators on the L-R smash product $A\natural H$ and give the
necessary and sufficient conditions to make $\overline{B}$ a Rota-Baxter
operator of $A\natural H$. Then we will consider the dual case, and construct a
Rota-Baxter co-operator on the L-R smash product $C\ltimes H$.
| [
{
"version": "v1",
"created": "Tue, 11 Mar 2025 04:02:10 GMT"
},
{
"version": "v2",
"created": "Mon, 21 Apr 2025 00:13:48 GMT"
}
] | 2025-04-22T00:00:00 | [
[
"Lu",
"Daowei",
""
],
[
"Wang",
"Dingguo",
""
]
] |
2503.08224 | Dongbin Zhang | Dongbin Zhang, Yunfei Liu, Lijian Lin, Ye Zhu, Kangjie Chen, Minghan
Qin, Yu Li and Haoqian Wang | HRAvatar: High-Quality and Relightable Gaussian Head Avatar | Accepted to CVPR 2025,Project page:
https://eastbeanzhang.github.io/HRAvatar | null | null | null | cs.CV | http://arxiv.org/licenses/nonexclusive-distrib/1.0/ | Reconstructing animatable and high-quality 3D head avatars from monocular
videos, especially with realistic relighting, is a valuable task. However, the
limited information from single-view input, combined with the complex head
poses and facial movements, makes this challenging. Previous methods achieve
real-time performance by combining 3D Gaussian Splatting with a parametric head
model, but the resulting head quality suffers from inaccurate face tracking and
limited expressiveness of the deformation model. These methods also fail to
produce realistic effects under novel lighting conditions. To address these
issues, we propose HRAvatar, a 3DGS-based method that reconstructs
high-fidelity, relightable 3D head avatars. HRAvatar reduces tracking errors
through end-to-end optimization and better captures individual facial
deformations using learnable blendshapes and learnable linear blend skinning.
Additionally, it decomposes head appearance into several physical properties
and incorporates physically-based shading to account for environmental
lighting. Extensive experiments demonstrate that HRAvatar not only reconstructs
superior-quality heads but also achieves realistic visual effects under varying
lighting conditions.
| [
{
"version": "v1",
"created": "Tue, 11 Mar 2025 09:42:40 GMT"
},
{
"version": "v2",
"created": "Mon, 21 Apr 2025 06:29:45 GMT"
}
] | 2025-04-22T00:00:00 | [
[
"Zhang",
"Dongbin",
""
],
[
"Liu",
"Yunfei",
""
],
[
"Lin",
"Lijian",
""
],
[
"Zhu",
"Ye",
""
],
[
"Chen",
"Kangjie",
""
],
[
"Qin",
"Minghan",
""
],
[
"Li",
"Yu",
""
],
[
"Wang",
"Haoqian",
""
]
] |
2503.08863 | Debajyoti Kar | Debajyoti Kar, Arindam Khan, Malin Rau | Improved Approximation Algorithms for Three-Dimensional Bin Packing | null | null | null | null | cs.CG cs.DS | http://creativecommons.org/licenses/by/4.0/ | We study three fundamental three-dimensional (3D) geometric packing problems:
3D (Geometric) Bin Packing (3D-BP), 3D Strip Packing (3D-SP), and Minimum
Volume Bounding Box (3D-MVBB), where given a set of 3D (rectangular) cuboids,
the goal is to find an axis-aligned nonoverlapping packing of all cuboids. In
3D-BP, we need to pack the given cuboids into the minimum number of unit cube
bins. In 3D-SP, we need to pack them into a 3D cuboid with a unit square base
and minimum height. Finally, in 3D-MVBB, the goal is to pack into a cuboid box
of minimum volume.
It is NP-hard to even decide whether a set of rectangles can be packed into a
unit square bin -- giving an (absolute) approximation hardness of 2 for 3D-BP
and 3D-SP. The previous best (absolute) approximation for all three problems is
by Li and Cheng (SICOMP, 1990), who gave algorithms with approximation ratios
of 13, $46/7$, and $46/7+\varepsilon$, respectively, for 3D-BP, 3D-SP, and
3D-MVBB. We provide improved approximation ratios of 6, 6, and $3+\varepsilon$,
respectively, for the three problems, for any constant $\varepsilon > 0$.
For 3D-BP, in the asymptotic regime, Bansal, Correa, Kenyon, and Sviridenko
(Math.~Oper.~Res., 2006) showed that there is no asymptotic polynomial-time
approximation scheme (APTAS) even when all items have the same height. Caprara
(Math.~Oper.~Res., 2008) gave an asymptotic approximation ratio of
$T_{\infty}^2 + \varepsilon\approx 2.86$, where $T_{\infty}$ is the well-known
Harmonic constant in Bin Packing. We provide an algorithm with an improved
asymptotic approximation ratio of $3 T_{\infty}/2 +\varepsilon \approx 2.54$.
Further, we show that unlike 3D-BP (and 3D-SP), 3D-MVBB admits an APTAS.
| [
{
"version": "v1",
"created": "Tue, 11 Mar 2025 20:07:42 GMT"
},
{
"version": "v2",
"created": "Sat, 19 Apr 2025 09:53:48 GMT"
}
] | 2025-04-22T00:00:00 | [
[
"Kar",
"Debajyoti",
""
],
[
"Khan",
"Arindam",
""
],
[
"Rau",
"Malin",
""
]
] |
2503.08873 | \v{Z}an Grad | \v{Z}an Grad | Covariant derivatives in the representation-valued Bott-Shulman-Stasheff
and Weil complex | 65 pages | null | null | null | math.DG | http://creativecommons.org/licenses/by/4.0/ | For a Lie groupoid $G$, the differential forms on its nerve comprise a double
complex. A natural question is if this statement extends to forms with values
in a representation $V$ of $G$. In this paper, we research two types of
covariant derivatives which commute with the simplicial differential, yielding
two types of "curved" double complexes of forms with coefficients in $V$. The
naive approach is to consider a linear connection $\nabla$ on $V$, in which
case $d^\nabla$ commutes with the simplicial differential if and only if
$\nabla$ satisfies a certain (restrictive) invariance condition. The heart of
this paper focuses on another, more compelling approach: using a multiplicative
Ehresmann connection for a bundle of ideals. In this case, we obtain a
geometrically richer curved double complex, where the cochain map is given by
the horizontal exterior covariant derivative $D$, which generalizes the
well-known operator from the theory of principal bundles. Moreover, both
differential operators $d^\nabla$ and $D$ are researched in the infinitesimal
setting of Lie algebroids, as well as their relationship with the van Est map.
We conclude by using the operator $D$ to study the curvature of an
(infinitesimal) multiplicative Ehresmann connection.
| [
{
"version": "v1",
"created": "Tue, 11 Mar 2025 20:37:52 GMT"
},
{
"version": "v2",
"created": "Mon, 21 Apr 2025 17:28:41 GMT"
}
] | 2025-04-22T00:00:00 | [
[
"Grad",
"Žan",
""
]
] |
2503.09002 | Chenyuan Yang | Chenyuan Yang, Zijie Zhao, Zichen Xie, Haoyu Li, Lingming Zhang | KNighter: Transforming Static Analysis with LLM-Synthesized Checkers | null | null | null | null | cs.SE cs.AI cs.CR cs.OS | http://creativecommons.org/licenses/by/4.0/ | Static analysis is a powerful technique for bug detection in critical systems
like operating system kernels. However, designing and implementing static
analyzers is challenging, time-consuming, and typically limited to predefined
bug patterns. While large language models (LLMs) have shown promise for static
analysis, directly applying them to scan large systems remains impractical due
to computational constraints and contextual limitations.
We present KNighter, the first approach that unlocks scalable LLM-based
static analysis by automatically synthesizing static analyzers from historical
bug patterns. Rather than using LLMs to directly analyze massive systems, our
key insight is leveraging LLMs to generate specialized static analyzers guided
by historical patch knowledge. KNighter implements this vision through a
multi-stage synthesis pipeline that validates checker correctness against
original patches and employs an automated refinement process to iteratively
reduce false positives. Our evaluation on the Linux kernel demonstrates that
KNighter generates high-precision checkers capable of detecting diverse bug
patterns overlooked by existing human-written analyzers. To date,
KNighter-synthesized checkers have discovered 92 new, critical, long-latent
bugs (average 4.3 years) in the Linux kernel; 77 are confirmed, 57 fixed, and
16 have been assigned CVE numbers. This work establishes an entirely new
paradigm for scalable, reliable, and traceable LLM-based static analysis for
real-world systems via checker synthesis.
| [
{
"version": "v1",
"created": "Wed, 12 Mar 2025 02:30:19 GMT"
},
{
"version": "v2",
"created": "Fri, 18 Apr 2025 21:40:39 GMT"
}
] | 2025-04-22T00:00:00 | [
[
"Yang",
"Chenyuan",
""
],
[
"Zhao",
"Zijie",
""
],
[
"Xie",
"Zichen",
""
],
[
"Li",
"Haoyu",
""
],
[
"Zhang",
"Lingming",
""
]
] |
2503.10566 | Egor Zverev | Egor Zverev, Evgenii Kortukov, Alexander Panfilov, Alexandra Volkova,
Soroush Tabesh, Sebastian Lapuschkin, Wojciech Samek, Christoph H. Lampert | ASIDE: Architectural Separation of Instructions and Data in Language
Models | ICLR 2025 Workshop on Building Trust in Language Models and
Applications | null | null | null | cs.LG | http://creativecommons.org/licenses/by/4.0/ | Despite their remarkable performance, large language models lack elementary
safety features, and this makes them susceptible to numerous malicious attacks.
In particular, previous work has identified the absence of an intrinsic
separation between instructions and data as a root cause for the success of
prompt injection attacks. In this work, we propose a method, ASIDE, that allows
the model to clearly separate between instructions and data on the level of
embeddings. ASIDE applies a fixed orthogonal rotation to the embeddings of data
tokens, thus creating distinct representations of instructions and data tokens
without introducing any additional parameters. We demonstrate the effectiveness
of our method by instruct-tuning LLMs with ASIDE and showing (1) highly
increased instruction-data separation scores without a loss in model
capabilities and (2) competitive results on prompt injection benchmarks, even
without dedicated safety training. Additionally, we study the working mechanism
behind our method through an analysis of model representations.
| [
{
"version": "v1",
"created": "Thu, 13 Mar 2025 17:17:17 GMT"
},
{
"version": "v2",
"created": "Mon, 21 Apr 2025 17:45:08 GMT"
}
] | 2025-04-22T00:00:00 | [
[
"Zverev",
"Egor",
""
],
[
"Kortukov",
"Evgenii",
""
],
[
"Panfilov",
"Alexander",
""
],
[
"Volkova",
"Alexandra",
""
],
[
"Tabesh",
"Soroush",
""
],
[
"Lapuschkin",
"Sebastian",
""
],
[
"Samek",
"Wojciech",
""
],
[
"Lampert",
"Christoph H.",
""
]
] |
2503.10596 | Tianheng Cheng | Rui Hu, Lianghui Zhu, Yuxuan Zhang, Tianheng Cheng, Lei Liu, Heng Liu,
Longjin Ran, Xiaoxin Chen, Wenyu Liu, Xinggang Wang | GroundingSuite: Measuring Complex Multi-Granular Pixel Grounding | Work in progress. Code: https://github.com/hustvl/GroundingSuite.
Update: add more results & polish the report | null | null | null | cs.CV | http://creativecommons.org/licenses/by-nc-sa/4.0/ | Pixel grounding, encompassing tasks such as Referring Expression Segmentation
(RES), has garnered considerable attention due to its immense potential for
bridging the gap between vision and language modalities. However, advancements
in this domain are currently constrained by limitations inherent in existing
datasets, including limited object categories, insufficient textual diversity,
and a scarcity of high-quality annotations. To mitigate these limitations, we
introduce GroundingSuite, which comprises: (1) an automated data annotation
framework leveraging multiple Vision-Language Model (VLM) agents; (2) a
large-scale training dataset encompassing 9.56 million diverse referring
expressions and their corresponding segmentations; and (3) a meticulously
curated evaluation benchmark consisting of 3,800 images. The GroundingSuite
training dataset facilitates substantial performance improvements, enabling
models trained on it to achieve state-of-the-art results. Specifically, a cIoU
of 68.9 on gRefCOCO and a gIoU of 55.3 on RefCOCOm. Moreover, the
GroundingSuite annotation framework demonstrates superior efficiency compared
to the current leading data annotation method, i.e., $4.5 \times$ faster than
the GLaMM.
| [
{
"version": "v1",
"created": "Thu, 13 Mar 2025 17:43:10 GMT"
},
{
"version": "v2",
"created": "Mon, 21 Apr 2025 14:25:51 GMT"
}
] | 2025-04-22T00:00:00 | [
[
"Hu",
"Rui",
""
],
[
"Zhu",
"Lianghui",
""
],
[
"Zhang",
"Yuxuan",
""
],
[
"Cheng",
"Tianheng",
""
],
[
"Liu",
"Lei",
""
],
[
"Liu",
"Heng",
""
],
[
"Ran",
"Longjin",
""
],
[
"Chen",
"Xiaoxin",
""
],
[
"Liu",
"Wenyu",
""
],
[
"Wang",
"Xinggang",
""
]
] |
2503.10664 | Timo Laine | Timo Aukusti Laine | Semantic Wave Functions: Exploring Meaning in Large Language Models
Through Quantum Formalism | 29 pages, 4 figures. Some corrections added | OA J Applied Sci Technol, 3(1), 01-22 (2025) | 10.33140/OAJAST | null | cs.CL cs.LG quant-ph | http://creativecommons.org/licenses/by/4.0/ | Large Language Models (LLMs) encode semantic relationships in
high-dimensional vector embeddings. This paper explores the analogy between LLM
embedding spaces and quantum mechanics, positing that LLMs operate within a
quantized semantic space where words and phrases behave as quantum states. To
capture nuanced semantic interference effects, we extend the standard
real-valued embedding space to the complex domain, drawing parallels to the
double-slit experiment. We introduce a "semantic wave function" to formalize
this quantum-derived representation and utilize potential landscapes, such as
the double-well potential, to model semantic ambiguity. Furthermore, we propose
a complex-valued similarity measure that incorporates both magnitude and phase
information, enabling a more sensitive comparison of semantic representations.
We develop a path integral formalism, based on a nonlinear Schr\"odinger
equation with a gauge field and Mexican hat potential, to model the dynamic
evolution of LLM behavior. This interdisciplinary approach offers a new
theoretical framework for understanding and potentially manipulating LLMs, with
the goal of advancing both artificial and natural language understanding.
| [
{
"version": "v1",
"created": "Sun, 9 Mar 2025 08:23:31 GMT"
},
{
"version": "v2",
"created": "Mon, 21 Apr 2025 13:04:29 GMT"
}
] | 2025-04-22T00:00:00 | [
[
"Laine",
"Timo Aukusti",
""
]
] |
2503.10735 | Najmeh Miri | Najmeh Miri, Shahrzad Khayatbashi, Jelena Zdravkovic, Amin Jalali | OCPM$^2$: Extending the Process Mining Methodology for Object-Centric
Event Data Extraction | null | null | null | null | cs.DB cs.AI | http://creativecommons.org/licenses/by/4.0/ | Object-Centric Process Mining (OCPM) enables business process analysis from
multiple perspectives. For example, an educational path can be examined from
the viewpoints of students, teachers, and groups. This analysis depends on
Object-Centric Event Data (OCED), which captures relationships between events
and object types, representing different perspectives. Unlike traditional
process mining techniques, extracting OCED minimizes the need for repeated log
extractions when shifting the analytical focus. However, recording these
complex relationships increases the complexity of the log extraction process.
To address this challenge, this paper proposes a methodology for extracting
OCED based on PM\inst{2}, a well-established process mining framework. Our
approach introduces a structured framework that guides data analysts and
engineers in extracting OCED for process analysis. We validate this framework
by applying it in a real-world educational setting, demonstrating its
effectiveness in extracting an Object-Centric Event Log (OCEL), which serves as
the standard format for recording OCED, from a learning management system and
an administrative grading system.
| [
{
"version": "v1",
"created": "Thu, 13 Mar 2025 15:30:10 GMT"
},
{
"version": "v2",
"created": "Mon, 21 Apr 2025 12:26:40 GMT"
}
] | 2025-04-22T00:00:00 | [
[
"Miri",
"Najmeh",
""
],
[
"Khayatbashi",
"Shahrzad",
""
],
[
"Zdravkovic",
"Jelena",
""
],
[
"Jalali",
"Amin",
""
]
] |
2503.11158 | Ulrich Jentschura | Ulrich D. Jentschura | Radiative Energy and Mass Shifts of Quantum Cyclotron States | 5 pages; RevTeX; Eur.Phys.J.D, in press; accepted for the PSAS 2024
Special Issue on Precision Physics | Eur.Phys.J. D 79 (2025) 31 | 10.1140/epjd/s10053-025-00982-3 | null | hep-ph | http://creativecommons.org/licenses/by/4.0/ | We discuss relativistic and radiative corrections to the energies of quantum
cyclotron states. In particular, it is shown analytically that the leading
logarithmic radiative (self-energy) correction to the bound-state energy levels
of quantum cyclotron states is state-independent, and must be interpreted as a
magnetic-field-dependent correction to the electron's mass in a Penning trap.
| [
{
"version": "v1",
"created": "Fri, 14 Mar 2025 07:53:08 GMT"
}
] | 2025-04-22T00:00:00 | [
[
"Jentschura",
"Ulrich D.",
""
]
] |
2503.11407 | Amol Aggarwal | Amol Aggarwal | Effective Velocities in the Toda Lattice | 70 pages, no figures. arXiv admin note: text overlap with
arXiv:2503.08018; Version 2: Edits to make terminology more consistent with
physics literature; Version 3: Added references and exposition | null | null | null | math-ph math.DS math.MP math.PR nlin.SI | http://creativecommons.org/licenses/by/4.0/ | In this paper we consider the Toda lattice $(\boldsymbol{p}(t);
\boldsymbol{q}(t))$ at thermal equilibrium, meaning that its variables $(p_i)$
and $(e^{q_i-q_{i+1}})$ are independent Gaussian and Gamma random variables,
respectively. This model can be thought of a dense collection of many
``quasiparticles'' that act as solitons. We establish a law of large numbers
for the trajectory of these quasiparticles, showing that they travel with
approximately constant velocities, which are explicit. Our proof is based on a
direct analysis of the asymptotic scattering relation, an equation (proven in
previous work of the author) that approximately governs the dynamics of
quasiparticles locations. This makes use of a regularization argument that
essentially linearizes this relation, together with concentration estimates for
the Toda lattice's (random) Lax matrix.
| [
{
"version": "v1",
"created": "Fri, 14 Mar 2025 13:48:24 GMT"
},
{
"version": "v2",
"created": "Thu, 10 Apr 2025 04:17:56 GMT"
},
{
"version": "v3",
"created": "Mon, 21 Apr 2025 13:34:14 GMT"
}
] | 2025-04-22T00:00:00 | [
[
"Aggarwal",
"Amol",
""
]
] |
2503.11873 | Zhuoqi Li | Zhuoqi Li, Shuyuan Huyan, Elizabeth C. Thompson, Tyler J. Slade,
Dongzhou Zhang, Young J. Ryu, Wenli Bi, Sergey L. Bud'ko, Paul C. Canfield | Superconducting dome and structural changes in LaRu$_3$Si$_2$ under
pressure | 12 pages, 16 figures | null | 10.1103/PhysRevB.111.144505 | null | cond-mat.supr-con | http://creativecommons.org/licenses/by/4.0/ | LaRu$_3$Si$_2$ is of current research interest as a kagome metal with a
superconducting transition temperature, $T_c\sim$7 K and higher temperature
charge density wave (CDW) orders. Here we report electrical transport and X-ray
diffraction measurements on LaRu$_3$Si$_2$ under pressure up to 65 GPa and 35
GPa respectively. The superconducting transition temperature $T_c$ first gets
slightly enhanced and reaches a maximum $\sim$8.7 K at $\sim$8.5 GPa. With
further applied pressure, $T_c$ is initially gradually suppressed, then more
rapidly suppressed, followed by gradual suppression, revealing a
superconducting dome. Two possible pressure-induced structural phase
transitions are also observed at room temperature, from original hexagonal
phase to another hexagonal structure above $\sim$11.5 GPa, and further to a
structure with lower symmetry above $\sim$23.5 GPa. These transition pressures
roughly correlate with features found in our pressure dependent transport data.
| [
{
"version": "v1",
"created": "Fri, 14 Mar 2025 20:56:15 GMT"
},
{
"version": "v2",
"created": "Fri, 18 Apr 2025 23:49:37 GMT"
}
] | 2025-04-22T00:00:00 | [
[
"Li",
"Zhuoqi",
""
],
[
"Huyan",
"Shuyuan",
""
],
[
"Thompson",
"Elizabeth C.",
""
],
[
"Slade",
"Tyler J.",
""
],
[
"Zhang",
"Dongzhou",
""
],
[
"Ryu",
"Young J.",
""
],
[
"Bi",
"Wenli",
""
],
[
"Bud'ko",
"Sergey L.",
""
],
[
"Canfield",
"Paul C.",
""
]
] |
2503.12033 | Li Jiaying | Jiaying Li, Yuanwei Liu, and Hong Xing | Unsupervised Learning for AoD Estimation in MISO Downlink LoS
Transmissions | 5 pages, 3 figures and 1 table, submitted for possible publication | null | null | null | eess.SP cs.IT math.IT | http://arxiv.org/licenses/nonexclusive-distrib/1.0/ | With the emergence of simultaneous localization and communication (SLAC), it
becomes more and more attractive to perform angle of departure (AoD) estimation
at the receiving Internet of Thing (IoT) user end for improved positioning
accuracy, flexibility and enhanced user privacy. To address challenges like a
large number of real-time measurements required for latency-critical
applications and enormous data collection for training deep learning models in
conventional AoD estimation methods, we propose in this letter an unsupervised
learning framework, which unifies training for both deterministic maximum
likelihood (DML) and stochastic maximum likelihood (SML) based AoD estimation
in multiple-input single-output (MISO) downlink (DL) wireless transmissions.
Specifically, under the line-of-sight (LoS) assumption, we incorporate both the
received signals and pilot-sequence information, as per its availability at the
DL user, into the input of the deep learning model, and adopt a common neural
network architecture compatible with input data in both DML and SML cases.
Extensive numerical results validate that the proposed unsupervised learning
based AoD estimation not only improves estimation accuracy, but also
significantly reduces required number of observations, thereby reducing both
estimation overhead and latency compared to various benchmarks.
| [
{
"version": "v1",
"created": "Sat, 15 Mar 2025 07:53:01 GMT"
},
{
"version": "v2",
"created": "Sat, 19 Apr 2025 16:48:45 GMT"
}
] | 2025-04-22T00:00:00 | [
[
"Li",
"Jiaying",
""
],
[
"Liu",
"Yuanwei",
""
],
[
"Xing",
"Hong",
""
]
] |
2503.12206 | Ans Munir | Ans Munir, Faisal Z. Qureshi, Muhammad Haris Khan, and Mohsen Ali | TLAC: Two-stage LMM Augmented CLIP for Zero-Shot Classification | Added code link in the abstract | null | null | null | cs.CV | http://creativecommons.org/licenses/by-nc-sa/4.0/ | Contrastive Language-Image Pretraining (CLIP) has shown impressive zero-shot
performance on image classification. However, state-of-the-art methods often
rely on fine-tuning techniques like prompt learning and adapter-based tuning to
optimize CLIP's performance. The necessity for fine-tuning significantly limits
CLIP's adaptability to novel datasets and domains. This requirement mandates
substantial time and computational resources for each new dataset. To overcome
this limitation, we introduce simple yet effective training-free approaches,
Single-stage LMM Augmented CLIP (SLAC) and Two-stage LMM Augmented CLIP (TLAC),
that leverages powerful Large Multimodal Models (LMMs), such as Gemini, for
image classification. The proposed methods leverages the capabilities of
pre-trained LMMs, allowing for seamless adaptation to diverse datasets and
domains without the need for additional training. Our approaches involve
prompting the LMM to identify objects within an image. Subsequently, the CLIP
text encoder determines the image class by identifying the dataset class with
the highest semantic similarity to the LLM predicted object. Our models
achieved superior accuracy on 9 of 11 base-to-novel datasets, including
ImageNet, SUN397, and Caltech101, while maintaining a strictly training-free
paradigm. Our TLAC model achieved an overall accuracy of 83.44%, surpassing the
previous state-of-the-art few-shot methods by a margin of 6.75%. Compared to
other training-free approaches, our TLAC method achieved 83.6% average accuracy
across 13 datasets, a 9.7% improvement over the previous methods. Our Code is
available at https://github.com/ans92/TLAC
| [
{
"version": "v1",
"created": "Sat, 15 Mar 2025 17:11:41 GMT"
},
{
"version": "v2",
"created": "Mon, 21 Apr 2025 04:44:14 GMT"
}
] | 2025-04-22T00:00:00 | [
[
"Munir",
"Ans",
""
],
[
"Qureshi",
"Faisal Z.",
""
],
[
"Khan",
"Muhammad Haris",
""
],
[
"Ali",
"Mohsen",
""
]
] |
2503.12555 | Mengdi Yin | Mengdi Yin and Jing Zhang and Dimitri D Vvedensky | Density-Functional Theory and Triply-Periodic Minimal Surfaces | null | null | null | null | cond-mat.mtrl-sci | http://arxiv.org/licenses/nonexclusive-distrib/1.0/ | Several authors have suggested that the surfaces of vanishing potential
generated by the electrostatic fields from a distribution of point charges
resemble triply periodic minimal surfaces (TPMS) corresponding to the positions
of the point charges. We provide a theoretical basis for this phenomenological
comparison by starting with the Boltzmann equation to show that the surface
corresponding to zero charge density is a minimal surface. We then use
density-functional calculations for elemental materials that differ
electronically and structurally, Na, Cu, and Al, to show that surfaces of
vanishing charge density converge to the corresponding TPMS.
| [
{
"version": "v1",
"created": "Sun, 16 Mar 2025 15:54:47 GMT"
},
{
"version": "v2",
"created": "Sat, 19 Apr 2025 02:57:51 GMT"
}
] | 2025-04-22T00:00:00 | [
[
"Yin",
"Mengdi",
""
],
[
"Zhang",
"Jing",
""
],
[
"Vvedensky",
"Dimitri D",
""
]
] |
2503.12602 | Teresa Head-Gordon | Kunyang Sun, Dorian Bagni, Joseph M. Cavanagh, Yingze Wang, Jacob M.
Sawyer, Andrew Gritsevskiy, Oufan Zhang, Teresa Head-Gordon | SynLlama: Generating Synthesizable Molecules and Their Analogs with
Large Language Models | null | null | null | null | cs.LG physics.bio-ph | http://creativecommons.org/licenses/by-sa/4.0/ | Generative machine learning models for small molecule drug discovery have
shown immense promise, but many molecules they generate are too difficult to
synthesize, making them impractical for further investigation or development.
In this work, we present a novel approach by fine-tuning Meta's Llama3 Large
Language Models (LLMs) to create SynLlama, which generates full synthetic
pathways made of commonly accessible building blocks and robust organic
reaction templates. SynLlama explores a large synthesizable space using
significantly less data compared to other state-of-the-art methods, and offers
strong performance in bottom-up synthesis, synthesizable analog generation, and
hit expansion, offering medicinal chemists a valuable tool for drug discovery
developments. We find that SynLlama, even without training on external building
blocks, can effectively generalize to unseen yet purchasable building blocks,
meaning that its reconstruction capabilities extend to a broader synthesizable
chemical space than the training data. We also demonstrate the use of SynLlama
in a pharmaceutical context for synthesis planning of analog molecules and hit
expansion leads for proposed inhibitors of target proteins.
| [
{
"version": "v1",
"created": "Sun, 16 Mar 2025 18:30:56 GMT"
},
{
"version": "v2",
"created": "Tue, 15 Apr 2025 20:08:03 GMT"
},
{
"version": "v3",
"created": "Fri, 18 Apr 2025 22:32:14 GMT"
}
] | 2025-04-22T00:00:00 | [
[
"Sun",
"Kunyang",
""
],
[
"Bagni",
"Dorian",
""
],
[
"Cavanagh",
"Joseph M.",
""
],
[
"Wang",
"Yingze",
""
],
[
"Sawyer",
"Jacob M.",
""
],
[
"Gritsevskiy",
"Andrew",
""
],
[
"Zhang",
"Oufan",
""
],
[
"Head-Gordon",
"Teresa",
""
]
] |
2503.13031 | Jakob Sponholz | Jakob Sponholz, Andreas Weilinghoff and Juliane Schopf | Halving transcription time: A fast, user-friendly and GDPR-compliant
workflow to create AI-assisted transcripts for content analysis | 10 pages, 1 table | null | null | null | cs.CL | http://creativecommons.org/licenses/by/4.0/ | In qualitative research, data transcription is often labor-intensive and
time-consuming. To expedite this process, a workflow utilizing artificial
intelligence (AI) was developed. This workflow not only enhances transcription
speed but also addresses the issue of AI-generated transcripts often lacking
compatibility with standard content analysis software. Within this workflow,
automatic speech recognition is employed to create initial transcripts from
audio recordings, which are then formatted to be compatible with content
analysis software such as ATLAS or MAXQDA. Empirical data from a study of 12
interviews suggests that this workflow can reduce transcription time by up to
76.4%. Furthermore, by using widely used standard software, this process is
suitable for both students and researchers while also being adaptable to a
variety of learning, teaching, and research environments. It is also
particularly beneficial for non-native speakers. In addition, the workflow is
GDPR-compliant and facilitates local, offline transcript generation, which is
crucial when dealing with sensitive data.
| [
{
"version": "v1",
"created": "Mon, 17 Mar 2025 10:33:39 GMT"
},
{
"version": "v2",
"created": "Sat, 19 Apr 2025 19:15:04 GMT"
}
] | 2025-04-22T00:00:00 | [
[
"Sponholz",
"Jakob",
""
],
[
"Weilinghoff",
"Andreas",
""
],
[
"Schopf",
"Juliane",
""
]
] |
2503.13395 | Erik Hoel | Erik Hoel | Causal Emergence 2.0: Quantifying emergent complexity | v3: minor wording edits; added Supplementary Information (1 new
figure); results unchanged | null | null | null | cs.IT math.IT | http://arxiv.org/licenses/nonexclusive-distrib/1.0/ | Complex systems can be described at myriad different scales, and their causal
workings often have multiscale structure (e.g., a computer can be described at
the microscale of its hardware circuitry, the mesoscale of its machine code,
and the macroscale of its operating system). While scientists study and model
systems across the full hierarchy of their scales, from microphysics to
macroeconomics, there is debate about what the macroscales of systems can
possibly add beyond mere compression. To resolve this longstanding issue, here
a new theory of emergence is introduced wherein the different scales of a
system are treated like slices of a higher-dimensional object. The theory can
distinguish which of these scales possess unique causal contributions, and
which are not causally relevant. Constructed from an axiomatic notion of
causation, the theory's application is demonstrated in coarse-grains of Markov
chains. It identifies all cases of macroscale causation: instances where
reduction to a microscale is possible, yet lossy about causation. Furthermore,
the theory posits a causal apportioning schema that calculates the causal
contribution of each scale, showing what each uniquely adds. Finally, it
reveals a novel measure of emergent complexity: how widely distributed a
system's causal workings are across its hierarchy of scales.
| [
{
"version": "v1",
"created": "Mon, 17 Mar 2025 17:28:46 GMT"
},
{
"version": "v2",
"created": "Mon, 31 Mar 2025 17:39:16 GMT"
},
{
"version": "v3",
"created": "Mon, 21 Apr 2025 17:51:57 GMT"
}
] | 2025-04-22T00:00:00 | [
[
"Hoel",
"Erik",
""
]
] |
2503.13959 | Vijay Kumar | Vijay Kumar, S. P. Ram, S. Singh, Kavish Bhardwaj, V. B. Tiwari and S.
R. Mishra | On damping of Rabi oscillations in two-photon Raman excitation in cold
$^{87}$Rb atoms | null | null | null | null | physics.atom-ph | http://arxiv.org/licenses/nonexclusive-distrib/1.0/ | Two-photon Raman excitation between the ground hyperfine states $|5 \
^2S_{1/2}, F = 2\rangle$ and $|5 \ ^2S_{1/2}, F = 1\rangle$ of $^{87}$Rb atom
has been experimentally studied. The Rabi coupling strengths of various
transition involved have been calculated in presence of a weak magnetic field.
A density matrix formalism has been developed to understand the experimentally
observed damping of Rabi oscillations of population in a hyperfine state of
$^{87}$Rb atom during interaction with the Raman laser beams. The observed
damping of Rabi oscillations has been attributed to the dephasing during the
light atom interaction.
| [
{
"version": "v1",
"created": "Tue, 18 Mar 2025 06:51:36 GMT"
},
{
"version": "v2",
"created": "Mon, 21 Apr 2025 03:54:52 GMT"
}
] | 2025-04-22T00:00:00 | [
[
"Kumar",
"Vijay",
""
],
[
"Ram",
"S. P.",
""
],
[
"Singh",
"S.",
""
],
[
"Bhardwaj",
"Kavish",
""
],
[
"Tiwari",
"V. B.",
""
],
[
"Mishra",
"S. R.",
""
]
] |
2503.13993 | Tatiana Todorova L | Temenoujka P. Peneva and Tatiana L. Todorova | On a Diophantine Inequality with Primes Yielding Square-Free Sums with
Given Numbers | null | null | null | null | math.NT | http://creativecommons.org/licenses/by/4.0/ | Let $\alpha\in \mathbb{R}\setminus\mathbb{Q}$ and $\beta\in \mathbb{R}$ be
given. Suppose that $a_1,\ldots,a_s$ are distinct positive integers that do not
contain a reduced residue system modulo $p^2$ for any prime $p$. We prove that
there exist infinitely many primes $p$ such that the inequality $||\alpha
p+\beta||<p^{-1/10}$ holds and all the numbers $p+a_1,\ldots,p+a_s$ are
square-free.
| [
{
"version": "v1",
"created": "Tue, 18 Mar 2025 07:54:34 GMT"
},
{
"version": "v2",
"created": "Sat, 19 Apr 2025 16:58:07 GMT"
}
] | 2025-04-22T00:00:00 | [
[
"Peneva",
"Temenoujka P.",
""
],
[
"Todorova",
"Tatiana L.",
""
]
] |
2503.14027 | Kanji Mori | Kanji Mori, Tomoya Takiwaki, Kazunori Kohri, Hiroki Nagakura | Core-collapse supernova explosions hindered by eV-mass sterile neutrinos | 14 pages, 12 figures, 1 table, accepted for publication in PRD | null | null | KEK-Cosmo-0376, KEK-TH-2709 | astro-ph.HE hep-ph | http://creativecommons.org/licenses/by/4.0/ | Light sterile neutrinos, $\nu_s$, are often introduced to explain an
anomalous deficit in the electron antineutrino flux from nuclear reactors. If
they exist, sterile neutrinos would also be produced in collapsing massive
stars through the active-sterile neutrino oscillation. In order to investigate
the impacts of sterile neutrinos on supernova dynamics, we perform
two-dimensional neutrino-radiation hydrodynamic simulations of stellar
core-collapse coupled with the active-sterile oscillation through the
Mikheyev-Smirnov-Wolfenstein effect. As the initial condition of our
simulations, we adopt a blue supergiant model that is tuned to reproduce
observational features of the SN 1987A progenitor to compare our models with
observations of the event. It is found that the active-sterile oscillation
reduces the $\nu_{e}$ and $\bar{\nu}_e$ fluxes and decreases the explosion
energy. We also find that, if the mixing angle $\theta$ and the mass difference
$\delta m_\mathrm{s}^2$ between $\nu_e$ and $\nu_s$ are large enough, the star
fails to explode. This suggests that these mixing parameters relevant to
sterile neutrinos could be constrained by supernova explodability, though other
uncertainties in supernova theory need to be addressed to refine them. In
addition, we predict neutrino signals from a nearby supernova event and find
that the neutrino event number can significantly decrease because the $\nu_e$
and $\bar{\nu}_e$ fluxes are reduced. In particular, DUNE observations of
$\nu_e$ will be useful to search for a signature of sterile neutrinos with a
tiny mixing angle because a smaller mixing angle leads to a larger effect on
the $\nu_e$ flux.
| [
{
"version": "v1",
"created": "Tue, 18 Mar 2025 08:37:20 GMT"
},
{
"version": "v2",
"created": "Thu, 20 Mar 2025 08:27:33 GMT"
},
{
"version": "v3",
"created": "Fri, 4 Apr 2025 09:03:58 GMT"
},
{
"version": "v4",
"created": "Sun, 20 Apr 2025 15:22:27 GMT"
}
] | 2025-04-22T00:00:00 | [
[
"Mori",
"Kanji",
""
],
[
"Takiwaki",
"Tomoya",
""
],
[
"Kohri",
"Kazunori",
""
],
[
"Nagakura",
"Hiroki",
""
]
] |
2503.14151 | Yong Zhong | Yong Zhong, Zhuoyi Yang, Jiayan Teng, Xiaotao Gu, Chongxuan Li | Concat-ID: Towards Universal Identity-Preserving Video Synthesis | null | null | null | null | cs.CV cs.AI | http://arxiv.org/licenses/nonexclusive-distrib/1.0/ | We present Concat-ID, a unified framework for identity-preserving video
generation. Concat-ID employs Variational Autoencoders to extract image
features, which are concatenated with video latents along the sequence
dimension, leveraging solely 3D self-attention mechanisms without the need for
additional modules. A novel cross-video pairing strategy and a multi-stage
training regimen are introduced to balance identity consistency and facial
editability while enhancing video naturalness. Extensive experiments
demonstrate Concat-ID's superiority over existing methods in both single and
multi-identity generation, as well as its seamless scalability to multi-subject
scenarios, including virtual try-on and background-controllable generation.
Concat-ID establishes a new benchmark for identity-preserving video synthesis,
providing a versatile and scalable solution for a wide range of applications.
| [
{
"version": "v1",
"created": "Tue, 18 Mar 2025 11:17:32 GMT"
},
{
"version": "v2",
"created": "Sat, 19 Apr 2025 09:26:43 GMT"
}
] | 2025-04-22T00:00:00 | [
[
"Zhong",
"Yong",
""
],
[
"Yang",
"Zhuoyi",
""
],
[
"Teng",
"Jiayan",
""
],
[
"Gu",
"Xiaotao",
""
],
[
"Li",
"Chongxuan",
""
]
] |
2503.14438 | Noam Soker | Shlomi Hillel, Ron Schreier, Noam Soker (Technion, Israel) | Forming a clumpy circumstellar material in energetic pre-supernova
activity | Accepted for publication in Research in Astronomy and Astrophysics | null | null | null | astro-ph.HE | http://creativecommons.org/licenses/by/4.0/ | We demonstrate by three-dimensional hydrodynamical simulations of energy
deposition into the envelope of a red supergiant (RSG) model the inflation of a
Rayleigh-Taylor unstable envelope that forms a compact clumpy circumstellar
material (CSM). Our simulations mimic vigorous core activity years to months
before a core-collapse supernova (CCSN) explosion that deposits energy to the
outer envelope. The fierce core nuclear activity in the pre-CCSN explosion
phase might excite waves that propagate to the envelope. The wave energy is
dissipated where envelope convection cannot carry the energy. We deposit this
energy into a shell in the outer envelope with a power of L(wave)=2.6e6Lo or
L(wave)=5.2e5Lo for 0.32 year. The energy-deposition shell expands while its
pressure is higher than its surroundings, but its density is lower. Therefore,
this expansion is Rayleigh-Taylor unstable and develops instability fingers.
Most of the inflated envelope does not reach the escape velocity in the year of
simulation but forms a compact and clumpy CSM. The high density of the inflated
envelope implies that if a companion is present in that zone, it will accrete
mass at a very high rate and power a pre-explosion outburst.
| [
{
"version": "v1",
"created": "Tue, 18 Mar 2025 17:15:47 GMT"
},
{
"version": "v2",
"created": "Mon, 21 Apr 2025 09:04:31 GMT"
}
] | 2025-04-22T00:00:00 | [
[
"Hillel",
"Shlomi",
"",
"Technion, Israel"
],
[
"Schreier",
"Ron",
"",
"Technion, Israel"
],
[
"Soker",
"Noam",
"",
"Technion, Israel"
]
] |
2503.14453 | Qiushuo Hou | Qiushuo Hou, Sangwoo Park, Matteo Zecchin, Yunlong Cai, Guanding Yu,
Osvaldo Simeone | Online Conformal Probabilistic Numerics via Adaptive Edge-Cloud
Offloading | This paper has been submitted to a conference | null | null | null | stat.ML cs.LG | http://arxiv.org/licenses/nonexclusive-distrib/1.0/ | Consider an edge computing setting in which a user submits queries for the
solution of a linear system to an edge processor, which is subject to
time-varying computing availability. The edge processor applies a probabilistic
linear solver (PLS) so as to be able to respond to the user's query within the
allotted time and computing budget. Feedback to the user is in the form of a
set of plausible solutions. Due to model misspecification, the
highest-probability-density (HPD) set obtained via a direct application of PLS
does not come with coverage guarantees with respect to the true solution of the
linear system. This work introduces a new method to calibrate the HPD sets
produced by PLS with the aim of guaranteeing long-term coverage requirements.
The proposed method, referred to as online conformal prediction-PLS (OCP-PLS),
assumes sporadic feedback from cloud to edge. This enables the online
calibration of uncertainty thresholds via online conformal prediction (OCP), an
online optimization method previously studied in the context of prediction
models. The validity of OCP-PLS is verified via experiments that bring insights
into trade-offs between coverage, prediction set size, and cloud usage.
| [
{
"version": "v1",
"created": "Tue, 18 Mar 2025 17:30:26 GMT"
},
{
"version": "v2",
"created": "Sat, 19 Apr 2025 14:40:19 GMT"
}
] | 2025-04-22T00:00:00 | [
[
"Hou",
"Qiushuo",
""
],
[
"Park",
"Sangwoo",
""
],
[
"Zecchin",
"Matteo",
""
],
[
"Cai",
"Yunlong",
""
],
[
"Yu",
"Guanding",
""
],
[
"Simeone",
"Osvaldo",
""
]
] |
2503.14747 | Ivan Canay | Federico A. Bugni, Ivan A. Canay, and Deborah Kim | Testing Conditional Stochastic Dominance at Target Points | null | null | null | null | econ.EM math.ST stat.TH | http://creativecommons.org/licenses/by/4.0/ | This paper introduces a novel test for conditional stochastic dominance (CSD)
at specific values of the conditioning covariates, referred to as target
points. The test is relevant for analyzing income inequality, evaluating
treatment effects, and studying discrimination. We propose a
Kolmogorov--Smirnov-type test statistic that utilizes induced order statistics
from independent samples. Notably, the test features a data-independent
critical value, eliminating the need for resampling techniques such as the
bootstrap. Our approach avoids kernel smoothing and parametric assumptions,
instead relying on a tuning parameter to select relevant observations. We
establish the asymptotic properties of our test, showing that the induced order
statistics converge to independent draws from the true conditional
distributions and that the test is asymptotically of level $\alpha$ under weak
regularity conditions. While our results apply to both continuous and discrete
data, in the discrete case, the critical value only provides a valid upper
bound. To address this, we propose a refined critical value that significantly
enhances power, requiring only knowledge of the support size of the
distributions. Additionally, we analyze the test's behavior in the limit
experiment, demonstrating that it reduces to a problem analogous to testing
unconditional stochastic dominance in finite samples. This framework allows us
to prove the validity of permutation-based tests for stochastic dominance when
the random variables are continuous. Monte Carlo simulations confirm the strong
finite-sample performance of our method.
| [
{
"version": "v1",
"created": "Tue, 18 Mar 2025 21:22:55 GMT"
},
{
"version": "v2",
"created": "Sun, 20 Apr 2025 15:15:52 GMT"
}
] | 2025-04-22T00:00:00 | [
[
"Bugni",
"Federico A.",
""
],
[
"Canay",
"Ivan A.",
""
],
[
"Kim",
"Deborah",
""
]
] |
2503.15084 | Zhengyan Liu | Zhengyan Liu, Ji-an Jiang, Wen Zhao | A constraint on superheavy elements of the GRB-kilonova AT 2023vfi | ApJL accepted | ApJL, 983, L34 (2025) | 10.3847/2041-8213/adc260 | null | astro-ph.HE | http://arxiv.org/licenses/nonexclusive-distrib/1.0/ | The discovery of the kilonova (KN) AT 2017gfo, accompanying the gravitational
wave event GW170817, provides crucial insight into the synthesis of heavy
elements during binary neutron star (BNS) mergers. Following this landmark
event, another KN was detected in association with the second-brightest
gamma-ray burst (GRB) observed to date, GRB 230307A, and subsequently confirmed
by observations of the James Webb Space Telescope (JWST). In this work, we
conduct an end-to-end simulation to analyze the temporal evolution of the KN AT
2023vfi associated with GRB 230307A, and constrain the abundances of superheavy
elements produced. We find that the temporal evolution of AT 2023vfi is similar
to AT 2017gfo in the first week post-burst. Additionally, the
\textit{r}-process nuclide abundances of lanthanide-rich ejecta, derived from
numerical relativity simulations of BNS mergers, can also successfully
interpret the temporal evolution of the KN with the lanthanide-rich ejecta mass
of $0.02 M_\odot$, which is consistent with the mass range of dynamical ejecta
from numerical simulations in literature. Both findings strongly suggest the
hypothesis that GRB 230307A originated from a BNS merger, similar to AT
2017gfo. Based on the first time observation of the KN for JWST, we are able to
constrain the superheavy elements of another KN following AT 2017gfo. The
pre-radioactive-decay abundances of the superheavy nuclides: $^{222}$Rn,
$^{223}$Ra, $^{224}$Ra and $^{225}$Ac, are estimated to be at least on the
order of $1 \times 10^{-5}$. These abundance estimates provide valuable insight
into the synthesis of superheavy elements in BNS mergers, contributing to our
understanding of astrophysical \textit{r}-process nucleosynthesis.
| [
{
"version": "v1",
"created": "Wed, 19 Mar 2025 10:32:32 GMT"
},
{
"version": "v2",
"created": "Fri, 21 Mar 2025 11:57:11 GMT"
}
] | 2025-04-22T00:00:00 | [
[
"Liu",
"Zhengyan",
""
],
[
"Jiang",
"Ji-an",
""
],
[
"Zhao",
"Wen",
""
]
] |
2503.15547 | Juhee Kim | Juhee Kim, Woohyuk Choi, Byoungyoung Lee | Prompt Flow Integrity to Prevent Privilege Escalation in LLM Agents | null | null | null | null | cs.CR cs.AI cs.MA | http://creativecommons.org/licenses/by/4.0/ | Large Language Models (LLMs) are combined with tools to create powerful LLM
agents that provide a wide range of services. Unlike traditional software, LLM
agent's behavior is determined at runtime by natural language prompts from
either user or tool's data. This flexibility enables a new computing paradigm
with unlimited capabilities and programmability, but also introduces new
security risks, vulnerable to privilege escalation attacks. Moreover, user
prompts are prone to be interpreted in an insecure way by LLM agents, creating
non-deterministic behaviors that can be exploited by attackers. To address
these security risks, we propose Prompt Flow Integrity (PFI), a system
security-oriented solution to prevent privilege escalation in LLM agents.
Analyzing the architectural characteristics of LLM agents, PFI features three
mitigation techniques -- i.e., agent isolation, secure untrusted data
processing, and privilege escalation guardrails. Our evaluation result shows
that PFI effectively mitigates privilege escalation attacks while successfully
preserving the utility of LLM agents.
| [
{
"version": "v1",
"created": "Mon, 17 Mar 2025 05:27:57 GMT"
},
{
"version": "v2",
"created": "Mon, 21 Apr 2025 02:10:14 GMT"
}
] | 2025-04-22T00:00:00 | [
[
"Kim",
"Juhee",
""
],
[
"Choi",
"Woohyuk",
""
],
[
"Lee",
"Byoungyoung",
""
]
] |
2503.16483 | Guowu Yuan | Mingming Wang, Guowu Yuan, Hailan He, Chengming Tan, Hao Wu, Hao Zhou | Multi-category solar radio burst detection based on task-aligned
one-stage object detection model | 16 pages, 10 figures (Accepted for publication in Astrophysics and
Space Science) | null | 10.1007/s10509-025-04414-x | null | astro-ph.IM | http://creativecommons.org/publicdomain/zero/1.0/ | Accurate identification of solar radio bursts (SRBs) is essential for
advancing research in solar physics and predicting space weather. However, the
majority of current studies mainly concentrate on detecting whether SRBs are
present or absent, often focusing on only one particular type of burst.
Moreover, the neural network models used for SRB detection are typically
complex, involving a large number of parameters, which results in slower
processing speeds. This study establishes a dataset encompassing Type II, Type
III, Type IIIs, Type IV, and Type V SRBs collected from e-CALLISTO, including
8,752 SRB spectrum images and achieving annotations for 10,822 SRBs. We propose
a multi-category SRB detection model based on task-aligned one-stage object
detection (TOOD). TOOD can solve the problem of inconsistent predictions in
classification and localization tasks, and it improves the detection recall
rate. This model aligns classification and localization tasks and optimizes the
neck network by incorporating a channel attention mechanism. This model
achieves higher recall and accuracy with fewer parameters. This model can
accurately detect five types of SBRs. The experimental results show that the
model achieved an accuracy of 79.9\% (AP50) and a recall rate of 95.1\% on the
SBRs dataset. A higher recall rate than other models means fewer SRBs are
missed in automatic detection. The model we propose has the potential to make a
substantial impact on solar physics research and space weather studies.
Additionally, the findings in this paper could provide valuable insights for
processing other small-sample astronomical datasets.The source code and data is
available at https://github.com/onewangqianqian/MobileNetVitv2-TOOD.git.
| [
{
"version": "v1",
"created": "Fri, 7 Mar 2025 14:38:19 GMT"
}
] | 2025-04-22T00:00:00 | [
[
"Wang",
"Mingming",
""
],
[
"Yuan",
"Guowu",
""
],
[
"He",
"Hailan",
""
],
[
"Tan",
"Chengming",
""
],
[
"Wu",
"Hao",
""
],
[
"Zhou",
"Hao",
""
]
] |
2503.16588 | Jan Reineke | Sophie Kahlen, Jan Reineke | A Unified Framework for Quantitative Cache Analysis | Extended version of RTAS 2025 paper | null | null | null | cs.PL | http://arxiv.org/licenses/nonexclusive-distrib/1.0/ | In this work we unify two existing lines of work towards cache analysis for
non-LRU policies. To this end, we extend the notion of competitiveness to block
competitiveness and systematically analyze the competitiveness and block
competitiveness of FIFO and MRU relative to LRU for arbitrary associativities.
We show how competitiveness and block competitiveness can be exploited in
state-of-the-art WCET analysis based on the results of existing persistence
analyses for LRU. Unlike prior work, our approach is applicable to
microarchitectures that exhibit timing anomalies. We experimentally evaluate
the precision and cost of our approach on benchmarks from TACLeBench. The
experiments demonstrate that quantitative cache analysis for FIFO and MRU comes
close to the precision of LRU.
| [
{
"version": "v1",
"created": "Thu, 20 Mar 2025 17:37:15 GMT"
},
{
"version": "v2",
"created": "Sun, 30 Mar 2025 09:46:34 GMT"
},
{
"version": "v3",
"created": "Mon, 21 Apr 2025 15:36:53 GMT"
}
] | 2025-04-22T00:00:00 | [
[
"Kahlen",
"Sophie",
""
],
[
"Reineke",
"Jan",
""
]
] |
2503.16791 | Zijian Ding | Zijian Ding, Michelle Brachman, Joel Chan, Werner Geyer | "The Diagram is like Guardrails": Structuring GenAI-assisted Hypotheses
Exploration with an Interactive Shared Representation | null | null | null | null | cs.HC cs.AI | http://creativecommons.org/licenses/by-nc-nd/4.0/ | Data analysis encompasses a spectrum of tasks, from high-level conceptual
reasoning to lower-level execution. While AI-powered tools increasingly support
execution tasks, there remains a need for intelligent assistance in conceptual
tasks. This paper investigates the design of an ordered node-link tree
interface augmented with AI-generated information hints and visualizations, as
a potential shared representation for hypothesis exploration. Through a design
probe (n=22), participants generated diagrams averaging 21.82 hypotheses. Our
findings showed that the node-link diagram acts as "guardrails" for hypothesis
exploration, facilitating structured workflows, providing comprehensive
overviews, and enabling efficient backtracking. The AI-generated information
hints, particularly visualizations, aided users in transforming abstract ideas
into data-backed concepts while reducing cognitive load. We further discuss how
node-link diagrams can support both parallel exploration and iterative
refinement in hypothesis formulation, potentially enhancing the breadth and
depth of human-AI collaborative data analysis.
| [
{
"version": "v1",
"created": "Fri, 21 Mar 2025 02:01:37 GMT"
},
{
"version": "v2",
"created": "Mon, 21 Apr 2025 16:05:54 GMT"
}
] | 2025-04-22T00:00:00 | [
[
"Ding",
"Zijian",
""
],
[
"Brachman",
"Michelle",
""
],
[
"Chan",
"Joel",
""
],
[
"Geyer",
"Werner",
""
]
] |
2503.16816 | Yi Niu | Yi Niu, Jiashuai Liu, Yingkang Zhan, Jiangbo Shi, Di Zhang, Marika
Reinius, Ines Machado, Mireia Crispin-Ortuzar, Jialun Wu, Chen Li, Zeyu Gao | PH2ST:ST-Prompt Guided Histological Hypergraph Learning for Spatial Gene
Expression Prediction | null | null | null | null | cs.CV | http://arxiv.org/licenses/nonexclusive-distrib/1.0/ | Spatial Transcriptomics (ST) reveals the spatial distribution of gene
expression in tissues, offering critical insights into biological processes and
disease mechanisms. However, the high cost, limited coverage, and technical
complexity of current ST technologies restrict their widespread use in clinical
and research settings, making obtaining high-resolution transcriptomic profiles
across large tissue areas challenging. Predicting ST from H\&E-stained
histology images has emerged as a promising alternative to address these
limitations but remains challenging due to the heterogeneous relationship
between histomorphology and gene expression, which is affected by substantial
variability across patients and tissue sections. In response, we propose PH2ST,
a prompt-guided hypergraph learning framework, which leverages limited ST
signals to guide multi-scale histological representation learning for accurate
and robust spatial gene expression prediction. Extensive evaluations on two
public ST datasets and multiple prompt sampling strategies simulating
real-world scenarios demonstrate that PH2ST not only outperforms existing
state-of-the-art methods, but also shows strong potential for practical
applications such as imputing missing spots, ST super-resolution, and
local-to-global prediction, highlighting its value for scalable and
cost-effective spatial gene expression mapping in biomedical contexts.
| [
{
"version": "v1",
"created": "Fri, 21 Mar 2025 03:10:43 GMT"
},
{
"version": "v2",
"created": "Sun, 20 Apr 2025 09:02:29 GMT"
}
] | 2025-04-22T00:00:00 | [
[
"Niu",
"Yi",
""
],
[
"Liu",
"Jiashuai",
""
],
[
"Zhan",
"Yingkang",
""
],
[
"Shi",
"Jiangbo",
""
],
[
"Zhang",
"Di",
""
],
[
"Reinius",
"Marika",
""
],
[
"Machado",
"Ines",
""
],
[
"Crispin-Ortuzar",
"Mireia",
""
],
[
"Wu",
"Jialun",
""
],
[
"Li",
"Chen",
""
],
[
"Gao",
"Zeyu",
""
]
] |
2503.17012 | Ziqi Ji | Ziqi Ji, Gang Du, Penghao Duan | Learning Non-Ideal Single Vortex Flows Using the Differentiable Vortex
Particle Method | null | null | null | null | physics.flu-dyn physics.comp-ph | http://creativecommons.org/licenses/by/4.0/ | This study extends the differentiable vortex particle method (DVPM) beyond
idealized flow scenarios to encompass more realistic, non-ideal conditions,
including viscous flow and flow subjected to non-conservative body forces. We
establish the Lamb-Oseen vortex as a benchmark case, representing a fundamental
viscous single vortex flow in fluid mechanics. This selection offers
significant analytical advantages, as the Lamb-Oseen vortex possesses an exact
analytical solution derived from the Navier-Stokes (NS) equations, thereby
providing definitive ground truth data for training and validation purposes.
Through rigorous evaluation across a spectrum of Reynolds numbers, we
demonstrate that DVPM achieves superior accuracy in modeling the Lamb-Oseen
vortex compared to conventional convolutional neural networks (CNNs) and
physics-informed neural networks (PINNs). Our results substantiate DVPM's
robust capabilities in modeling non-ideal single vortex flows, establishing its
distinct advantages over contemporary deep learning methodologies in fluid
dynamics applications. The dataset and source code are publicly available on
GitHub at the following link:
https://github.com/jh36714753/Learning_Non-Ideal_Single_Vortex_Flows.
| [
{
"version": "v1",
"created": "Fri, 21 Mar 2025 10:22:34 GMT"
},
{
"version": "v2",
"created": "Sat, 19 Apr 2025 09:18:34 GMT"
}
] | 2025-04-22T00:00:00 | [
[
"Ji",
"Ziqi",
""
],
[
"Du",
"Gang",
""
],
[
"Duan",
"Penghao",
""
]
] |
2503.17119 | Junchen Pei | Haoyu Shang, Jiawei Chen, Rongzhe Hu, Xin Zhen, Chongji Jiang, Junchen
Pei | Many-Body Effects on Nuclear Short Range Correlations | 5 pages, 4 figures | null | null | null | nucl-th | http://creativecommons.org/licenses/by/4.0/ | We reveal nuclear many-body effects on short range correlations by ab initio
no-core shell model calculations of the scaling factor a2. The factor a2
characterizes the abundance of SRC pairs and is linearly related to the EMC
effect. Our study employs the fifth-order N4LO chiral nuclear force without
softening, enabling to distinguish the influences of nuclear states with
different quantum numbers on SRC. It is striking to find that a2 is reduced and
close in triplet isobaric analog states of neighboring nuclei, indicating that
it is insufficient to estimate SRC abundances by considering only mean-field
shell structures. This is explained as specific nuclear states suppress the
formation of deuteron-like component, impacting our understandings of the link
between high-energy partonic properties and low-energy nuclear physics.
| [
{
"version": "v1",
"created": "Fri, 21 Mar 2025 13:12:38 GMT"
},
{
"version": "v2",
"created": "Sat, 19 Apr 2025 08:28:08 GMT"
}
] | 2025-04-22T00:00:00 | [
[
"Shang",
"Haoyu",
""
],
[
"Chen",
"Jiawei",
""
],
[
"Hu",
"Rongzhe",
""
],
[
"Zhen",
"Xin",
""
],
[
"Jiang",
"Chongji",
""
],
[
"Pei",
"Junchen",
""
]
] |
2503.17211 | Zilin Dai | Zilin Dai, Lehong Wang, Fangzhou Lin, Yidong Wang, Zhigang Li,
Kazunori D Yamada, Ziming Zhang, Wang Lu | A Language Anchor-Guided Method for Robust Noisy Domain Generalization | null | null | null | null | cs.CL cs.CV cs.LG | http://creativecommons.org/licenses/by/4.0/ | Real-world machine learning applications often struggle with two major
challenges: distribution shift and label noise. Models tend to overfit by
focusing on redundant and uninformative features in the training data, which
makes it hard for them to generalize to the target domain. Noisy data worsens
this problem by causing further overfitting to the noise, meaning that existing
methods often fail to tell the difference between true, invariant features and
misleading, spurious ones. To tackle these issues, we introduce Anchor
Alignment and Adaptive Weighting (A3W). This new algorithm uses sample
reweighting guided by natural language processing (NLP) anchors to extract more
representative features. In simple terms, A3W leverages semantic
representations from natural language models as a source of domain-invariant
prior knowledge. Additionally, it employs a weighted loss function that adjusts
each sample's contribution based on its similarity to the corresponding NLP
anchor. This adjustment makes the model more robust to noisy labels. Extensive
experiments on standard benchmark datasets show that A3W consistently
outperforms state-of-the-art domain generalization methods, offering
significant improvements in both accuracy and robustness across different
datasets and noise levels.
| [
{
"version": "v1",
"created": "Fri, 21 Mar 2025 15:20:28 GMT"
},
{
"version": "v2",
"created": "Sun, 20 Apr 2025 19:49:45 GMT"
}
] | 2025-04-22T00:00:00 | [
[
"Dai",
"Zilin",
""
],
[
"Wang",
"Lehong",
""
],
[
"Lin",
"Fangzhou",
""
],
[
"Wang",
"Yidong",
""
],
[
"Li",
"Zhigang",
""
],
[
"Yamada",
"Kazunori D",
""
],
[
"Zhang",
"Ziming",
""
],
[
"Lu",
"Wang",
""
]
] |
2503.17401 | Torsten Tiltack | Torsten Tiltack | AIJIM: A Scalable Model for Real-Time AI in Environmental Journalism | 22 pages, 10 figures, 5 tables. Keywords: Artificial Intelligence,
Environmental Journalism, Real-Time Reporting, Vision Transformers, Image
Recognition, Crowdsourced Validation, GPT-4, Automated News Generation, GIS
Integration, Data Privacy Compliance, Explainable AI (XAI), AI Ethics,
Sustainable Development | null | null | null | cs.CY cs.AI cs.HC | http://arxiv.org/licenses/nonexclusive-distrib/1.0/ | Environmental journalism is vital for raising awareness of ecological crises
and supporting evidence-based policy, yet traditional methods suffer from
delays, limited scalability, and lack of coverage in under-monitored regions.
This paper introduces the Artificial Intelligence Journalism Integration Model
(AIJIM), a conceptual and transferable theoretical model that structures
real-time, AI-supported environmental journalism workflows. AIJIM combines
citizen-sourced image data, automated hazard detection, dual-level validation
(visual and textual), and AI-generated reporting. Validated through a pilot
study in Mallorca, AIJIM achieved significant improvements in reporting speed
and accuracy, while maintaining transparency and ethical oversight through
Explainable AI (XAI), GDPR compliance, and community review. The model
demonstrates high transferability and offers a new benchmark for scalable,
responsible, and participatory journalism at the intersection of environmental
communication and artificial intelligence.
| [
{
"version": "v1",
"created": "Wed, 19 Mar 2025 19:00:24 GMT"
},
{
"version": "v2",
"created": "Sun, 30 Mar 2025 11:33:03 GMT"
},
{
"version": "v3",
"created": "Tue, 8 Apr 2025 06:26:24 GMT"
},
{
"version": "v4",
"created": "Mon, 21 Apr 2025 12:45:07 GMT"
}
] | 2025-04-22T00:00:00 | [
[
"Tiltack",
"Torsten",
""
]
] |
2503.17515 | Italo Romani de Oliveira | \'Italo Romani de Oliveira, Samet Ayhan, Glaucia Balvedi, Michael
Biglin, Pablo Costas, Euclides C. Pinto Neto, Alexandre Leite, Felipe C. F.
de Azevedo | A Predictive Services Architecture for Efficient Airspace Operations | null | null | null | null | cs.LG cs.AI cs.DB cs.SY eess.SY | http://arxiv.org/licenses/nonexclusive-distrib/1.0/ | Predicting air traffic congestion and flow management is essential for
airlines and Air Navigation Service Providers (ANSP) to enhance operational
efficiency. Accurate estimates of future airport capacity and airspace density
are vital for better airspace management, reducing air traffic controller
workload and fuel consumption, ultimately promoting sustainable aviation. While
existing literature has addressed these challenges, data management and query
processing remain complex due to the vast volume of high-rate air traffic data.
Many analytics use cases require a common pre-processing infrastructure, as
ad-hoc approaches are insufficient. Additionally, linear prediction models
often fall short, necessitating more advanced techniques.
This paper presents a data processing and predictive services architecture
that ingests large, uncorrelated, and noisy streaming data to forecast future
airspace system states. The system continuously collects raw data, periodically
compresses it, and stores it in NoSQL databases for efficient query processing.
For prediction, the system learns from historical traffic by extracting key
features such as airport arrival and departure events, sector boundary
crossings, weather parameters, and other air traffic data. These features are
input into various regression models, including linear, non-linear, and
ensemble models, with the best-performing model selected for predictions. We
evaluate this infrastructure across three prediction use cases in the US
National Airspace System (NAS) and a segment of European airspace, using
extensive real operations data, confirming that our system can predict future
system states efficiently and accurately.
| [
{
"version": "v1",
"created": "Fri, 21 Mar 2025 19:57:38 GMT"
},
{
"version": "v2",
"created": "Fri, 18 Apr 2025 19:15:45 GMT"
}
] | 2025-04-22T00:00:00 | [
[
"de Oliveira",
"Ítalo Romani",
""
],
[
"Ayhan",
"Samet",
""
],
[
"Balvedi",
"Glaucia",
""
],
[
"Biglin",
"Michael",
""
],
[
"Costas",
"Pablo",
""
],
[
"Neto",
"Euclides C. Pinto",
""
],
[
"Leite",
"Alexandre",
""
],
[
"de Azevedo",
"Felipe C. F.",
""
]
] |
2503.17549 | Andrzej Rostworowski | Andrzej Rostworowski | Vacuum axisymmetric gravitational collapse revisited: preliminary
investigation | 9 pages, 2 figures, a sign error pointeded out by a Referee in Eqs.
(57,58) corrected, Eq. (56) corrected, signs in the comments below eqs.
(51,56) corrected, numerous typos corrected | null | null | null | gr-qc | http://creativecommons.org/licenses/by/4.0/ | Validating the results of [A.M. Abrahams and C.R. Evans, Phys. Rev. Lett. 70,
2980] poses a numerical challenge and has been inspiring a lot of research. We
join these efforts and present our first steps to achieve this goal: we discuss
a formulation of Einstein equations for a vacuum axisymmetric spacetime with
vanishing twist in spherical-polar coordinates, its linearised approximation,
and identify some problems in achieving numerically stable evolution at the
threshold of a black hole formation.
| [
{
"version": "v1",
"created": "Fri, 21 Mar 2025 21:49:26 GMT"
},
{
"version": "v2",
"created": "Fri, 18 Apr 2025 19:10:30 GMT"
}
] | 2025-04-22T00:00:00 | [
[
"Rostworowski",
"Andrzej",
""
]
] |
2503.17859 | Mishel Khirk | M. Khirk (1 and 2), L.V. Grigorenko (1, 3 and 4), E.Yu. Nikolskii (4
and 1), P.G. Sharov (1 and 5) ((1) Flerov Laboratory of Nuclear Reactions,
JINR, Dubna, (2) M.V. Lomonosov Moscow State University, Skobeltsyn Institute
of Nuclear Physics, Moscow, (3) National Research Nuclear University
"MEPhI'', Moscow, (4) National Research Centre "Kurchatov Institute'',
Moscow, (5) Institute of Physics, Silesian University in Opava, Opava) | Correlation studies of the $^{7}$He excited states | 19 pages, 23 figures | null | null | null | nucl-ex nucl-th | http://creativecommons.org/licenses/by/4.0/ | The unbound nucleus $^{7}$He was recently studied in the
$^2$H($^{6}$He,$^1$H)$^{7}$He reaction at 29 $A\,$MeV beam energy in Ref.[M.S.
Golovkov et al., Phys. Rev. C 109, L061602 (2024)]. The excitation spectrum of
$^{7}$He was measured up to $E_T=8$ MeV ($E_T$ is energy above the $^{6}$He-$n$
threshold). Angular distribution for the $^{6}$He-$n$ decay of the $^{7}$He
$3/2^-$ ground state can be explained by a strong spin alignment induced by a
reaction mechanism. The correlation information for the higher-lying $^{7}$He
excitations is available as backward-forward asymmetry for the $^{6}$He-$n$
decay in the $^{7}$He frame. The asymmetry function has an expressed energy
profile which may be explained by using quite restrictive assumptions about
structure of $^{7}$He excitations or/and peculiarities of the reaction
mechanism. In the analysis of [M.S. Golovkov et al., Phys. Rev. C 109, L061602
(2024)] the observation the $s_{1/2}$ state in $^{7}$He is declared with $E_r
\approx 2.0$ MeV. Our work is based on the same $^{7}$He data. However, the
data analysis was improved and also the data interpretation is substantiated
with the detailed PWBA reaction studies and coupled-channel calculations of the
$^{7}$He continuous spectrum. The idea of the $s_{1/2}$ resonant state with
$E_r \approx 2.0$ is rejected. In addition, the position of the $1/2^-$ state
in $^{7}$He is confined to the interval $E_r=2.2-3.0$ MeV, with ``preferred''
value 2.6 MeV. There is indication on the second $3/2^-$ state in the data with
$E_r \sim 4.5$ MeV and with the lower resonance energy limit $E_r \gtrsim 3.5$
MeV. Importance and prospects of more detailed correlation studies of $^{7}$He
continuum are discussed.
| [
{
"version": "v1",
"created": "Sat, 22 Mar 2025 20:35:29 GMT"
},
{
"version": "v2",
"created": "Fri, 18 Apr 2025 20:55:08 GMT"
}
] | 2025-04-22T00:00:00 | [
[
"Khirk",
"M.",
"",
"1 and 2"
],
[
"Grigorenko",
"L. V.",
"",
"1, 3 and 4"
],
[
"Nikolskii",
"E. Yu.",
"",
"4\n and 1"
],
[
"Sharov",
"P. G.",
"",
"1 and 5"
]
] |
2503.17987 | Chenyu Zhang | Chenyu Zhang, Lanjun Wang, Yiwen Ma, Wenhui Li, and An-An Liu | Reason2Attack: Jailbreaking Text-to-Image Models via LLM Reasoning | This paper includes model-generated content that may contain
offensive or distressing material | null | null | null | cs.CR cs.AI cs.CV | http://arxiv.org/licenses/nonexclusive-distrib/1.0/ | Text-to-Image(T2I) models typically deploy safety filters to prevent the
generation of sensitive images. Unfortunately, recent jailbreaking attack
methods manually design prompts for the LLM to generate adversarial prompts,
which effectively bypass safety filters while producing sensitive images,
exposing safety vulnerabilities of T2I models. However, due to the LLM's
limited understanding of the T2I model and its safety filters, existing methods
require numerous queries to achieve a successful attack, limiting their
practical applicability. To address this issue, we propose Reason2Attack(R2A),
which aims to enhance the LLM's reasoning capabilities in generating
adversarial prompts by incorporating the jailbreaking attack into the
post-training process of the LLM. Specifically, we first propose a CoT example
synthesis pipeline based on Frame Semantics, which generates adversarial
prompts by identifying related terms and corresponding context illustrations.
Using CoT examples generated by the pipeline, we fine-tune the LLM to
understand the reasoning path and format the output structure. Subsequently, we
incorporate the jailbreaking attack task into the reinforcement learning
process of the LLM and design an attack process reward that considers prompt
length, prompt stealthiness, and prompt effectiveness, aiming to further
enhance reasoning accuracy. Extensive experiments on various T2I models show
that R2A achieves a better attack success ratio while requiring fewer queries
than baselines. Moreover, our adversarial prompts demonstrate strong attack
transferability across both open-source and commercial T2I models.
| [
{
"version": "v1",
"created": "Sun, 23 Mar 2025 08:40:39 GMT"
},
{
"version": "v2",
"created": "Sat, 19 Apr 2025 07:17:52 GMT"
}
] | 2025-04-22T00:00:00 | [
[
"Zhang",
"Chenyu",
""
],
[
"Wang",
"Lanjun",
""
],
[
"Ma",
"Yiwen",
""
],
[
"Li",
"Wenhui",
""
],
[
"Liu",
"An-An",
""
]
] |
2503.18977 | Xudong Huai | Xudong Huai, Luke Pritchard Cairns, Bridget Delles, Michal J.
Winiarski, Maurice Sorolla II, Xinshu Zhang, Youzhe Chen, Stuart Calder,
Tatenda Kanyowa, Anshul Kogar, Huibo Cao, Danielle Yahne, Robert Birgeneau,
James Analytis, Thao T. Tran | Jeff = 1/2 Diamond Magnet CaCo2TeO6: A Pathway toward New Spin Physics
and Quantum Functions | null | null | null | null | cond-mat.mtrl-sci physics.chem-ph physics.comp-ph quant-ph | http://creativecommons.org/licenses/by/4.0/ | Diamond lattice magnets, formed by a framework of corner-sharing tetrahedra
of magnetic cations, offer unique opportunities to realize novel states of
matter for potential utility in information technology. However, research has
mostly focused on AB2X4 spinels with Td magnetic ions. This hinders the
atomically enabled tunability of competing interactions at different energy
scales and the ability to harness many-body electronic states in quantum
materials, making the discovery of quantum fluctuations and spin dynamics less
accessible. We discover a new material CaCo2TeO6 featuring a diamond lattice of
two distinct Oh-Co2+ sites. This material displays strong quantum fluctuations,
increased competing magnetic exchange interactions, and field-induced
tunability of magnetic structures. The results demonstrate how simple,
fundamental refinements in ligand fields can profoundly influence the phase
space of quantum matter.
| [
{
"version": "v1",
"created": "Sun, 23 Mar 2025 01:57:51 GMT"
},
{
"version": "v2",
"created": "Mon, 21 Apr 2025 01:27:00 GMT"
}
] | 2025-04-22T00:00:00 | [
[
"Huai",
"Xudong",
""
],
[
"Cairns",
"Luke Pritchard",
""
],
[
"Delles",
"Bridget",
""
],
[
"Winiarski",
"Michal J.",
""
],
[
"Sorolla",
"Maurice",
"II"
],
[
"Zhang",
"Xinshu",
""
],
[
"Chen",
"Youzhe",
""
],
[
"Calder",
"Stuart",
""
],
[
"Kanyowa",
"Tatenda",
""
],
[
"Kogar",
"Anshul",
""
],
[
"Cao",
"Huibo",
""
],
[
"Yahne",
"Danielle",
""
],
[
"Birgeneau",
"Robert",
""
],
[
"Analytis",
"James",
""
],
[
"Tran",
"Thao T.",
""
]
] |
2503.19082 | Felix Burt | Felix Burt, Kuan-Cheng Chen, Kin K. Leung | A Multilevel Framework for Partitioning Quantum Circuits | 42 pages, 30 figures | null | null | null | quant-ph | http://creativecommons.org/licenses/by/4.0/ | Executing quantum algorithms over distributed quantum systems requires
quantum circuits to be divided into sub-circuits which communicate via
entanglement-based teleportation. Naively mapping circuits to qubits over
multiple quantum processing units (QPUs) results in large communication
overhead, increasing both execution time and noise. This can be minimised by
optimising the assignment of qubits to QPUs and the methods used for covering
non-local operations. Formulations that are general enough to capture the
spectrum of teleportation possibilities lead to complex problem instances which
can be difficult to solve effectively. This highlights a need to exploit the
wide range of heuristic techniques used in the graph partitioning literature.
This paper formalises and extends existing constructions for graphical quantum
circuit partitioning and designs a new objective function that captures further
possibilities for non-local operations via nested state teleportation. We adapt
the well-known Fiduccia-Mattheyses heuristic to the constraints and problem
objective and explore multilevel techniques that coarsen hypergraphs and
partition at multiple levels of granularity. We find that this reduces runtime
and improves solution quality of standard partitioning. We place these
techniques within a larger framework, through which we can extract full
distributed quantum circuits including teleportation instructions. We compare
the entanglement requirements and runtimes with state-of-the-art methods,
finding that we can achieve the lowest entanglement costs in most cases, while
always being close to the best performing method. We achieve an average
improvement of 33% over the next best performing method across a wide range of
circuits. We also find that our techniques can scale to much larger circuit
sizes than state-of-the-art methods, provided the number of partitions is not
too large.
| [
{
"version": "v1",
"created": "Mon, 24 Mar 2025 19:13:55 GMT"
},
{
"version": "v2",
"created": "Thu, 27 Mar 2025 14:11:06 GMT"
},
{
"version": "v3",
"created": "Fri, 18 Apr 2025 18:14:00 GMT"
}
] | 2025-04-22T00:00:00 | [
[
"Burt",
"Felix",
""
],
[
"Chen",
"Kuan-Cheng",
""
],
[
"Leung",
"Kin K.",
""
]
] |
2503.19294 | Jingtao Zhang | Jingtao Zhang and Xi Chen | Multilevel Monte Carlo Metamodeling for Variance Function Estimation | null | null | null | null | stat.ME | http://arxiv.org/licenses/nonexclusive-distrib/1.0/ | This work introduces a novel multilevel Monte Carlo (MLMC) metamodeling
approach for variance function estimation. Although devising an efficient
experimental design for simulation metamodeling can be elusive, the MLMC-based
approach addresses this challenge by dynamically adjusting the number of design
points and budget allocation at each level, thereby automatically creating an
efficient design. Theoretical analyses show that, under mild conditions, the
proposed MLMC metamodeling approach for variance function estimation can
achieve superior computational efficiency compared to standard Monte Carlo
metamodeling while achieving the desired level of accuracy. Additionally, this
work establishes the asymptotic normality of the MLMC metamodeling estimator
under certain sufficient conditions, providing valuable insights for
uncertainty quantification. Finally, two MLMC metamodeling procedures are
proposed for variance function estimation: one to achieve a target accuracy
level and another to efficiently utilize a fixed computational budget.
Numerical evaluations support the theoretical results and demonstrate the
potential of the proposed approach in facilitating global sensitivity analysis.
| [
{
"version": "v1",
"created": "Tue, 25 Mar 2025 02:50:25 GMT"
},
{
"version": "v2",
"created": "Sat, 19 Apr 2025 00:37:18 GMT"
}
] | 2025-04-22T00:00:00 | [
[
"Zhang",
"Jingtao",
""
],
[
"Chen",
"Xi",
""
]
] |
2503.19343 | Victor Vassiliev | V.A.Vassiliev | Equilevel algebras | null | null | null | null | math.AG math.AT | http://arxiv.org/licenses/nonexclusive-distrib/1.0/ | An equilevel algebra is a subalgebra of the space of smooth functions $f: M
\to {\mathbb R}$ distinguished in this space by finitely many conditions of the
type $f(x_i) = f(\tilde x_i)$, $x_i \neq \tilde x_i \in M$, or approximated by
such subalgebras. For $M = S^1$ or ${\mathbb R}^1$, the regular points of the
variety of equilevel algebras of codimension $k$ are known in knot theory as
$k$-chord diagrams. The whole of this variety completes the space of chord
diagrams in the same way as the Hilbert schemes complete the configuration
spaces. We describe cell structures of the varieties of all equilevel algebras
up to the codimension three in the space $C^\infty(S^1, {\mathbb R})$ and
compute their homology groups and characteristic classes of canonical vector
bundles on them.
| [
{
"version": "v1",
"created": "Tue, 25 Mar 2025 04:35:48 GMT"
},
{
"version": "v2",
"created": "Sat, 12 Apr 2025 09:43:02 GMT"
},
{
"version": "v3",
"created": "Sun, 20 Apr 2025 03:07:21 GMT"
}
] | 2025-04-22T00:00:00 | [
[
"Vassiliev",
"V. A.",
""
]
] |
2503.19371 | Daniel Saragih | Daniel Saragih, Deyu Cao, Tejas Balaji, Ashwin Santhosh | Flow to Learn: Flow Matching on Neural Network Parameters | Accepted at the ICLR Workshop on Neural Network Weights as a New Data
Modality 2025 | null | null | null | cs.LG cs.AI | http://creativecommons.org/licenses/by/4.0/ | Foundational language models show a remarkable ability to learn new concepts
during inference via context data. However, similar work for images lag behind.
To address this challenge, we introduce FLoWN, a flow matching model that
learns to generate neural network parameters for different tasks. Our approach
models the flow on latent space, while conditioning the process on context
data. Experiments verify that FLoWN attains various desiderata for a
meta-learning model. In addition, it matches or exceeds baselines on
in-distribution tasks, provides better initializations for classifier training,
and is performant on out-of-distribution few-shot tasks while having a
fine-tuning mechanism to improve performance.
| [
{
"version": "v1",
"created": "Tue, 25 Mar 2025 05:57:50 GMT"
},
{
"version": "v2",
"created": "Sat, 19 Apr 2025 13:42:41 GMT"
}
] | 2025-04-22T00:00:00 | [
[
"Saragih",
"Daniel",
""
],
[
"Cao",
"Deyu",
""
],
[
"Balaji",
"Tejas",
""
],
[
"Santhosh",
"Ashwin",
""
]
] |
2503.19902 | Fernando Julio Cendra | Fernando Julio Cendra, Kai Han | ICE: Intrinsic Concept Extraction from a Single Image via Diffusion
Models | CVPR 2025, Project page: https://visual-ai.github.io/ice | null | null | null | cs.CV | http://creativecommons.org/licenses/by-nc-sa/4.0/ | The inherent ambiguity in defining visual concepts poses significant
challenges for modern generative models, such as the diffusion-based
Text-to-Image (T2I) models, in accurately learning concepts from a single
image. Existing methods lack a systematic way to reliably extract the
interpretable underlying intrinsic concepts. To address this challenge, we
present ICE, short for Intrinsic Concept Extraction, a novel framework that
exclusively utilises a T2I model to automatically and systematically extract
intrinsic concepts from a single image. ICE consists of two pivotal stages. In
the first stage, ICE devises an automatic concept localization module to
pinpoint relevant text-based concepts and their corresponding masks within the
image. This critical stage streamlines concept initialization and provides
precise guidance for subsequent analysis. The second stage delves deeper into
each identified mask, decomposing the object-level concepts into intrinsic
concepts and general concepts. This decomposition allows for a more granular
and interpretable breakdown of visual elements. Our framework demonstrates
superior performance on intrinsic concept extraction from a single image in an
unsupervised manner. Project page: https://visual-ai.github.io/ice
| [
{
"version": "v1",
"created": "Tue, 25 Mar 2025 17:58:29 GMT"
},
{
"version": "v2",
"created": "Mon, 21 Apr 2025 17:59:26 GMT"
}
] | 2025-04-22T00:00:00 | [
[
"Cendra",
"Fernando Julio",
""
],
[
"Han",
"Kai",
""
]
] |
2503.19924 | Athokpam Langlen Chanu | Athokpam Langlen Chanu, Youngjai Park, Younghwa Cha, UnCheol Lee,
Joon-Young Moon, and Jong-Min Park | EEG relative phase-based analysis unveils the complexity and
universality of human brain dynamics: integrative insights from general
anesthesia and ADHD | null | null | null | null | q-bio.NC nlin.AO | http://creativecommons.org/licenses/by/4.0/ | Understanding brain wave patterns is fundamental to uncovering neural
information processing mechanisms, making quantifying complexity across brain
states an important line of investigation. We present a comprehensive analysis
of the complexity of electroencephalography (EEG) signals, integrating data
from seven distinct states experienced by participants undergoing general
anesthesia, and resting-state recordings from individuals with inattentive-type
ADHD alongside healthy control groups. Departing from prior studies that
primarily focus on EEG amplitude dynamics, we adopt a novel relative phase
approach to extract patterns of information flow directionality based on EEG
phase dynamics. We quantify the complexity of these relative phase
directionality patterns across various states using permutation entropy (PE)
and statistical complexity measure within ordinal pattern framework. Our
analysis: (i) PE is inversely correlated with the level of consciousness during
general anesthesia, reflecting a dynamic interplay between anesthetic depth and
shifts in directional information flow; (ii) healthy subjects consistently show
higher PE than inADHD participants; (iii) when mapped onto the
complexity-entropy causality plane, all brain states, regardless of condition
or individual differences, align along a single curve, suggesting an underlying
universal pattern in brain dynamics; and (iv) brain data consistently exhibit
higher complexity than standard stochastic processes, likely due to greater
multifractal scaling. These findings highlight that neural information
propagation, as captured by EEG relative phase dynamics, is governed by
self-organizing principles that are fundamentally more complex than the
stochastic processes. Our EEG relative phase-based characterization provides
new insight into the complexity of neural information flow directionality.
| [
{
"version": "v1",
"created": "Wed, 12 Mar 2025 15:00:51 GMT"
},
{
"version": "v2",
"created": "Thu, 27 Mar 2025 05:50:21 GMT"
},
{
"version": "v3",
"created": "Mon, 21 Apr 2025 15:15:14 GMT"
}
] | 2025-04-22T00:00:00 | [
[
"Chanu",
"Athokpam Langlen",
""
],
[
"Park",
"Youngjai",
""
],
[
"Cha",
"Younghwa",
""
],
[
"Lee",
"UnCheol",
""
],
[
"Moon",
"Joon-Young",
""
],
[
"Park",
"Jong-Min",
""
]
] |
2503.20010 | Seungki Kim | Seokho Jin, Seungki Kim | A truncated inner product formula in the geometry of numbers | Theorem 1.3 is now improved and it now holds without smoothing | null | null | null | math.NT | http://creativecommons.org/licenses/by/4.0/ | We provide an asymptotic inner product formula for truncations of the
pseudo-Eisenstein series on $\SL(n,\Z)\backslash\SL(n,\R)$ of the form
\begin{align*} E_{f}(g) = \sum_{L \subseteq \Z^n} f(\det Lg) \end{align*} for
$f \in C^\infty_0(\R_{\geq 0})$, where the sum ranges over all primitive rank
$k < n$ sublattices of $\Z^n$. We consider both the Arthur and the ``harsh''
truncations, with uses in the geometry of numbers in mind. As an application,
we present an improvement over an estimate concerning the number of the
primitive sublattices of a fixed lattice of bounded determinant, or
equivalently, the number of the rational points on a Grassmannian variety of
bounded height.
| [
{
"version": "v1",
"created": "Tue, 25 Mar 2025 18:49:13 GMT"
},
{
"version": "v2",
"created": "Sat, 19 Apr 2025 06:14:42 GMT"
}
] | 2025-04-22T00:00:00 | [
[
"Jin",
"Seokho",
""
],
[
"Kim",
"Seungki",
""
]
] |
2503.20090 | Michael Suleymanov | Michael Suleymanov, Avishy Carmi, Eliahu Cohen | Relativity of Quantum Correlations: Invariant Quantities and
Frame-Dependent Measures | 21 pages, 5 figures. The updated version includes some new results
and a few minor changes | null | null | null | quant-ph | http://arxiv.org/licenses/nonexclusive-distrib/1.0/ | Viewing frames of reference as physical systems, subject to the same laws as
the systems they describe, is central to the relational approach in physics.
Under the assumption that quantum mechanics universally governs all physical
entities, this perspective naturally leads to the concept of quantum reference
frames (QRFs). We investigate the perspective-dependence of position and
momentum uncertainties, correlations, covariance matrices, and entanglement
within the QRF formalism. We show that the Robertson-Schr\"odinger uncertainty
relations are frame-dependent, and so are correlations and variances, which
satisfy various constraints described as inequalities. However, the determinant
of the total covariance matrix, linked to the uncertainty volume in phase
space, as well as variance-based entanglement criteria, remains invariant under
changes of reference frame. Under specific conditions, the purities of
subsystems are also invariant for different QRFs, but in general, they are
perspective-dependent. These invariants suggest fundamental, robust measures of
uncertainty and entanglement that persist despite changes in observational
perspective, potentially inspiring dedicated quantum information protocols as
well as further foundational studies.
| [
{
"version": "v1",
"created": "Tue, 25 Mar 2025 22:02:41 GMT"
},
{
"version": "v2",
"created": "Sat, 19 Apr 2025 22:32:19 GMT"
}
] | 2025-04-22T00:00:00 | [
[
"Suleymanov",
"Michael",
""
],
[
"Carmi",
"Avishy",
""
],
[
"Cohen",
"Eliahu",
""
]
] |
2503.20314 | Shiwei Zhang | Team Wan, Ang Wang, Baole Ai, Bin Wen, Chaojie Mao, Chen-Wei Xie, Di
Chen, Feiwu Yu, Haiming Zhao, Jianxiao Yang, Jianyuan Zeng, Jiayu Wang,
Jingfeng Zhang, Jingren Zhou, Jinkai Wang, Jixuan Chen, Kai Zhu, Kang Zhao,
Keyu Yan, Lianghua Huang, Mengyang Feng, Ningyi Zhang, Pandeng Li, Pingyu Wu,
Ruihang Chu, Ruili Feng, Shiwei Zhang, Siyang Sun, Tao Fang, Tianxing Wang,
Tianyi Gui, Tingyu Weng, Tong Shen, Wei Lin, Wei Wang, Wei Wang, Wenmeng
Zhou, Wente Wang, Wenting Shen, Wenyuan Yu, Xianzhong Shi, Xiaoming Huang,
Xin Xu, Yan Kou, Yangyu Lv, Yifei Li, Yijing Liu, Yiming Wang, Yingya Zhang,
Yitong Huang, Yong Li, You Wu, Yu Liu, Yulin Pan, Yun Zheng, Yuntao Hong,
Yupeng Shi, Yutong Feng, Zeyinzi Jiang, Zhen Han, Zhi-Fan Wu, Ziyu Liu | Wan: Open and Advanced Large-Scale Video Generative Models | 60 pages, 33 figures | null | null | null | cs.CV | http://arxiv.org/licenses/nonexclusive-distrib/1.0/ | This report presents Wan, a comprehensive and open suite of video foundation
models designed to push the boundaries of video generation. Built upon the
mainstream diffusion transformer paradigm, Wan achieves significant
advancements in generative capabilities through a series of innovations,
including our novel VAE, scalable pre-training strategies, large-scale data
curation, and automated evaluation metrics. These contributions collectively
enhance the model's performance and versatility. Specifically, Wan is
characterized by four key features: Leading Performance: The 14B model of Wan,
trained on a vast dataset comprising billions of images and videos,
demonstrates the scaling laws of video generation with respect to both data and
model size. It consistently outperforms the existing open-source models as well
as state-of-the-art commercial solutions across multiple internal and external
benchmarks, demonstrating a clear and significant performance superiority.
Comprehensiveness: Wan offers two capable models, i.e., 1.3B and 14B
parameters, for efficiency and effectiveness respectively. It also covers
multiple downstream applications, including image-to-video, instruction-guided
video editing, and personal video generation, encompassing up to eight tasks.
Consumer-Grade Efficiency: The 1.3B model demonstrates exceptional resource
efficiency, requiring only 8.19 GB VRAM, making it compatible with a wide range
of consumer-grade GPUs. Openness: We open-source the entire series of Wan,
including source code and all models, with the goal of fostering the growth of
the video generation community. This openness seeks to significantly expand the
creative possibilities of video production in the industry and provide academia
with high-quality video foundation models. All the code and models are
available at https://github.com/Wan-Video/Wan2.1.
| [
{
"version": "v1",
"created": "Wed, 26 Mar 2025 08:25:43 GMT"
},
{
"version": "v2",
"created": "Sat, 19 Apr 2025 02:22:42 GMT"
}
] | 2025-04-22T00:00:00 | [
[
"Wan",
"Team",
""
],
[
"Wang",
"Ang",
""
],
[
"Ai",
"Baole",
""
],
[
"Wen",
"Bin",
""
],
[
"Mao",
"Chaojie",
""
],
[
"Xie",
"Chen-Wei",
""
],
[
"Chen",
"Di",
""
],
[
"Yu",
"Feiwu",
""
],
[
"Zhao",
"Haiming",
""
],
[
"Yang",
"Jianxiao",
""
],
[
"Zeng",
"Jianyuan",
""
],
[
"Wang",
"Jiayu",
""
],
[
"Zhang",
"Jingfeng",
""
],
[
"Zhou",
"Jingren",
""
],
[
"Wang",
"Jinkai",
""
],
[
"Chen",
"Jixuan",
""
],
[
"Zhu",
"Kai",
""
],
[
"Zhao",
"Kang",
""
],
[
"Yan",
"Keyu",
""
],
[
"Huang",
"Lianghua",
""
],
[
"Feng",
"Mengyang",
""
],
[
"Zhang",
"Ningyi",
""
],
[
"Li",
"Pandeng",
""
],
[
"Wu",
"Pingyu",
""
],
[
"Chu",
"Ruihang",
""
],
[
"Feng",
"Ruili",
""
],
[
"Zhang",
"Shiwei",
""
],
[
"Sun",
"Siyang",
""
],
[
"Fang",
"Tao",
""
],
[
"Wang",
"Tianxing",
""
],
[
"Gui",
"Tianyi",
""
],
[
"Weng",
"Tingyu",
""
],
[
"Shen",
"Tong",
""
],
[
"Lin",
"Wei",
""
],
[
"Wang",
"Wei",
""
],
[
"Wang",
"Wei",
""
],
[
"Zhou",
"Wenmeng",
""
],
[
"Wang",
"Wente",
""
],
[
"Shen",
"Wenting",
""
],
[
"Yu",
"Wenyuan",
""
],
[
"Shi",
"Xianzhong",
""
],
[
"Huang",
"Xiaoming",
""
],
[
"Xu",
"Xin",
""
],
[
"Kou",
"Yan",
""
],
[
"Lv",
"Yangyu",
""
],
[
"Li",
"Yifei",
""
],
[
"Liu",
"Yijing",
""
],
[
"Wang",
"Yiming",
""
],
[
"Zhang",
"Yingya",
""
],
[
"Huang",
"Yitong",
""
],
[
"Li",
"Yong",
""
],
[
"Wu",
"You",
""
],
[
"Liu",
"Yu",
""
],
[
"Pan",
"Yulin",
""
],
[
"Zheng",
"Yun",
""
],
[
"Hong",
"Yuntao",
""
],
[
"Shi",
"Yupeng",
""
],
[
"Feng",
"Yutong",
""
],
[
"Jiang",
"Zeyinzi",
""
],
[
"Han",
"Zhen",
""
],
[
"Wu",
"Zhi-Fan",
""
],
[
"Liu",
"Ziyu",
""
]
] |
2503.20519 | Jinnan Chen | Jinnan Chen, Lingting Zhu, Zeyu Hu, Shengju Qian, Yugang Chen, Xin
Wang, Gim Hee Lee | MAR-3D: Progressive Masked Auto-regressor for High-Resolution 3D
Generation | CVPR 2025 Highlight: https://jinnan-chen.github.io/projects/MAR-3D/ | null | null | null | cs.CV | http://arxiv.org/licenses/nonexclusive-distrib/1.0/ | Recent advances in auto-regressive transformers have revolutionized
generative modeling across different domains, from language processing to
visual generation, demonstrating remarkable capabilities. However, applying
these advances to 3D generation presents three key challenges: the unordered
nature of 3D data conflicts with sequential next-token prediction paradigm,
conventional vector quantization approaches incur substantial compression loss
when applied to 3D meshes, and the lack of efficient scaling strategies for
higher resolution latent prediction. To address these challenges, we introduce
MAR-3D, which integrates a pyramid variational autoencoder with a cascaded
masked auto-regressive transformer (Cascaded MAR) for progressive latent
upscaling in the continuous space. Our architecture employs random masking
during training and auto-regressive denoising in random order during inference,
naturally accommodating the unordered property of 3D latent tokens.
Additionally, we propose a cascaded training strategy with condition
augmentation that enables efficiently up-scale the latent token resolution with
fast convergence. Extensive experiments demonstrate that MAR-3D not only
achieves superior performance and generalization capabilities compared to
existing methods but also exhibits enhanced scaling capabilities compared to
joint distribution modeling approaches (e.g., diffusion transformers).
| [
{
"version": "v1",
"created": "Wed, 26 Mar 2025 13:00:51 GMT"
},
{
"version": "v2",
"created": "Thu, 27 Mar 2025 12:39:55 GMT"
},
{
"version": "v3",
"created": "Sun, 20 Apr 2025 08:36:32 GMT"
}
] | 2025-04-22T00:00:00 | [
[
"Chen",
"Jinnan",
""
],
[
"Zhu",
"Lingting",
""
],
[
"Hu",
"Zeyu",
""
],
[
"Qian",
"Shengju",
""
],
[
"Chen",
"Yugang",
""
],
[
"Wang",
"Xin",
""
],
[
"Lee",
"Gim Hee",
""
]
] |
2503.20574 | Gong-Bo Zhao | Weibing Zhang, Ruiyang Zhao, Xiaoyong Mu, Kazuya Koyama, Ryuichi
Takahashi, Yuting Wang, Gong-Bo Zhao | A cosmological full-shape power spectra analysis using pre- and
post-reconstructed density fields | 11 + 4 pages; 3 + 3 figures; version accepted to RAA | null | null | null | astro-ph.CO | http://creativecommons.org/licenses/by/4.0/ | In this work, we investigate a joint fitting approach based on theoretical
models of power spectra associated with density-field reconstruction.
Specifically, we consider the matter auto-power spectra before and after baryon
acoustic oscillation (BAO) reconstruction, as well as the cross-power spectrum
between the pre- and post-reconstructed density fields. We present
redshift-space models for these three power spectra at the one-loop level
within the framework of standard perturbation theory (SPT), and perform a joint
analysis using three types of power spectra, and quantify their impact on
parameter constraints. When restricting the analysis to wavenumbers $k \leq
0.2\,h\,\mathrm{Mpc}^{-1}$ and adopting a smoothing scale of $R_{\mathrm{s}} =
15\,h^{-1}\,\mathrm{Mpc}$, we find that incorporating all three power spectra
improves parameter constraints by approximately $11\%\text{--}16\%$ compared to
using only the post-reconstruction power spectrum, with the Figure of Merit
(FoM) increasing by $10.5\%$. These results highlight the advantages of
leveraging multiple power spectra in BAO reconstruction, ultimately enabling
more precise cosmological parameter estimation.
| [
{
"version": "v1",
"created": "Wed, 26 Mar 2025 14:23:30 GMT"
},
{
"version": "v2",
"created": "Mon, 21 Apr 2025 06:55:36 GMT"
}
] | 2025-04-22T00:00:00 | [
[
"Zhang",
"Weibing",
""
],
[
"Zhao",
"Ruiyang",
""
],
[
"Mu",
"Xiaoyong",
""
],
[
"Koyama",
"Kazuya",
""
],
[
"Takahashi",
"Ryuichi",
""
],
[
"Wang",
"Yuting",
""
],
[
"Zhao",
"Gong-Bo",
""
]
] |
2503.20727 | Kartik Prasanna | Atsushi Ichino, Kartik Prasanna | Representations of $\mathrm{GL}_2$ over $\mathbb{Z}/p^n\mathbb{Z}$ and
congruences for binomial coefficients | 16 pages, 1 figure Introduction, acknowledgements and references
updated, minor typos corrected | null | null | null | math.RT math.NT | http://arxiv.org/licenses/nonexclusive-distrib/1.0/ | For an odd prime $p$, we realize the trivial representation of
$\mathrm{GL}_2(\mathbb{Z}/p^n\mathbb{Z})$ on the free $\mathbb{Z}/p^n
\mathbb{Z}$-module of rank one as a subquotient of a direct sum of symmetric
power representations (twisted by appropriate powers of the determinant) of
rank strictly greater than one. The proof eventually reduces to establishing
some novel supercongruences for binomial coefficients.
| [
{
"version": "v1",
"created": "Wed, 26 Mar 2025 17:09:28 GMT"
},
{
"version": "v2",
"created": "Mon, 21 Apr 2025 14:07:07 GMT"
}
] | 2025-04-22T00:00:00 | [
[
"Ichino",
"Atsushi",
""
],
[
"Prasanna",
"Kartik",
""
]
] |
2503.20749 | Yuxuan Lu | Yuxuan Lu, Jing Huang, Yan Han, Bennet Bei, Yaochen Xie, Dakuo Wang,
Jessie Wang, Qi He | LLM Agents That Act Like Us: Accurate Human Behavior Simulation with
Real-World Data | null | null | null | null | cs.CL | http://arxiv.org/licenses/nonexclusive-distrib/1.0/ | Recent research shows that LLMs can simulate ``believable'' human behaviors
to power LLM agents via prompt-only methods. In this work, we focus on
evaluating and improving LLM's objective ``accuracy'' rather than the
subjective ``believability'' in the web action generation task, leveraging a
large-scale, real-world dataset collected from online shopping human actions.
We present the first comprehensive quantitative evaluation of state-of-the-art
LLMs (e.g., DeepSeek-R1, Llama, and Claude) on the task of web action
generation. Our results show that fine-tuning LLMs on real-world behavioral
data substantially improves their ability to generate actions compared to
prompt-only methods. Furthermore, incorporating synthesized reasoning traces
into model training leads to additional performance gains, demonstrating the
value of explicit rationale in behavior modeling. This work establishes a new
benchmark for evaluating LLMs in behavior simulation and offers actionable
insights into how real-world action data and reasoning augmentation can enhance
the fidelity of LLM agents.
| [
{
"version": "v1",
"created": "Wed, 26 Mar 2025 17:33:27 GMT"
},
{
"version": "v2",
"created": "Thu, 27 Mar 2025 02:42:03 GMT"
},
{
"version": "v3",
"created": "Sat, 5 Apr 2025 02:45:14 GMT"
},
{
"version": "v4",
"created": "Mon, 21 Apr 2025 05:12:56 GMT"
}
] | 2025-04-22T00:00:00 | [
[
"Lu",
"Yuxuan",
""
],
[
"Huang",
"Jing",
""
],
[
"Han",
"Yan",
""
],
[
"Bei",
"Bennet",
""
],
[
"Xie",
"Yaochen",
""
],
[
"Wang",
"Dakuo",
""
],
[
"Wang",
"Jessie",
""
],
[
"He",
"Qi",
""
]
] |
2503.21015 | Mark Neyrinck | Mark Neyrinck, Miguel Arag\'on-Calvo, and Istv\'an Szapudi | Galaxy and Halo Root Systems: Fingerprints of Mass Assembly | Several visualizations, e.g.
https://neyrinck.github.io/haloroots2d.mp4 . Comments welcome! | null | null | null | astro-ph.CO astro-ph.GA | http://creativecommons.org/licenses/by/4.0/ | We discuss what we call halo or galaxy root systems, collections of particle
pathlines that show the infall of matter from the initial uniform distribution
into a collapsed structure. The matter clumps as it falls in; projected through
time, it produces filamentary density enhancements analogous to tree roots and
branches, blood vessels, or even human transportation infrastructure in cities
and regions. This relates to the larger-scale cosmic web, but is defined
locally about one of its nodes: a physical, geometric version of a merger tree.
We find dark-matter-halo root systems on average to exhibit more roots and root
branches for the largest cluster haloes than in small haloes. This may relate
to the `cosmic-web detachment' mechanism that likely contributes to
star-formation quenching in galaxy groups and clusters. We also find that high
spin manifests in these root systems as curvier roots.
| [
{
"version": "v1",
"created": "Wed, 26 Mar 2025 22:03:57 GMT"
},
{
"version": "v2",
"created": "Fri, 18 Apr 2025 19:54:40 GMT"
}
] | 2025-04-22T00:00:00 | [
[
"Neyrinck",
"Mark",
""
],
[
"Aragón-Calvo",
"Miguel",
""
],
[
"Szapudi",
"István",
""
]
] |
2503.21074 | Ooha Lakkadi Reddy | Ooha Lakkadi Reddy | Rerouting Connection: Hybrid Computer Vision Analysis Reveals Visual
Similarity Between Indus and Tibetan-Yi Corridor Writing Systems | 107 pages (43 main text, 6 references, 58 appendices). 21 figures, 4
tables in main text; 106 figures, 8 tables total. Code available at
https://github.com/oohalakkadi/ivc2tyc. Undergraduate thesis at Duke Kunshan
University. Accepted for presentation at the 52nd International Conference
for Computer Applications & Quantitative Methods in Archaeology (CAA 2025),
Athens, Greece | null | null | null | cs.CV cs.AI cs.CL cs.LG | http://creativecommons.org/licenses/by/4.0/ | This thesis employs a hybrid CNN-Transformer architecture, alongside a
detailed anthropological framework, to investigate potential historical
connections between the visual morphology of the Indus Valley script and
pictographic systems of the Tibetan-Yi Corridor. Through an ensemble
methodology of three target scripts across 15 independently trained models, we
demonstrate that Tibetan-Yi Corridor scripts exhibit approximately six-fold
higher visual similarity to the Indus script (0.635) than to the Bronze Age
Proto-Cuneiform (0.102) or Proto-Elamite (0.078).
Contrary to expectations, when measured through direct script-to-script
embedding comparisons, the Indus script maps closer to Tibetan-Yi Corridor
scripts with a mean cosine similarity of 0.930 (CI: [0.917, 0.942]) than to
contemporaneous West Asian signaries, which recorded mean similarities of 0.887
(CI: [0.863, 0.911]) and 0.855 (CI: [0.818, 0.891]). Across dimensionality
reduction and clustering methods, the Indus script consistently clusters
closest to Tibetan-Yi Corridor scripts.
These computational findings align with observed pictorial parallels in
numeral systems, gender markers, and iconographic elements. Archaeological
evidence of contact networks along the ancient Shu-Shendu road, coinciding with
the Indus Civilization's decline, provides a plausible transmission pathway.
While alternate explanations cannot be ruled out, the specificity and
consistency of similarities suggest more complex cultural transmission networks
between South and East Asia than previously recognized.
| [
{
"version": "v1",
"created": "Thu, 27 Mar 2025 01:19:47 GMT"
},
{
"version": "v2",
"created": "Tue, 1 Apr 2025 01:56:42 GMT"
},
{
"version": "v3",
"created": "Sat, 19 Apr 2025 09:18:11 GMT"
}
] | 2025-04-22T00:00:00 | [
[
"Reddy",
"Ooha Lakkadi",
""
]
] |
2503.21085 | Pengtao Song | Shengyong Li, Yidian Fan, Xiang Li, Xinhui Ruan, Qianchuan Zhao,
Zhihui Peng, Re-Bing Wu, Jing Zhang and Pengtao Song | Robust Quantum Control using Reinforcement Learning from Demonstration | 11 pages, 5 figures;typos corrected, references added | null | null | null | quant-ph | http://creativecommons.org/licenses/by/4.0/ | Quantum control requires high-precision and robust control pulses to ensure
optimal system performance. However, control sequences generated with a system
model may suffer from model bias, leading to low fidelity. While model-free
reinforcement learning (RL) methods have been developed to avoid such biases,
training an RL agent from scratch can be time-consuming, often taking hours to
gather enough samples for convergence. This challenge has hindered the broad
application of RL techniques to larger and more complex quantum control issues,
limiting their adaptability. In this work, we use Reinforcement Learning from
Demonstration (RLfD) to leverage the control sequences generated with system
models and further optimize them with RL to avoid model bias. By avoiding
learning from scratch and starting with reasonable control pulse shapes, this
approach can increase sample efficiency by reducing the number of samples,
which can significantly reduce the training time. Thus, this method can
effectively handle pulse shapes that are discretized into more than 1000 pieces
without compromising final fidelity. We have simulated the preparation of
several high-fidelity non-classical states using the RLfD method. We also find
that the training process is more stable when using RLfD. In addition, this
method is suitable for fast gate calibration using reinforcement learning.
| [
{
"version": "v1",
"created": "Thu, 27 Mar 2025 02:01:28 GMT"
},
{
"version": "v2",
"created": "Sun, 20 Apr 2025 03:39:39 GMT"
}
] | 2025-04-22T00:00:00 | [
[
"Li",
"Shengyong",
""
],
[
"Fan",
"Yidian",
""
],
[
"Li",
"Xiang",
""
],
[
"Ruan",
"Xinhui",
""
],
[
"Zhao",
"Qianchuan",
""
],
[
"Peng",
"Zhihui",
""
],
[
"Wu",
"Re-Bing",
""
],
[
"Zhang",
"Jing",
""
],
[
"Song",
"Pengtao",
""
]
] |
2503.21088 | Ningyu Zhang | Haoming Xu, Shuxun Wang, Yanqiu Zhao, Yi Zhong, Ziyan Jiang, Ningyuan
Zhao, Shumin Deng, Huajun Chen, Ningyu Zhang | ZJUKLAB at SemEval-2025 Task 4: Unlearning via Model Merging | SemEval@ACL 2025 | null | null | null | cs.CL cs.AI cs.CV cs.LG cs.MM | http://arxiv.org/licenses/nonexclusive-distrib/1.0/ | This paper presents the ZJUKLAB team's submission for SemEval-2025 Task 4:
Unlearning Sensitive Content from Large Language Models. This task aims to
selectively erase sensitive knowledge from large language models, avoiding both
over-forgetting and under-forgetting issues. We propose an unlearning system
that leverages Model Merging (specifically TIES-Merging), combining two
specialized models into a more balanced unlearned model. Our system achieves
competitive results, ranking second among 26 teams, with an online score of
0.944 for Task Aggregate and 0.487 for overall Aggregate. In this paper, we
also conduct local experiments and perform a comprehensive analysis of the
unlearning process, examining performance trajectories, loss dynamics, and
weight perspectives, along with several supplementary experiments, to
understand the effectiveness of our method. Furthermore, we analyze the
shortcomings of our method and evaluation metrics, emphasizing that MIA scores
and ROUGE-based metrics alone are insufficient to fully evaluate successful
unlearning. Finally, we emphasize the need for more comprehensive evaluation
methodologies and rethinking of unlearning objectives in future research. Code
is available at https://github.com/zjunlp/unlearn/tree/main/semeval25.
| [
{
"version": "v1",
"created": "Thu, 27 Mar 2025 02:03:25 GMT"
},
{
"version": "v2",
"created": "Sun, 20 Apr 2025 00:37:24 GMT"
}
] | 2025-04-22T00:00:00 | [
[
"Xu",
"Haoming",
""
],
[
"Wang",
"Shuxun",
""
],
[
"Zhao",
"Yanqiu",
""
],
[
"Zhong",
"Yi",
""
],
[
"Jiang",
"Ziyan",
""
],
[
"Zhao",
"Ningyuan",
""
],
[
"Deng",
"Shumin",
""
],
[
"Chen",
"Huajun",
""
],
[
"Zhang",
"Ningyu",
""
]
] |
2503.21138 | Hedong Yan | Hedong Yan | A Computational Theory for Efficient Model Evaluation with Causal
Guarantees | null | null | null | null | cs.AI cs.LG math.ST stat.ML stat.TH | http://arxiv.org/licenses/nonexclusive-distrib/1.0/ | In order to reduce the cost of experimental evaluation for models, we
introduce a computational theory of evaluation for prediction and decision
models: build evaluation model to accelerate the evaluation procedures. We
prove upper bounds of generalized error and generalized causal effect error of
given evaluation models. We also prove efficiency, and consistency to estimated
causal effect from deployed subject to evaluation metric by prediction. To
learn evaluation models, we propose a meta-learner to handle heterogeneous
evaluation subjects space problem. Comparing with existed evaluation
approaches, our (conditional) evaluation model reduced 24.1\%-99.0\% evaluation
errors across 12 scenes, including individual medicine, scientific simulation,
social experiment, business activity, and quantum trade. The evaluation time is
reduced 3-7 order of magnitude comparing with experiments or simulations.
| [
{
"version": "v1",
"created": "Thu, 27 Mar 2025 04:00:49 GMT"
},
{
"version": "v2",
"created": "Wed, 16 Apr 2025 07:26:19 GMT"
},
{
"version": "v3",
"created": "Sat, 19 Apr 2025 04:06:47 GMT"
}
] | 2025-04-22T00:00:00 | [
[
"Yan",
"Hedong",
""
]
] |
2503.21533 | Malay K. Nandy | Nihar Ranjan Ghosh and Malay K. Nandy | Geodesically Complete Regularized Schwarzschild Black Holes | null | null | null | null | gr-qc | http://creativecommons.org/licenses/by/4.0/ | Classical general relativity predicts a singularity at the center of a black
hole, where known laws of physics break down. This suggests the existence of
deeper, yet unknown principles of Nature. Among various theoretical
possibilities, one of the most promising proposals is a transition to a de
Sitter phase at the BH core. This transition, originally proposed by Gliner and
Sakharov, ensures the regularity of metric coefficients and avoids the
singularity. In search for such a regular BH solution with finite curvature
scalar, we propose a metric $g_{rr}$ that exhibits a dS core in the central
region. An appealing feature of this metric is the existence of a $single$
event horizon resembling the Schwarzschild black hole. Furthermore, the entire
spacetime geometry is determined by the black hole mass alone, in agreement
with the Isarel-Carter $no-hair$ $theorem$ for a charge-less, non-rotating
black hole. To determine the gravitational action consistent with such a
solution, we consider a general Lagrangian density $f(R)$ in place of the
Einstein-Hilbert action. By numerically solving the resulting field equation,
we find that, in addition to the Einstein-Hilbert term, a Pad\'e approximant in
the Ricci scalar $R$ can produce such regular black hole solutions. To assess
the physical viability of these black hole solutions, we verify that the
proposed metric satisfies the principal energy conditions: DEC, WEC, and NEC,
throughout spacetime. Furthermore, in agreement with Zaslavskii's regularity
criterion, the metric satisfies the SEC in the range $r\geq r_h/2$, where $r_h$
is the event horizon. Furthermore, with the proposed regularized metric, the
expansion scalar in the Raychaudhuri equation remains finite and its derivative
vanishes at $r=0$, thereby preventing formation of caustic. This confirms that
the spacetime is geodesically complete and free from true physical
singularities.
| [
{
"version": "v1",
"created": "Thu, 27 Mar 2025 14:24:58 GMT"
},
{
"version": "v2",
"created": "Sun, 20 Apr 2025 08:27:49 GMT"
}
] | 2025-04-22T00:00:00 | [
[
"Ghosh",
"Nihar Ranjan",
""
],
[
"Nandy",
"Malay K.",
""
]
] |
2503.22080 | Matthias von Davier | Matthias von Davier | An Improved Satterthwaite Effective Degrees of Freedom Correction for
Weighted Syntheses of Variance | null | null | null | null | stat.ME math.ST stat.TH | http://creativecommons.org/licenses/by-nc-nd/4.0/ | This article presents an improved approximation for the effective degrees of
freedom in the Satterthwaite (1941, 1946) method which estimates the
distribution of a weighted combination of variance components The standard
Satterthwaite approximation assumes a scaled chisquare distribution for the
composite variance estimator but is known to be biased downward when component
degrees of freedom are small. Building on recent work by von Davier (2025), we
propose an adjusted estimator that corrects this bias by modifying both the
numerator and denominator of the traditional formula. The new approximation
incorporates a weighted average of component degrees of freedom and a scaling
factor that ensures consistency as the number of components or their degrees of
freedom increases. We demonstrate the utility of this adjustment in practical
settings, including Rubin's (1987) total variance estimation in multiple
imputations, where weighted variance combinations are common. The proposed
estimator generalizes and further improves von Davier's (2025) unweighted case
and more accurately approximates synthetic variance estimators with arbitrary
weights.
| [
{
"version": "v1",
"created": "Fri, 28 Mar 2025 01:57:15 GMT"
},
{
"version": "v2",
"created": "Sun, 20 Apr 2025 11:16:34 GMT"
}
] | 2025-04-22T00:00:00 | [
[
"von Davier",
"Matthias",
""
]
] |
2503.22314 | Helge Maakestad Dr. | Helge {\O}ystein Maakestad | A universal characteristic class for vector bundles with a connection | null | null | null | null | math.AG math.AC math.AT math.DG | http://creativecommons.org/licenses/by-nc-nd/4.0/ | In the paper I introduce a new characteristic class $c(E)$ for a finite rank
vector bundle $E$ on an affine scheme $S:=Spec(A)$ - the fundamental class of
$E$. The class $c(E)$ is not a characteristic class in the classical sense in
the sense that it lives in a pointed cohomology torsor $\operatorname{Ext}^1(L,
\operatorname{End}_A(E))$. Most characteristic classes lives in a cohomology
group. The pointed cohomology torsor $\operatorname{Ext}^1(L,
\operatorname{End}_A(E))$ is a torsor on the abelian group
$\operatorname{H}^2(L, Z(\operatorname{End}_A(E)))$ where
$Z(\operatorname{End}_A(E))$ is the center of the ring of endomorphisms of $E$
and where the cohomology is the Lie-Rinehart cohomology of the center. The
class $c(E)$ is trivial if and only if $E$ has a flat algebraic connection.
Hence the class $c(T_S)$ where $T_S$ is the tangent bundle, is an an
obstruction for $S$ to be algebraically parallelizable. I use a connection
$\nabla$ to define $c(E)$ and I also prove the class $c(E)$ is independent of
choice of connection, hence $c(E)$ is an invariant of the vector bundle $E$.
The class generalize the Chern class, the Pontryagin class, the Euler class and
the Teleman characteristic class. I prove using an explicit example that the
class $c(E)$ is stronger than the Chern class and the Euler class. I also give
a new proof of a formula for the curvature of a connection $\nabla$ in terms of
an idempotent endomorphism $\phi$ defining $E$. This formula was claimed and
proved in a paper put out on the arXiv in 2011, and in this paper I give a new
proof that is easier to read. The class may be interesting in the study of the
"cancellation problem" in affine algebraic geometry and the problem of giving
algebraic formulas for the topological Euler characteristic. I also calculate
the algebraic deRham cohomology of the complex two sphere and prove it is
infinite dimensional.
| [
{
"version": "v1",
"created": "Fri, 28 Mar 2025 10:43:38 GMT"
},
{
"version": "v2",
"created": "Sun, 20 Apr 2025 10:21:22 GMT"
}
] | 2025-04-22T00:00:00 | [
[
"Maakestad",
"Helge Øystein",
""
]
] |
2503.22695 | Pengfei Ma | Pengfei Ma and Li Cai and Xuan Wang and Hao Gao | Fully GPU-Accelerated Immersed Boundary Method for Fluid-Structure
Interaction in Complex Cardiac Models | null | null | null | null | physics.comp-ph | http://arxiv.org/licenses/nonexclusive-distrib/1.0/ | Fluid-structure interaction (FSI) plays a crucial role in cardiac mechanics,
where the strong coupling between fluid flow and deformable structures presents
significant computational challenges. The immersed boundary (IB) method
efficiently handles large deformations and contact without requiring mesh
regeneration. However, solving complex FSI problems demands high computational
efficiency, making GPU acceleration essential to leverage massive parallelism,
high throughput, and memory bandwidth. We present a fully GPU-accelerated
algorithm for the IB method to solve FSI problems in complex cardiac models.
The Navier-Stokes equations are discretized using the finite difference method,
while the finite element method is employed for structural mechanics.
Traditionally, IB methods are not GPU-friendly due to irregular memory access
and limited parallelism. The novelty of this work lies in eliminating sparse
matrix storage and operations entirely, significantly improving memory access
efficiency and fully utilizing GPU computational capability. Additionally, the
structural materials can be modeled using general hyperelastic constitutive
laws, including fiber-reinforced anisotropic biological tissues such as the
Holzapfel-Ogden (HO) model. Furthermore, a combined geometric multigrid solver
is used to accelerate the convergence. The full FSI system, consisting of
millions of degrees of freedom, achieves a per-timestep computation time of
just 0.1 seconds. We conduct FSI simulations of the left ventricle, mitral
valve, and aortic valve, achieving results with high consistency. Compared to
single-core CPU computations, our fully GPU-accelerated approach delivers
speedups exceeding 100 times.
| [
{
"version": "v1",
"created": "Fri, 14 Mar 2025 09:32:30 GMT"
},
{
"version": "v2",
"created": "Mon, 21 Apr 2025 12:53:24 GMT"
}
] | 2025-04-22T00:00:00 | [
[
"Ma",
"Pengfei",
""
],
[
"Cai",
"Li",
""
],
[
"Wang",
"Xuan",
""
],
[
"Gao",
"Hao",
""
]
] |
2503.22792 | Xu Yang | Xu Yang, Ryan Buechele and Nandini Trivedi | Detection of anyon braiding through pump-probe spectroscopy | 17 pages, 11 figures. References updated; discussion in "Time Scales"
section expanded | null | null | null | cond-mat.str-el | http://arxiv.org/licenses/nonexclusive-distrib/1.0/ | We show that the braiding of anyons in a quantum spin liquid leaves a
distinct dynamical signature in the nonlinear pump-probe response. Using a
combination of exact diagonalization and matrix product state techniques, we
study the nonlinear pump-probe response of the toric code in a magnetic field,
a model which hosts mobile electric $e$ and magnetic $m$ anyonic excitations.
While the linear response signal oscillates and decays with time like $\sim
t^{-1.3}$, the amplitude of the nonlinear signal for $\chi^{(3)}_{XZZ}$
features a linear-in-time enhancement at early times. The comparison between
$\chi^{(3)}_{XZZ}$, which involves the non-trivial braiding of $e$ and $m$
anyons, and $\chi^{(3)}_{XXX}$ that involves the trivial braiding of the same
types of anyons, serves to distinguish the braiding statistics of anyons. We
support our analysis by constructing a hard-core anyon model with statistical
gauge fields to develop further insights into the time dependence of the
pump-probe response. Pump-probe spectroscopy provides a distinctive new probe
of quantum spin liquid states, beyond the inconclusive broad features observed
in single spin flip inelastic neutron scattering.
| [
{
"version": "v1",
"created": "Fri, 28 Mar 2025 18:00:03 GMT"
},
{
"version": "v2",
"created": "Mon, 21 Apr 2025 06:53:02 GMT"
}
] | 2025-04-22T00:00:00 | [
[
"Yang",
"Xu",
""
],
[
"Buechele",
"Ryan",
""
],
[
"Trivedi",
"Nandini",
""
]
] |
2503.22995 | Mamtaj Akter | Mamtaj Akter, Jinkyung Katie Park, Pamela J. Wisniewski | Moving Beyond Parental Control toward Community-based Approaches to
Adolescent Online Safety | null | Mobile Technology and Teens Workshop of the 2025 CHI Conference on
Human Factors in Computing Systems | null | null | cs.HC | http://creativecommons.org/licenses/by/4.0/ | In this position paper, we discuss the paradigm shift that moves away from
parental mediation approaches toward collaborative approaches to promote
adolescents' online safety. We present empirical studies that highlight the
limitations of traditional parental control models and advocate for
collaborative, community-driven solutions that prioritize teen empowerment.
Specifically, we explore how extending oversight beyond the immediate family to
include trusted community members can provide crucial support for teens in
managing their online lives. We discuss the potential benefits and challenges
of this expanded approach, emphasizing the importance of granular privacy
controls and reciprocal support within these networks. Finally, we pose open
questions for the research community to consider during the workshop, focusing
on the design of "teen-centered" online safety solutions that foster autonomy,
awareness, and self-regulation.
| [
{
"version": "v1",
"created": "Sat, 29 Mar 2025 07:00:27 GMT"
},
{
"version": "v2",
"created": "Sat, 5 Apr 2025 07:18:24 GMT"
},
{
"version": "v3",
"created": "Sat, 19 Apr 2025 04:19:52 GMT"
}
] | 2025-04-22T00:00:00 | [
[
"Akter",
"Mamtaj",
""
],
[
"Park",
"Jinkyung Katie",
""
],
[
"Wisniewski",
"Pamela J.",
""
]
] |
2503.23512 | Jianhui Wang | Qiang Yi, Yangfan He, Jianhui Wang, Xinyuan Song, Shiyao Qian, Xinhang
Yuan, Miao Zhang, Li Sun, Keqin Li, Kuan Lu, Menghao Huo, Jiaqi Chen, Tianyu
Shi | SCORE: Story Coherence and Retrieval Enhancement for AI Narratives | null | null | null | null | cs.CL | http://arxiv.org/licenses/nonexclusive-distrib/1.0/ | Large Language Models (LLMs) can generate creative and engaging narratives
from user-specified input, but maintaining coherence and emotional depth
throughout these AI-generated stories remains a challenge. In this work, we
propose SCORE, a framework for Story Coherence and Retrieval Enhancement,
designed to detect and resolve narrative inconsistencies. By tracking key item
statuses and generating episode summaries, SCORE uses a Retrieval-Augmented
Generation (RAG) approach, incorporating TF-IDF and cosine similarity to
identify related episodes and enhance the overall story structure. Results from
testing multiple LLM-generated stories demonstrate that SCORE significantly
improves the consistency and stability of narrative coherence compared to
baseline GPT models, providing a more robust method for evaluating and refining
AI-generated narratives.
| [
{
"version": "v1",
"created": "Sun, 30 Mar 2025 16:48:27 GMT"
},
{
"version": "v2",
"created": "Mon, 21 Apr 2025 05:40:00 GMT"
}
] | 2025-04-22T00:00:00 | [
[
"Yi",
"Qiang",
""
],
[
"He",
"Yangfan",
""
],
[
"Wang",
"Jianhui",
""
],
[
"Song",
"Xinyuan",
""
],
[
"Qian",
"Shiyao",
""
],
[
"Yuan",
"Xinhang",
""
],
[
"Zhang",
"Miao",
""
],
[
"Sun",
"Li",
""
],
[
"Li",
"Keqin",
""
],
[
"Lu",
"Kuan",
""
],
[
"Huo",
"Menghao",
""
],
[
"Chen",
"Jiaqi",
""
],
[
"Shi",
"Tianyu",
""
]
] |
2503.23611 | Yuhong Zhong | Yuhong Zhong, Daniel S. Berger, Pantea Zardoshti, Enrique Saurez,
Jacob Nelson, Antonis Psistakis, Joshua Fried, Asaf Cidon | My CXL Pool Obviates Your PCIe Switch | null | null | 10.1145/3713082.3730393 | null | cs.OS | http://arxiv.org/licenses/nonexclusive-distrib/1.0/ | Pooling PCIe devices across multiple hosts offers a promising solution to
mitigate stranded I/O resources, enhance device utilization, address device
failures, and reduce total cost of ownership. The only viable option today are
PCIe switches, which decouple PCIe devices from hosts by connecting them
through a hardware switch. However, the high cost and limited flexibility of
PCIe switches hinder their widespread adoption beyond specialized datacenter
use cases.
This paper argues that PCIe device pooling can be effectively implemented in
software using CXL memory pools. CXL memory pools improve memory utilization
and already have positive return on investment. We find that, once CXL pools
are in place, they can serve as a building block for pooling any kind of PCIe
device. We demonstrate that PCIe devices can directly use CXL memory as I/O
buffers without device modifications, which enables routing PCIe traffic
through CXL pool memory. This software-based approach is deployable on today's
hardware and is more flexible than hardware PCIe switches. In particular, we
explore how disaggregating devices such as NICs can transform datacenter
infrastructure.
| [
{
"version": "v1",
"created": "Sun, 30 Mar 2025 22:25:18 GMT"
},
{
"version": "v2",
"created": "Mon, 14 Apr 2025 14:23:00 GMT"
},
{
"version": "v3",
"created": "Mon, 21 Apr 2025 16:57:36 GMT"
}
] | 2025-04-22T00:00:00 | [
[
"Zhong",
"Yuhong",
""
],
[
"Berger",
"Daniel S.",
""
],
[
"Zardoshti",
"Pantea",
""
],
[
"Saurez",
"Enrique",
""
],
[
"Nelson",
"Jacob",
""
],
[
"Psistakis",
"Antonis",
""
],
[
"Fried",
"Joshua",
""
],
[
"Cidon",
"Asaf",
""
]
] |
2503.23765 | Yun Li | Yun Li, Yiming Zhang, Tao Lin, XiangRui Liu, Wenxiao Cai, Zheng Liu,
Bo Zhao | STI-Bench: Are MLLMs Ready for Precise Spatial-Temporal World
Understanding? | null | null | null | null | cs.CV | http://arxiv.org/licenses/nonexclusive-distrib/1.0/ | The use of Multimodal Large Language Models (MLLMs) as an end-to-end solution
for Embodied AI and Autonomous Driving has become a prevailing trend. While
MLLMs have been extensively studied for visual semantic understanding tasks,
their ability to perform precise and quantitative spatial-temporal
understanding in real-world applications remains largely unexamined, leading to
uncertain prospects. To evaluate models' Spatial-Temporal Intelligence, we
introduce STI-Bench, a benchmark designed to evaluate MLLMs' spatial-temporal
understanding through challenging tasks such as estimating and predicting the
appearance, pose, displacement, and motion of objects. Our benchmark
encompasses a wide range of robot and vehicle operations across desktop,
indoor, and outdoor scenarios. The extensive experiments reveals that the
state-of-the-art MLLMs still struggle in real-world spatial-temporal
understanding, especially in tasks requiring precise distance estimation and
motion analysis.
| [
{
"version": "v1",
"created": "Mon, 31 Mar 2025 06:30:35 GMT"
},
{
"version": "v2",
"created": "Wed, 9 Apr 2025 18:07:39 GMT"
},
{
"version": "v3",
"created": "Mon, 21 Apr 2025 13:43:53 GMT"
}
] | 2025-04-22T00:00:00 | [
[
"Li",
"Yun",
""
],
[
"Zhang",
"Yiming",
""
],
[
"Lin",
"Tao",
""
],
[
"Liu",
"XiangRui",
""
],
[
"Cai",
"Wenxiao",
""
],
[
"Liu",
"Zheng",
""
],
[
"Zhao",
"Bo",
""
]
] |
2503.23869 | Yongle Li | Yongle Li, Bo Liu, Sheng Huang, ZHeng ZHang, Xiaotong Yuan, and
Richang Hong | Communication-Efficient and Personalized Federated Foundation Model
Fine-Tuning via Tri-Matrix Adaptation | null | null | null | null | cs.LG | http://arxiv.org/licenses/nonexclusive-distrib/1.0/ | In federated learning, fine-tuning pre-trained foundation models poses
significant challenges, particularly regarding high communication cost and
suboptimal model performance due to data heterogeneity between the clients. To
address these issues, this paper introduces communication-efficient federated
LoRA adaption (CE-LoRA), a method that employs a tri-factorization low-rank
adaptation approach with personalized model parameter aggregation. We first
presents a novel LoRA parameter factorization by introducing a small-size dense
matrix, which can significantly reduce the communication cost and achieve
comparable empirical performance than transferring the low-rank parameter
matrix used by existing methods. Without violating data privacy, the server
considers the client similarity in both training dataset and model parameter
space, and learns personalized weights for model aggregation. Our experiments
on various LLM and VLM fine-tuning tasks demonstrate that CE-LoRA not only
significantly reduces communication overhead but also improves performance
under not independently and identically distributed data conditions. In
addition, CE-LoRA improves data privacy protection, effectively mitigating
gradient-based data reconstruction attacks.
| [
{
"version": "v1",
"created": "Mon, 31 Mar 2025 09:18:42 GMT"
},
{
"version": "v2",
"created": "Sun, 20 Apr 2025 01:10:23 GMT"
}
] | 2025-04-22T00:00:00 | [
[
"Li",
"Yongle",
""
],
[
"Liu",
"Bo",
""
],
[
"Huang",
"Sheng",
""
],
[
"ZHang",
"ZHeng",
""
],
[
"Yuan",
"Xiaotong",
""
],
[
"Hong",
"Richang",
""
]
] |
2503.24331 | Keshav Das Agarwal | Keshav Das Agarwal, Tanoy Kanti Konar, Leela Ganesh Chandra Lakkaraju,
Aditi Sen De | Critical quantum metrology using non-Hermitian spin model with
RT-symmetry | v1 : 7 pages, 3 figures; v2 : 9 pages, 4 figures, new results on
dynamics added, references updated | null | null | null | quant-ph cond-mat.str-el | http://arxiv.org/licenses/nonexclusive-distrib/1.0/ | The non-Hermitian transverse $XY$ model with
Kaplan-Shekhtman-Entin-Wohlman-Aharony (KSEA) interaction having
$\mathcal{RT}$-symmetry, referred to as $iKSEA$ model, possesses both an
exceptional point at which eigenvectors coalesce and a quantum critical point
where gap-closing occurs. To precisely estimate the magnetic field of the
system, we prove that the quantum Fisher information (QFI) of the ground state
of the $iKSEA$ model, which is a lower bound of the precision quantified by the
root mean square error, scales as $N^2$, with $N$ being the system-size. This
provides Heisenberg scaling both at the quantum critical point and the
exceptional point in the thermodynamic limit. It indicates that reservoir
engineering can provide enhanced precision of system parameters when the system
is in contact with the bath, resulting in this non-Hermitian model.
Additionally, we demonstrate analytically that, in contrast to Hermitian
systems, QFI surpasses the Heisenberg limit and achieves super-Heisenberg
scaling ($\sim N^6$), when the strength of the KSEA interaction approaches the
anisotropy parameter, permitting competition between non-hermiticity and
hermiticity features, as long as the system size is moderate. Moreover, we
illustrate that starting from a product state, the non-Hermitian evolving
Hamiltonian can create the dynamical state that surpasses the standard quantum
limit in the broken regime.
| [
{
"version": "v1",
"created": "Mon, 31 Mar 2025 17:19:31 GMT"
},
{
"version": "v2",
"created": "Sun, 20 Apr 2025 11:25:50 GMT"
}
] | 2025-04-22T00:00:00 | [
[
"Agarwal",
"Keshav Das",
""
],
[
"Konar",
"Tanoy Kanti",
""
],
[
"Lakkaraju",
"Leela Ganesh Chandra",
""
],
[
"De",
"Aditi Sen",
""
]
] |
2504.00142 | Srinitish Srinivasan | Srinitish Srinivasan and Omkumar CU | Lorentzian Graph Isomorphic Network | Preprint. Under Review | null | null | null | cs.LG cs.AI | http://creativecommons.org/licenses/by/4.0/ | We introduce the Lorentzian Graph Isomorphic Network (LGIN), a novel graph
neural network (GNN) designed to operate in hyperbolic spaces, leveraging the
Lorentzian model to enhance graph representation learning. Existing GNNs
primarily operate in Euclidean spaces, which can limit their ability to capture
hierarchical and multi-relational structures inherent to complex graphs. LGIN
addresses this by incorporating curvature-aware aggregation functions that
preserve the Lorentzian metric tensor, ensuring embeddings remain constrained
within the hyperbolic space by proposing a new update rule that effectively
captures both local neighborhood interactions and global structural properties,
enabling LGIN to distinguish non-isomorphic graphs with expressiveness at least
as powerful as the Weisfeiler-Lehman test. Through extensive evaluation across
nine benchmark datasets, including molecular and protein structures, LGIN
consistently outperforms or matches state-of-the-art GNNs, demonstrating its
robustness and efficacy in modeling complex graph structures. To the best of
our knowledge, this is the first study to extend the concept of a powerful
graph neural network to Riemannian manifolds, paving the way for future
advancements in hyperbolic graph learning. The code for our paper can be found
at https://github.com/Deceptrax123/LGIN.
| [
{
"version": "v1",
"created": "Mon, 31 Mar 2025 18:49:34 GMT"
},
{
"version": "v2",
"created": "Mon, 21 Apr 2025 07:52:27 GMT"
}
] | 2025-04-22T00:00:00 | [
[
"Srinivasan",
"Srinitish",
""
],
[
"CU",
"Omkumar",
""
]
] |
2504.00364 | Yi-Hsuan Chen | Yi-Hsuan Chen, Shuo Liu, Wei Xiao, Calin Belta, Michael Otte | Control Barrier Functions via Minkowski Operations for Safe Navigation
among Polytopic Sets | 8 pages, 3 figures | null | null | null | cs.RO math.OC | http://arxiv.org/licenses/nonexclusive-distrib/1.0/ | Safely navigating around obstacles while respecting the dynamics, control,
and geometry of the underlying system is a key challenge in robotics. Control
Barrier Functions (CBFs) generate safe control policies by considering system
dynamics and geometry when calculating safe forward-invariant sets. Existing
CBF-based methods often rely on conservative shape approximations, like spheres
or ellipsoids, which have explicit and differentiable distance functions. In
this paper, we propose an optimization-defined CBF that directly considers the
exact Signed Distance Function (SDF) between a polytopic robot and polytopic
obstacles. Inspired by the Gilbert-Johnson-Keerthi (GJK) algorithm, we
formulate both (i) minimum distance and (ii) penetration depth between
polytopic sets as convex optimization problems in the space of Minkowski
difference operations (the MD-space). Convenient geometric properties of the
MD-space enable the derivatives of implicit SDF between two polytopes to be
computed via differentiable optimization. We demonstrate the proposed framework
in three scenarios including pure translation, initialization inside an unsafe
set, and multi-obstacle avoidance. These three scenarios highlight the
generation of a non-conservative maneuver, a recovery after starting in
collision, and the consideration of multiple obstacles via pairwise CBF
constraint, respectively.
| [
{
"version": "v1",
"created": "Tue, 1 Apr 2025 02:29:11 GMT"
},
{
"version": "v2",
"created": "Sun, 20 Apr 2025 17:21:22 GMT"
}
] | 2025-04-22T00:00:00 | [
[
"Chen",
"Yi-Hsuan",
""
],
[
"Liu",
"Shuo",
""
],
[
"Xiao",
"Wei",
""
],
[
"Belta",
"Calin",
""
],
[
"Otte",
"Michael",
""
]
] |
2504.00387 | Weijia Li | Zilong Huang, Jun He, Junyan Ye, Lihan Jiang, Weijia Li, Yiping Chen,
Ting Han | Scene4U: Hierarchical Layered 3D Scene Reconstruction from Single
Panoramic Image for Your Immerse Exploration | CVPR 2025, 11 pages, 7 figures | null | null | null | cs.CV | http://arxiv.org/licenses/nonexclusive-distrib/1.0/ | The reconstruction of immersive and realistic 3D scenes holds significant
practical importance in various fields of computer vision and computer
graphics. Typically, immersive and realistic scenes should be free from
obstructions by dynamic objects, maintain global texture consistency, and allow
for unrestricted exploration. The current mainstream methods for image-driven
scene construction involves iteratively refining the initial image using a
moving virtual camera to generate the scene. However, previous methods struggle
with visual discontinuities due to global texture inconsistencies under varying
camera poses, and they frequently exhibit scene voids caused by
foreground-background occlusions. To this end, we propose a novel layered 3D
scene reconstruction framework from panoramic image, named Scene4U.
Specifically, Scene4U integrates an open-vocabulary segmentation model with a
large language model to decompose a real panorama into multiple layers. Then,
we employs a layered repair module based on diffusion model to restore occluded
regions using visual cues and depth information, generating a hierarchical
representation of the scene. The multi-layer panorama is then initialized as a
3D Gaussian Splatting representation, followed by layered optimization, which
ultimately produces an immersive 3D scene with semantic and structural
consistency that supports free exploration. Scene4U outperforms
state-of-the-art method, improving by 24.24% in LPIPS and 24.40% in BRISQUE,
while also achieving the fastest training speed. Additionally, to demonstrate
the robustness of Scene4U and allow users to experience immersive scenes from
various landmarks, we build WorldVista3D dataset for 3D scene reconstruction,
which contains panoramic images of globally renowned sites. The implementation
code and dataset will be released at https://github.com/LongHZ140516/Scene4U .
| [
{
"version": "v1",
"created": "Tue, 1 Apr 2025 03:17:24 GMT"
},
{
"version": "v2",
"created": "Mon, 21 Apr 2025 02:40:21 GMT"
}
] | 2025-04-22T00:00:00 | [
[
"Huang",
"Zilong",
""
],
[
"He",
"Jun",
""
],
[
"Ye",
"Junyan",
""
],
[
"Jiang",
"Lihan",
""
],
[
"Li",
"Weijia",
""
],
[
"Chen",
"Yiping",
""
],
[
"Han",
"Ting",
""
]
] |
2504.00529 | Chuangyin Dang | Yiyin Cao and Chuangyin Dang | A Characterization of Nash Equilibrium in Behavioral Strategies through
Local Sequential Rationality | null | null | null | null | econ.TH | http://arxiv.org/licenses/nonexclusive-distrib/1.0/ | The concept of Nash equilibrium in behavioral strategies (NashEBS) was
formulated By Nash~\cite{Nash (1951)} for an extensive-form game through global
rationality of nonconvex payoff functions. Kuhn's payoff equivalence theorem
resolves the nonconvexity issue, but it overlooks that one Nash equilibrium of
the associated normal-form game can correspond to infinitely many NashEBSs of
an extensive-form game. To remedy this multiplicity, the traditional approach
as documented in Myerson~\cite{Myerson (1991)} involves a two-step process:
identifying a Nash equilibrium of the agent normal-form representation,
followed by verifying whether the corresponding mixed strategy profile is a
Nash equilibrium of the associated normal-form game, which often scales
exponentially with the size of the extensive-form game tree. In response to
these challenges, this paper develops a characterization of NashEBS through the
incorporation of an extra behavioral strategy profile and beliefs, which meet
local sequential rationality of linear payoff functions and self-independent
consistency. This characterization allows one to analytically determine all
NashEBSs for small extensive-form games. Building upon this characterization,
we acquire a polynomial system serving as a necessary and sufficient condition
for determining whether a behavioral strategy profile is a NashEBS. An
application of the characterization yields differentiable path-following
methods for computing such an equilibrium.
| [
{
"version": "v1",
"created": "Tue, 1 Apr 2025 08:21:48 GMT"
},
{
"version": "v2",
"created": "Sun, 20 Apr 2025 02:16:14 GMT"
}
] | 2025-04-22T00:00:00 | [
[
"Cao",
"Yiyin",
""
],
[
"Dang",
"Chuangyin",
""
]
] |
2504.00695 | Zhouhong Gu | Xiaoxuan Zhu, Zhouhong Gu, Baiqian Wu, Suhang Zheng, Tao Wang, Tianyu
Li, Hongwei Feng, Yanghua Xiao | ToReMi: Topic-Aware Data Reweighting for Dynamic Pre-Training Data
Selection | null | null | null | null | cs.CL | http://creativecommons.org/licenses/by/4.0/ | Pre-training large language models (LLMs) necessitates enormous diverse
textual corpora, making effective data selection a key challenge for balancing
computational resources and model performance. Current methodologies primarily
emphasize data quality metrics and mixing proportions, yet they fail to
adequately capture the underlying semantic connections between training samples
and quality disparities within individual domains. We introduce ToReMi
(Topic-based Reweighting for Model improvement), a novel two-stage framework
that dynamically adjusts training sample weights according to their topical
associations and observed learning patterns. Our comprehensive experiments
reveal that ToReMi variants consistently achieve superior performance over
conventional pre-training approaches, demonstrating accelerated perplexity
reduction across multiple domains and enhanced capabilities on downstream
evaluation tasks. Code is available at https://github.com/zxx000728/ToReMi.
| [
{
"version": "v1",
"created": "Tue, 1 Apr 2025 12:06:42 GMT"
},
{
"version": "v2",
"created": "Mon, 7 Apr 2025 06:33:33 GMT"
},
{
"version": "v3",
"created": "Sun, 20 Apr 2025 04:18:23 GMT"
}
] | 2025-04-22T00:00:00 | [
[
"Zhu",
"Xiaoxuan",
""
],
[
"Gu",
"Zhouhong",
""
],
[
"Wu",
"Baiqian",
""
],
[
"Zheng",
"Suhang",
""
],
[
"Wang",
"Tao",
""
],
[
"Li",
"Tianyu",
""
],
[
"Feng",
"Hongwei",
""
],
[
"Xiao",
"Yanghua",
""
]
] |
2504.00730 | Peiqi Li | Jiayuan She, Lin Shi, Peiqi Li, Ziling Dong, Renxing Li, Shengkai Li,
Liping Gu, Zhao Tong, Zhuochang Yang, Yajie Ji, Liang Feng, Jiangang Chen | Detection of Disease on Nasal Breath Sound by New Lightweight
Architecture: Using COVID-19 as An Example | 14 pages, 5 figures, 6 tables | null | null | null | cs.LG | http://arxiv.org/licenses/nonexclusive-distrib/1.0/ | Background. Infectious diseases, particularly COVID-19, continue to be a
significant global health issue. Although many countries have reduced or
stopped large-scale testing measures, the detection of such diseases remains a
propriety. Objective. This study aims to develop a novel, lightweight deep
neural network for efficient, accurate, and cost-effective detection of
COVID-19 using a nasal breathing audio data collected via smartphones.
Methodology. Nasal breathing audio from 128 patients diagnosed with the Omicron
variant was collected. Mel-Frequency Cepstral Coefficients (MFCCs), a widely
used feature in speech and sound analysis, were employed for extracting
important characteristics from the audio signals. Additional feature selection
was performed using Random Forest (RF) and Principal Component Analysis (PCA)
for dimensionality reduction. A Dense-ReLU-Dropout model was trained with
K-fold cross-validation (K=3), and performance metrics like accuracy,
precision, recall, and F1-score were used to evaluate the model. Results. The
proposed model achieved 97% accuracy in detecting COVID-19 from nasal breathing
sounds, outperforming state-of-the-art methods such as those by [23] and [13].
Our Dense-ReLU-Dropout model, using RF and PCA for feature selection, achieves
high accuracy with greater computational efficiency compared to existing
methods that require more complex models or larger datasets. Conclusion. The
findings suggest that the proposed method holds significant potential for
clinical implementation, advancing smartphone-based diagnostics in infectious
diseases. The Dense-ReLU-Dropout model, combined with innovative feature
processing techniques, offers a promising approach for efficient and accurate
COVID-19 detection, showcasing the capabilities of mobile device-based
diagnostics
| [
{
"version": "v1",
"created": "Tue, 1 Apr 2025 12:41:53 GMT"
},
{
"version": "v2",
"created": "Sat, 19 Apr 2025 06:43:31 GMT"
}
] | 2025-04-22T00:00:00 | [
[
"She",
"Jiayuan",
""
],
[
"Shi",
"Lin",
""
],
[
"Li",
"Peiqi",
""
],
[
"Dong",
"Ziling",
""
],
[
"Li",
"Renxing",
""
],
[
"Li",
"Shengkai",
""
],
[
"Gu",
"Liping",
""
],
[
"Tong",
"Zhao",
""
],
[
"Yang",
"Zhuochang",
""
],
[
"Ji",
"Yajie",
""
],
[
"Feng",
"Liang",
""
],
[
"Chen",
"Jiangang",
""
]
] |
2504.00837 | Shuyu Li | Shuyu Li, Shulei Ji, Zihao Wang, Songruoyao Wu, Jiaxing Yu, Kejun
Zhang | A Survey on Music Generation from Single-Modal, Cross-Modal, and
Multi-Modal Perspectives | null | null | null | null | cs.SD cs.AI cs.MM | http://creativecommons.org/licenses/by-nc-sa/4.0/ | Multi-modal music generation, using multiple modalities like text, images,
and video alongside musical scores and audio as guidance, is an emerging
research area with broad applications. This paper reviews this field,
categorizing music generation systems from the perspective of modalities. The
review covers modality representation, multi-modal data alignment, and their
utilization to guide music generation. Current datasets and evaluation methods
are also discussed. Key challenges in this area include effective multi-modal
integration, large-scale comprehensive datasets, and systematic evaluation
methods. Finally, an outlook on future research directions is provided,
focusing on creativity, efficiency, multi-modal alignment, and evaluation.
| [
{
"version": "v1",
"created": "Tue, 1 Apr 2025 14:26:25 GMT"
},
{
"version": "v2",
"created": "Sun, 20 Apr 2025 12:55:44 GMT"
}
] | 2025-04-22T00:00:00 | [
[
"Li",
"Shuyu",
""
],
[
"Ji",
"Shulei",
""
],
[
"Wang",
"Zihao",
""
],
[
"Wu",
"Songruoyao",
""
],
[
"Yu",
"Jiaxing",
""
],
[
"Zhang",
"Kejun",
""
]
] |
2504.00904 | Yi-Tang Chen | Yi-Tang Chen, Haoyu Li, Neng Shi, Xihaier Luo, Wei Xu, Han-Wei Shen | Explorable INR: An Implicit Neural Representation for Ensemble
Simulation Enabling Efficient Spatial and Parameter Exploration | Accepted by IEEE Transactions on Visualization and Computer Graphics
(TVCG) | null | null | null | cs.GR cs.LG | http://creativecommons.org/licenses/by-nc-sa/4.0/ | With the growing computational power available for high-resolution ensemble
simulations in scientific fields such as cosmology and oceanology, storage and
computational demands present significant challenges. Current surrogate models
fall short in the flexibility of point- or region-based predictions as the
entire field reconstruction is required for each parameter setting, hence
hindering the efficiency of parameter space exploration. Limitations exist in
capturing physical attribute distributions and pinpointing optimal parameter
configurations. In this work, we propose Explorable INR, a novel implicit
neural representation-based surrogate model, designed to facilitate exploration
and allow point-based spatial queries without computing full-scale field data.
In addition, to further address computational bottlenecks of spatial
exploration, we utilize probabilistic affine forms (PAFs) for uncertainty
propagation through Explorable INR to obtain statistical summaries,
facilitating various ensemble analysis and visualization tasks that are
expensive with existing models. Furthermore, we reformulate the parameter
exploration problem as optimization tasks using gradient descent and KL
divergence minimization that ensures scalability. We demonstrate that the
Explorable INR with the proposed approach for spatial and parameter exploration
can significantly reduce computation and memory costs while providing effective
ensemble analysis.
| [
{
"version": "v1",
"created": "Tue, 1 Apr 2025 15:33:28 GMT"
},
{
"version": "v2",
"created": "Mon, 21 Apr 2025 17:27:05 GMT"
}
] | 2025-04-22T00:00:00 | [
[
"Chen",
"Yi-Tang",
""
],
[
"Li",
"Haoyu",
""
],
[
"Shi",
"Neng",
""
],
[
"Luo",
"Xihaier",
""
],
[
"Xu",
"Wei",
""
],
[
"Shen",
"Han-Wei",
""
]
] |
2504.00957 | Rachmad Vidya Wicaksana Putra | Rachmad Vidya Wicaksana Putra, Pasindu Wickramasinghe, Muhammad
Shafique | Enabling Efficient Processing of Spiking Neural Networks with On-Chip
Learning on Commodity Neuromorphic Processors for Edge AI Systems | Accepted at the International Joint Conference on Neural Networks
(IJCNN) 2025 in Rome, Italy | null | null | null | cs.NE cs.AI cs.AR cs.LG | http://arxiv.org/licenses/nonexclusive-distrib/1.0/ | The rising demand for energy-efficient edge AI systems (e.g., mobile
agents/robots) has increased the interest in neuromorphic computing, since it
offers ultra-low power/energy AI computation through spiking neural network
(SNN) algorithms on neuromorphic processors. However, their efficient
implementation strategy has not been comprehensively studied, hence limiting
SNN deployments for edge AI systems. Toward this, we propose a design
methodology to enable efficient SNN processing on commodity neuromorphic
processors. To do this, we first study the key characteristics of targeted
neuromorphic hardware (e.g., memory and compute budgets), and leverage this
information to perform compatibility analysis for network selection. Afterward,
we employ a mapping strategy for efficient SNN implementation on the targeted
processor. Furthermore, we incorporate an efficient on-chip learning mechanism
to update the systems' knowledge for adapting to new input classes and dynamic
environments. The experimental results show that the proposed methodology leads
the system to achieve low latency of inference (i.e., less than 50ms for image
classification, less than 200ms for real-time object detection in video
streaming, and less than 1ms in keyword recognition) and low latency of on-chip
learning (i.e., less than 2ms for keyword recognition), while incurring less
than 250mW of processing power and less than 15mJ of energy consumption across
the respective different applications and scenarios. These results show the
potential of the proposed methodology in enabling efficient edge AI systems for
diverse application use-cases.
| [
{
"version": "v1",
"created": "Tue, 1 Apr 2025 16:52:03 GMT"
},
{
"version": "v2",
"created": "Sat, 19 Apr 2025 05:17:37 GMT"
}
] | 2025-04-22T00:00:00 | [
[
"Putra",
"Rachmad Vidya Wicaksana",
""
],
[
"Wickramasinghe",
"Pasindu",
""
],
[
"Shafique",
"Muhammad",
""
]
] |
2504.00975 | Muhammad Umer | Muhammad Umer, Muhammad Ahmed Mohsin, Huma Ghafoor, and Syed Ali
Hassan | Resource Allocation for RIS-Assisted CoMP-NOMA Networks using
Reinforcement Learning | null | null | null | null | eess.SP cs.AI | http://arxiv.org/licenses/nonexclusive-distrib/1.0/ | This thesis delves into the forefront of wireless communication by exploring
the synergistic integration of three transformative technologies: STAR-RIS,
CoMP, and NOMA. Driven by the ever-increasing demand for higher data rates,
improved spectral efficiency, and expanded coverage in the evolving landscape
of 6G development, this research investigates the potential of these
technologies to revolutionize future wireless networks.
The thesis analyzes the performance gains achievable through strategic
deployment of STAR-RIS, focusing on mitigating inter-cell interference,
enhancing signal strength, and extending coverage to cell-edge users. Resource
sharing strategies for STAR-RIS elements are explored, optimizing both
transmission and reflection functionalities. Analytical frameworks are
developed to quantify the benefits of STAR-RIS assisted CoMP-NOMA networks
under realistic channel conditions, deriving key performance metrics such as
ergodic rates and outage probabilities. Additionally, the research delves into
energy-efficient design approaches for CoMP-NOMA networks incorporating RIS,
proposing novel RIS configurations and optimization algorithms to achieve a
balance between performance and energy consumption. Furthermore, the
application of Deep Reinforcement Learning (DRL) techniques for intelligent and
adaptive optimization in aerial RIS-assisted CoMP-NOMA networks is explored,
aiming to maximize network sum rate while meeting user quality of service
requirements. Through a comprehensive investigation of these technologies and
their synergistic potential, this thesis contributes valuable insights into the
future of wireless communication, paving the way for the development of more
efficient, reliable, and sustainable networks capable of meeting the demands of
our increasingly connected world.
| [
{
"version": "v1",
"created": "Tue, 1 Apr 2025 17:14:01 GMT"
},
{
"version": "v2",
"created": "Sat, 19 Apr 2025 12:03:00 GMT"
}
] | 2025-04-22T00:00:00 | [
[
"Umer",
"Muhammad",
""
],
[
"Mohsin",
"Muhammad Ahmed",
""
],
[
"Ghafoor",
"Huma",
""
],
[
"Hassan",
"Syed Ali",
""
]
] |
2504.01318 | Arun Kuchibhotla | Abhishek Chakrabortty and Arun K. Kuchibhotla | Tail Bounds for Canonical $U$-Statistics and $U$-Processes with
Unbounded Kernels | This is a slightly edited version of the 2018 draft available at
https://faculty.wharton.upenn.edu/wp-content/uploads/2018/10/Chakrabortty-UStat-Draft.pdf.
Added more comments on the assumptions and the proof technique of Theorem 1.
Corrected a few typos. More improvements to follow in the future for the
U-process results | null | null | null | math.ST math.PR stat.TH | http://creativecommons.org/licenses/by-nc-nd/4.0/ | In this paper, we prove exponential tail bounds for canonical (or degenerate)
$U$-statistics and $U$-processes under exponential-type tail assumptions on the
kernels. Most of the existing results in the relevant literature often assume
bounded kernels or obtain sub-optimal tail behavior under unbounded kernels. We
obtain sharp rates and optimal tail behavior under sub-Weibull kernel
functions. Some examples from nonparametric and semiparametric statistics
literature are considered.
| [
{
"version": "v1",
"created": "Wed, 2 Apr 2025 03:05:28 GMT"
},
{
"version": "v2",
"created": "Mon, 21 Apr 2025 03:25:44 GMT"
}
] | 2025-04-22T00:00:00 | [
[
"Chakrabortty",
"Abhishek",
""
],
[
"Kuchibhotla",
"Arun K.",
""
]
] |
2504.01337 | Mohan Zhang | Mohan Zhang, Pingzhi Li, Jie Peng, Mufan Qiu, Tianlong Chen | Advancing MoE Efficiency: A Collaboration-Constrained Routing (C2R)
Strategy for Better Expert Parallelism Design | NAACL 2025, SAC award for Low-resource Methods for NLP | null | null | null | cs.LG cs.AI cs.CL cs.DC | http://creativecommons.org/licenses/by/4.0/ | Mixture-of-Experts (MoE) has successfully scaled up models while maintaining
nearly constant computing costs. By employing a gating network to route input
tokens, it selectively activates a subset of expert networks to process the
corresponding token embeddings. However, in practice, the efficiency of MoE is
challenging to achieve due to two key reasons: imbalanced expert activation,
which leads to substantial idle time during model or expert parallelism, and
insufficient capacity utilization; massive communication overhead, induced by
numerous expert routing combinations in expert parallelism at the system level.
Previous works typically formulate it as the load imbalance issue characterized
by the gating network favoring certain experts over others or attribute it to
static execution which fails to adapt to the dynamic expert workload at
runtime. In this paper, we exploit it from a brand new perspective, a
higher-order view and analysis of MoE routing policies: expert collaboration
and specialization where some experts tend to activate broadly with others
(collaborative), while others are more likely to activate only with a specific
subset of experts (specialized). Our experiments reveal that most experts tend
to be overly collaborative, leading to increased communication overhead from
repeatedly sending tokens to different accelerators. To this end, we propose a
novel collaboration-constrained routing (C2R) strategy to encourage more
specialized expert groups, as well as to improve expert utilization, and
present an efficient implementation of MoE that further leverages expert
specialization. We achieve an average performance improvement of 0.51% and
0.33% on LLaMA-MoE and Qwen-MoE respectively across ten downstream NLP
benchmarks, and reduce the all2all communication costs between GPUs, bringing
an extra 20%-30% total running time savings on top of the existing SoTA, i.e.
MegaBlocks.
| [
{
"version": "v1",
"created": "Wed, 2 Apr 2025 03:51:59 GMT"
},
{
"version": "v2",
"created": "Sun, 20 Apr 2025 15:13:33 GMT"
}
] | 2025-04-22T00:00:00 | [
[
"Zhang",
"Mohan",
""
],
[
"Li",
"Pingzhi",
""
],
[
"Peng",
"Jie",
""
],
[
"Qiu",
"Mufan",
""
],
[
"Chen",
"Tianlong",
""
]
] |
2504.01338 | Manolo Canales | Manolo Canales Cuba and Vin\'icius do Carmo Mel\'icio and Jo\~ao Paulo
Gois | FlowMotion: Target-Predictive Conditional Flow Matching for
Jitter-Reduced Text-Driven Human Motion Generation | null | null | null | null | cs.GR cs.LG | http://arxiv.org/licenses/nonexclusive-distrib/1.0/ | Achieving high-fidelity and temporally smooth 3D human motion generation
remains a challenge, particularly within resource-constrained environments. We
introduce FlowMotion, a novel method leveraging Conditional Flow Matching
(CFM). FlowMotion incorporates a training objective within CFM that focuses on
more accurately predicting target motion in 3D human motion generation,
resulting in enhanced generation fidelity and temporal smoothness while
maintaining the fast synthesis times characteristic of flow-matching-based
methods. FlowMotion achieves state-of-the-art jitter performance, achieving the
best jitter in the KIT dataset and the second-best jitter in the HumanML3D
dataset, and a competitive FID value in both datasets. This combination
provides robust and natural motion sequences, offering a promising equilibrium
between generation quality and temporal naturalness.
| [
{
"version": "v1",
"created": "Wed, 2 Apr 2025 03:55:21 GMT"
},
{
"version": "v2",
"created": "Sun, 20 Apr 2025 00:50:16 GMT"
}
] | 2025-04-22T00:00:00 | [
[
"Cuba",
"Manolo Canales",
""
],
[
"Melício",
"Vinícius do Carmo",
""
],
[
"Gois",
"João Paulo",
""
]
] |
2504.01365 | Kamil Mudru\v{n}ka | Kamil Mudrunka, Kazunori Nakayama | Inflation with Gauss-Bonnet correction: beyond slow-roll | 31 pages, 14 figures, added missing references and chapter 6 | null | null | TU-1256, KEK-QUP-2025-0005 | gr-qc astro-ph.CO | http://creativecommons.org/licenses/by/4.0/ | If a coupling between the inflaton and the Gauss-Bonnet term is introduced,
many models of inflation that were ruled out by the most recent Planck data can
be made viable again. The predictions for the scalar spectral index and
tensor-to-scalar ratio are typically computed using the slow-roll
approximation. In this paper we instead study the full equations of motion and
determine the necessary initial conditions for reasonable inflation epoch. We
derive the conditions under which the Friedmann equation admits positive
solutions for the Hubble parameter. Then we study the possibility of the
inflaton becoming trapped in a local potential minimum induced by the
Gauss-Bonnet term. Finally we demonstrate the results on monomial potential
models with a quadratic and a quartic potential and show that the slow-roll
approximation becomes imprecise in the quartic case.
| [
{
"version": "v1",
"created": "Wed, 2 Apr 2025 05:18:15 GMT"
},
{
"version": "v2",
"created": "Sun, 20 Apr 2025 11:47:35 GMT"
}
] | 2025-04-22T00:00:00 | [
[
"Mudrunka",
"Kamil",
""
],
[
"Nakayama",
"Kazunori",
""
]
] |
2504.01701 | Gabriele Rembado | Jean Dou\c{c}ot, Gabriele Rembado, Daisuke Yamakawa | Twisted local G-wild mapping class groups | v2: added a section and references, comments welcome! | null | null | null | math.GT math.AG nlin.SI | http://arxiv.org/licenses/nonexclusive-distrib/1.0/ | We consider the (universal) local isomonodromic deformations of
irregular-singular connections defined on principal bundles over complex
curves: for any (connected) complex reductive structure group G and any pole
order, allowing for twisted/ramified formal normal forms at each pole and for
twists in the interior of the curve. This covers the general case, and we
particularly study the fundamental groups of the spaces of admissible
deformations of irregular types/classes, in the viewpoint of (twisted/nonsplit)
reflections cosets.
| [
{
"version": "v1",
"created": "Wed, 2 Apr 2025 13:01:09 GMT"
},
{
"version": "v2",
"created": "Sat, 19 Apr 2025 18:06:04 GMT"
}
] | 2025-04-22T00:00:00 | [
[
"Douçot",
"Jean",
""
],
[
"Rembado",
"Gabriele",
""
],
[
"Yamakawa",
"Daisuke",
""
]
] |
2504.01724 | Tianshu Hu | Yuxuan Luo, Zhengkun Rong, Lizhen Wang, Longhao Zhang, Tianshu Hu,
Yongming Zhu | DreamActor-M1: Holistic, Expressive and Robust Human Image Animation
with Hybrid Guidance | null | null | null | null | cs.CV cs.AI | http://creativecommons.org/licenses/by-nc-sa/4.0/ | While recent image-based human animation methods achieve realistic body and
facial motion synthesis, critical gaps remain in fine-grained holistic
controllability, multi-scale adaptability, and long-term temporal coherence,
which leads to their lower expressiveness and robustness. We propose a
diffusion transformer (DiT) based framework, DreamActor-M1, with hybrid
guidance to overcome these limitations. For motion guidance, our hybrid control
signals that integrate implicit facial representations, 3D head spheres, and 3D
body skeletons achieve robust control of facial expressions and body movements,
while producing expressive and identity-preserving animations. For scale
adaptation, to handle various body poses and image scales ranging from
portraits to full-body views, we employ a progressive training strategy using
data with varying resolutions and scales. For appearance guidance, we integrate
motion patterns from sequential frames with complementary visual references,
ensuring long-term temporal coherence for unseen regions during complex
movements. Experiments demonstrate that our method outperforms the
state-of-the-art works, delivering expressive results for portraits,
upper-body, and full-body generation with robust long-term consistency. Project
Page: https://grisoon.github.io/DreamActor-M1/.
| [
{
"version": "v1",
"created": "Wed, 2 Apr 2025 13:30:32 GMT"
},
{
"version": "v2",
"created": "Thu, 3 Apr 2025 14:51:10 GMT"
},
{
"version": "v3",
"created": "Sun, 20 Apr 2025 11:52:01 GMT"
}
] | 2025-04-22T00:00:00 | [
[
"Luo",
"Yuxuan",
""
],
[
"Rong",
"Zhengkun",
""
],
[
"Wang",
"Lizhen",
""
],
[
"Zhang",
"Longhao",
""
],
[
"Hu",
"Tianshu",
""
],
[
"Zhu",
"Yongming",
""
]
] |
2504.01754 | Gleb Fedorov P. | Alexander Yeremeyev, Aleksei Tolstobrov, Gleb Fedorov, Shtefan
Sanduleanu, Peter Shlykov, Sergey Samarin, Shamil Kadyrmetov, Artyom
Muraviev, Aleksey Bolgar, Daria Kalacheva, Viktor Lubsanov, Aleksei Dmitriev,
Evgenia Alekseeva, and Oleg V. Astafiev | Variational preparation of entangled states in a system of transmon
qubits | 6 pages, 4 figures + supplementary | null | null | null | quant-ph | http://creativecommons.org/licenses/by/4.0/ | The conventional method for generating entangled states in qubit systems
relies on applying precise two-qubit entangling gates alongside single-qubit
rotations. However, achieving high-fidelity entanglement demands high accuracy
in two-qubit operations, requiring complex calibration protocols. In this work,
we use a minimally calibrated two-qubit iSwap-like gate, tuned via
straightforward parameter optimization (flux pulse amplitude and duration), to
prepare Bell states and GHZ states experimentally in systems of two and three
transmon qubits. By integrating this gate into a variational quantum algorithm
(VQA), we bypass the need for intricate calibration while maintaining high
fidelity. Our proposed methodology employs variational quantum algorithms
(VQAs) to create the target quantum state through imperfect multiqubit
operations. Furthermore, we experimentally demonstrate a violation of the
Clauser-Horne-Shimony-Holt (CHSH) inequality for Bell states, confirming their
high fidelity of preparation.
| [
{
"version": "v1",
"created": "Wed, 2 Apr 2025 14:09:26 GMT"
},
{
"version": "v2",
"created": "Mon, 21 Apr 2025 15:09:23 GMT"
}
] | 2025-04-22T00:00:00 | [
[
"Yeremeyev",
"Alexander",
""
],
[
"Tolstobrov",
"Aleksei",
""
],
[
"Fedorov",
"Gleb",
""
],
[
"Sanduleanu",
"Shtefan",
""
],
[
"Shlykov",
"Peter",
""
],
[
"Samarin",
"Sergey",
""
],
[
"Kadyrmetov",
"Shamil",
""
],
[
"Muraviev",
"Artyom",
""
],
[
"Bolgar",
"Aleksey",
""
],
[
"Kalacheva",
"Daria",
""
],
[
"Lubsanov",
"Viktor",
""
],
[
"Dmitriev",
"Aleksei",
""
],
[
"Alekseeva",
"Evgenia",
""
],
[
"Astafiev",
"Oleg V.",
""
]
] |
2504.01980 | Ludvig Ericson | Ludvig Ericson, Jos\'e Pedro, Patric Jensfelt | Information Gain Is Not All You Need | 9 pages, 6 figures, under review | null | null | null | cs.RO cs.AI | http://arxiv.org/licenses/nonexclusive-distrib/1.0/ | Autonomous exploration in mobile robotics often involves a trade-off between
two objectives: maximizing environmental coverage and minimizing the total path
length. In the widely used information gain paradigm, exploration is guided by
the expected value of observations. While this approach is effective under
budget-constrained settings--where only a limited number of observations can be
made--it fails to align with quality-constrained scenarios, in which the robot
must fully explore the environment to a desired level of certainty or quality.
In such cases, total information gain is effectively fixed, and maximizing it
per step can lead to inefficient, greedy behavior and unnecessary backtracking.
This paper argues that information gain should not serve as an optimization
objective in quality-constrained exploration. Instead, it should be used to
filter viable candidate actions. We propose a novel heuristic, distance
advantage, which selects candidate frontiers based on a trade-off between
proximity to the robot and remoteness from other frontiers. This heuristic aims
to reduce future detours by prioritizing exploration of isolated regions before
the robot's opportunity to visit them efficiently has passed. We evaluate our
method in simulated environments against classical frontier-based exploration
and gain-maximizing approaches. Results show that distance advantage
significantly reduces total path length across a variety of environments, both
with and without access to prior map predictions. Our findings challenge the
assumption that more accurate gain estimation improves performance and offer a
more suitable alternative for the quality-constrained exploration paradigm.
| [
{
"version": "v1",
"created": "Fri, 28 Mar 2025 15:03:52 GMT"
},
{
"version": "v2",
"created": "Thu, 10 Apr 2025 13:43:00 GMT"
},
{
"version": "v3",
"created": "Sun, 20 Apr 2025 13:01:02 GMT"
}
] | 2025-04-22T00:00:00 | [
[
"Ericson",
"Ludvig",
""
],
[
"Pedro",
"José",
""
],
[
"Jensfelt",
"Patric",
""
]
] |
2504.02051 | Alfonso Amayuelas | Alfonso Amayuelas, Jingbo Yang, Saaket Agashe, Ashwin Nagarajan,
Antonis Antoniades, Xin Eric Wang, William Wang | Self-Resource Allocation in Multi-Agent LLM Systems | null | null | null | null | cs.MA cs.AI cs.CL | http://creativecommons.org/licenses/by/4.0/ | With the development of LLMs as agents, there is a growing interest in
connecting multiple agents into multi-agent systems to solve tasks
concurrently, focusing on their role in task assignment and coordination. This
paper explores how LLMs can effectively allocate computational tasks among
multiple agents, considering factors such as cost, efficiency, and performance.
In this work, we address key questions, including the effectiveness of LLMs as
orchestrators and planners, comparing their effectiveness in task assignment
and coordination. Our experiments demonstrate that LLMs can achieve high
validity and accuracy in resource allocation tasks. We find that the planner
method outperforms the orchestrator method in handling concurrent actions,
resulting in improved efficiency and better utilization of agents.
Additionally, we show that providing explicit information about worker
capabilities enhances the allocation strategies of planners, particularly when
dealing with suboptimal workers.
| [
{
"version": "v1",
"created": "Wed, 2 Apr 2025 18:15:41 GMT"
},
{
"version": "v2",
"created": "Sat, 19 Apr 2025 19:05:03 GMT"
}
] | 2025-04-22T00:00:00 | [
[
"Amayuelas",
"Alfonso",
""
],
[
"Yang",
"Jingbo",
""
],
[
"Agashe",
"Saaket",
""
],
[
"Nagarajan",
"Ashwin",
""
],
[
"Antoniades",
"Antonis",
""
],
[
"Wang",
"Xin Eric",
""
],
[
"Wang",
"William",
""
]
] |
2504.02287 | Trung Thanh Nguyen | Trung Thanh Nguyen, Yasutomo Kawanishi, Vijay John, Takahiro Komamizu,
Ichiro Ide | MultiSensor-Home: A Wide-area Multi-modal Multi-view Dataset for Action
Recognition and Transformer-based Sensor Fusion | The 19th IEEE International Conference on Automatic Face and Gesture
Recognition (FG 2025) | null | null | null | cs.CV | http://creativecommons.org/licenses/by-nc-nd/4.0/ | Multi-modal multi-view action recognition is a rapidly growing field in
computer vision, offering significant potential for applications in
surveillance. However, current datasets often fail to address real-world
challenges such as wide-area distributed settings, asynchronous data streams,
and the lack of frame-level annotations. Furthermore, existing methods face
difficulties in effectively modeling inter-view relationships and enhancing
spatial feature learning. In this paper, we introduce the MultiSensor-Home
dataset, a novel benchmark designed for comprehensive action recognition in
home environments, and also propose the Multi-modal Multi-view
Transformer-based Sensor Fusion (MultiTSF) method. The proposed
MultiSensor-Home dataset features untrimmed videos captured by distributed
sensors, providing high-resolution RGB and audio data along with detailed
multi-view frame-level action labels. The proposed MultiTSF method leverages a
Transformer-based fusion mechanism to dynamically model inter-view
relationships. Furthermore, the proposed method integrates a human detection
module to enhance spatial feature learning, guiding the model to prioritize
frames with human activity to enhance action the recognition accuracy.
Experiments on the proposed MultiSensor-Home and the existing MM-Office
datasets demonstrate the superiority of MultiTSF over the state-of-the-art
methods. Quantitative and qualitative results highlight the effectiveness of
the proposed method in advancing real-world multi-modal multi-view action
recognition.
| [
{
"version": "v1",
"created": "Thu, 3 Apr 2025 05:23:08 GMT"
},
{
"version": "v2",
"created": "Mon, 21 Apr 2025 03:46:22 GMT"
}
] | 2025-04-22T00:00:00 | [
[
"Nguyen",
"Trung Thanh",
""
],
[
"Kawanishi",
"Yasutomo",
""
],
[
"John",
"Vijay",
""
],
[
"Komamizu",
"Takahiro",
""
],
[
"Ide",
"Ichiro",
""
]
] |
2504.02438 | Chuanqi Cheng | Chuanqi Cheng, Jian Guan, Wei Wu, Rui Yan | Scaling Video-Language Models to 10K Frames via Hierarchical
Differential Distillation | null | null | null | null | cs.CL cs.AI | http://creativecommons.org/licenses/by-nc-sa/4.0/ | Long-form video processing fundamentally challenges vision-language models
(VLMs) due to the high computational costs of handling extended temporal
sequences. Existing token pruning and feature merging methods often sacrifice
critical temporal dependencies or dilute semantic information. We introduce
differential distillation, a principled approach that systematically preserves
task-relevant information while suppressing redundancy. Based on this
principle, we develop ViLaMP, a hierarchical video-language model that
processes hour-long videos at ``mixed precision'' through two key mechanisms:
(1) differential keyframe selection that maximizes query relevance while
maintaining temporal distinctiveness at the frame level and (2) differential
feature merging that preserves query-salient features in non-keyframes at the
patch level. Hence, ViLaMP retains full information in keyframes while reducing
non-keyframes to their most salient features, resembling mixed-precision
training. Extensive experiments demonstrate ViLaMP's superior performance
across four video understanding benchmarks, particularly on long-form content.
Notably, ViLaMP can process ultra-long videos (up to 10K frames) on a single
NVIDIA A100 GPU, achieving substantial computational efficiency while
maintaining state-of-the-art performance.
| [
{
"version": "v1",
"created": "Thu, 3 Apr 2025 09:55:09 GMT"
},
{
"version": "v2",
"created": "Tue, 8 Apr 2025 15:36:11 GMT"
},
{
"version": "v3",
"created": "Mon, 21 Apr 2025 15:12:34 GMT"
}
] | 2025-04-22T00:00:00 | [
[
"Cheng",
"Chuanqi",
""
],
[
"Guan",
"Jian",
""
],
[
"Wu",
"Wei",
""
],
[
"Yan",
"Rui",
""
]
] |
2504.02450 | Jingyi Wang | Jingyi Wang, Duanfeng Chu, Zejian Deng, Liping Lu and Pan Zhou | CHARMS: A Cognitive Hierarchical Agent for Reasoning and Motion
Stylization in Autonomous Driving | null | null | null | null | cs.RO cs.AI cs.LG | http://arxiv.org/licenses/nonexclusive-distrib/1.0/ | To address the challenges of limited behavioral intelligence and overly
simplified vehicle behavior modeling in autonomous driving simulations, this
paper proposes the Cognitive Hierarchical Agent for Reasoning and Motion
Stylization (CHARMS). Leveraging Level-k game theory, we model human driver
decision-making using reinforcement learning pretraining and supervised
fine-tuning. This enables the resulting models to exhibit diverse behaviors,
improving the intelligence and realism of surrounding vehicles in simulation.
Building upon this capability, we further develop a scenario generation
framework that utilizes the Poisson cognitive hierarchy theory to control the
distribution of vehicles with different driving styles through Poisson and
binomial sampling. Experimental results demonstrate that CHARMS is capable of
both making intelligent decisions as an ego vehicle and generating diverse,
realistic driving scenarios as surrounding vehicles. The code for CHARMS will
be released at https://github.com/WUTAD-Wjy/CHARMS.
| [
{
"version": "v1",
"created": "Thu, 3 Apr 2025 10:15:19 GMT"
},
{
"version": "v2",
"created": "Sun, 20 Apr 2025 13:17:56 GMT"
}
] | 2025-04-22T00:00:00 | [
[
"Wang",
"Jingyi",
""
],
[
"Chu",
"Duanfeng",
""
],
[
"Deng",
"Zejian",
""
],
[
"Lu",
"Liping",
""
],
[
"Zhou",
"Pan",
""
]
] |
2504.02908 | Arianna Foschi | The GRAVITY Collaboration: K. Abd El Dayem, R. Abuter, N. Aimar, P.
Amaro Seoane, A. Amorim, J.P. Berger, H. Bonnet, G. Bourdarot, W. Brandner,
V. Cardoso, Y. Cl\'enet, R. Davies, P.T. de Zeeuw, A. Drescher, A. Eckart, F.
Eisenhauer, H. Feuchtgruber, G. Finger, N.M. F\"orster Schreiber, A. Foschi,
P. Garcia, E. Gendron, R. Genzel, S. Gillessen, M. Hartl, X. Haubois, F.
Haussmann, T. Henning, S. Hippler, M. Horrobin, L. Jochum, L. Jocou, A.
Kaufer, P. Kervella, S. Lacour, V. Lapeyr\`ere, J.-B. Le Bouquin, P. L\'ena,
D. Lutz, F. Mang, N. More, J. Osorno, T. Ott, T. Paumard, K. Perraut, G.
Perrin, S. Rabien, D.C. Ribeiro, M. Sadun Bordoni, S. Scheithauer, J.
Shangguan, T. Shimizu, J. Stadler, O. Straub, C. Straubmeier, E. Sturm, L.J.
Tacconi, I. Urso, F. Vincent, S.D. von Fellenberg, E. Wieprecht, J. Woillez | On the presence of a fifth force at the Galactic Center | Updated to match published version. Accepted on A&A | null | null | null | astro-ph.GA astro-ph.IM gr-qc | http://creativecommons.org/licenses/by/4.0/ | Aims: The presence of a Yukawa-like correction to Newtonian gravity is
investigated at the Galactic Center, leading to a new upper limit for the
intensity of such a correction. Methods: We perform a Markov Chain Monte Carlo
analysis using the astrometric and spectroscopic data of star S$2$ collected at
the Very Large Telescope by GRAVITY, NACO and SINFONI instruments, covering the
period from $1992$ to $2022$. Results: The precision of the GRAVITY instrument
allows us to derive the most stringent upper limit at the Galactic Center for
the intensity of the Yukawa contribution ($\propto \, \alpha e^{- \lambda r}$)
to be $|\alpha| < 0.003$ for a scale length $\lambda = 3 \cdot 10^{13}\, \rm
m\, (\sim 200 \, \rm AU)$. This improves by roughly one order of magnitude all
estimates obtained in previous works.
| [
{
"version": "v1",
"created": "Thu, 3 Apr 2025 09:38:33 GMT"
},
{
"version": "v2",
"created": "Fri, 18 Apr 2025 10:52:19 GMT"
},
{
"version": "v3",
"created": "Mon, 21 Apr 2025 08:07:21 GMT"
}
] | 2025-04-22T00:00:00 | [
[
"The GRAVITY Collaboration",
"",
""
],
[
"Dayem",
"K. Abd El",
""
],
[
"Abuter",
"R.",
""
],
[
"Aimar",
"N.",
""
],
[
"Seoane",
"P. Amaro",
""
],
[
"Amorim",
"A.",
""
],
[
"Berger",
"J. P.",
""
],
[
"Bonnet",
"H.",
""
],
[
"Bourdarot",
"G.",
""
],
[
"Brandner",
"W.",
""
],
[
"Cardoso",
"V.",
""
],
[
"Clénet",
"Y.",
""
],
[
"Davies",
"R.",
""
],
[
"de Zeeuw",
"P. T.",
""
],
[
"Drescher",
"A.",
""
],
[
"Eckart",
"A.",
""
],
[
"Eisenhauer",
"F.",
""
],
[
"Feuchtgruber",
"H.",
""
],
[
"Finger",
"G.",
""
],
[
"Schreiber",
"N. M. Förster",
""
],
[
"Foschi",
"A.",
""
],
[
"Garcia",
"P.",
""
],
[
"Gendron",
"E.",
""
],
[
"Genzel",
"R.",
""
],
[
"Gillessen",
"S.",
""
],
[
"Hartl",
"M.",
""
],
[
"Haubois",
"X.",
""
],
[
"Haussmann",
"F.",
""
],
[
"Henning",
"T.",
""
],
[
"Hippler",
"S.",
""
],
[
"Horrobin",
"M.",
""
],
[
"Jochum",
"L.",
""
],
[
"Jocou",
"L.",
""
],
[
"Kaufer",
"A.",
""
],
[
"Kervella",
"P.",
""
],
[
"Lacour",
"S.",
""
],
[
"Lapeyrère",
"V.",
""
],
[
"Bouquin",
"J. -B. Le",
""
],
[
"Léna",
"P.",
""
],
[
"Lutz",
"D.",
""
],
[
"Mang",
"F.",
""
],
[
"More",
"N.",
""
],
[
"Osorno",
"J.",
""
],
[
"Ott",
"T.",
""
],
[
"Paumard",
"T.",
""
],
[
"Perraut",
"K.",
""
],
[
"Perrin",
"G.",
""
],
[
"Rabien",
"S.",
""
],
[
"Ribeiro",
"D. C.",
""
],
[
"Bordoni",
"M. Sadun",
""
],
[
"Scheithauer",
"S.",
""
],
[
"Shangguan",
"J.",
""
],
[
"Shimizu",
"T.",
""
],
[
"Stadler",
"J.",
""
],
[
"Straub",
"O.",
""
],
[
"Straubmeier",
"C.",
""
],
[
"Sturm",
"E.",
""
],
[
"Tacconi",
"L. J.",
""
],
[
"Urso",
"I.",
""
],
[
"Vincent",
"F.",
""
],
[
"von Fellenberg",
"S. D.",
""
],
[
"Wieprecht",
"E.",
""
],
[
"Woillez",
"J.",
""
]
] |
2504.02989 | Andreino Simonelli | The KM3NeT collaboration | Evaluation of the upgraded 3-inch Hamamatsu photomultiplier for the
KM3NeT Neutrino Telescope | null | null | null | null | hep-ex astro-ph.IM physics.ins-det | http://creativecommons.org/licenses/by/4.0/ | The 3-inch Hamamatsu R14374-02 photomultiplier tube is an improved version of
the R12199-02 model and its successor in the construction of the KM3NeT
neutrino telescope. A total of 1000 photomultipliers were analysed to assess
their dark count rate, transit time spread, and spurious pulses. A subset of
200 photomultipliers were further evaluated to determine their quantum
efficiency which is an essential parameter for Monte Carlo simulations of the
detector response. The measurements show that R14374-02 model has better
quantum efficiency homogeneity over the photocatode and better time properties
than the R12199-02.
| [
{
"version": "v1",
"created": "Thu, 3 Apr 2025 19:29:02 GMT"
},
{
"version": "v2",
"created": "Fri, 18 Apr 2025 22:39:53 GMT"
}
] | 2025-04-22T00:00:00 | [
[
"The KM3NeT collaboration",
"",
""
]
] |
2504.03099 | Jinfan Yang | Jinfan Yang, Leo Foord-Kelcey, Suzuran Takikawa, Nicholas Vining,
Niloy Mitra, Alla Sheffer | Learning Human Perspective in Line Drawings from Single Sketches | null | null | null | null | cs.GR | http://arxiv.org/licenses/nonexclusive-distrib/1.0/ | Artist-drawn sketches only loosely conform to analytical models of
perspective projection. This deviation of human-drawn perspective from
analytical perspective models is persistent and well known, but has yet to be
algorithmically replicated or even well understood. Capturing human perspective
can benefit many computer graphics applications, including sketch-based
modeling and non-photorealistic rendering. We propose the first dedicated
method for learning and replicating human perspective. A core challenge in
learning this perspective is the lack of suitable large-scale data, as well as
the heterogeneity of human drawing choices. We overcome the data paucity by
learning, in a one-shot setup, from a single artist sketch of a given 3D shape
and a best matching analytical camera view of the same shape. We match the
contours of the depicted shape in this view to corresponding artist strokes. We
then learn a spatially continuous local perspective deviation function that
modifies the camera perspective projecting the contours to their corresponding
strokes while retaining key geometric properties that artists strive to
preserve when depicting 3D content. We leverage the observation that artists
employ similar perspectives when depicting shapes from slightly different view
angles to algorithmically augment our training data. First, we use the
perspective function learned from the single example to generate more
human-like contour renders from nearby views; then, we pair these renders with
the analytical camera contours from these views and use these pairs as
additional training data. The resulting learned perspective functions are well
aligned with the training sketch perspectives and are consistent across views.
We compare our results to potential alternatives, demonstrating the superiority
of the proposed approach, and showcasing applications that benefit from learned
human perspective.
| [
{
"version": "v1",
"created": "Fri, 4 Apr 2025 00:57:48 GMT"
},
{
"version": "v2",
"created": "Sat, 19 Apr 2025 23:45:23 GMT"
}
] | 2025-04-22T00:00:00 | [
[
"Yang",
"Jinfan",
""
],
[
"Foord-Kelcey",
"Leo",
""
],
[
"Takikawa",
"Suzuran",
""
],
[
"Vining",
"Nicholas",
""
],
[
"Mitra",
"Niloy",
""
],
[
"Sheffer",
"Alla",
""
]
] |
2504.03595 | Fabio Lilliu | Fabio Lilliu (1), Amir Laadhar (2), Christian Thomsen (3), Diego
Reforgiato Recupero (1), Torben Bach Pedersen (3) ((1) University of
Cagliari, (2) PANTOPIX GmbH & Co. KG, (3) Aalborg University) | Extending the SAREF4ENER Ontology with Flexibility Based on FlexOffers | 13 pages, 5 figures, 4 tables. Submitted to SmartGridComm 2025 | null | null | null | cs.CL | http://creativecommons.org/licenses/by/4.0/ | A key element to support the increased amounts of renewable energy in the
energy system is flexibility, i.e., the possibility of changing energy loads in
time and amount. Many flexibility models have been designed; however, exact
models fail to scale for long time horizons or many devices. Because of this,
the FlexOffer (FOs) model has been designed, to provide device-independent
approximations of flexibility with good accuracy, and much better scaling for
long time horizons and many devices. An important aspect of the real-life
implementation of energy flexibility is enabling flexible data exchange with
many types of smart energy appliances and market systems, e.g., in smart
buildings. For this, ontologies standardizing data formats are required.
However, the current industry standard ontology for integrating smart devices
for energy purposes, SAREF for Energy Flexibility (SAREF4ENER) only has limited
support for flexibility and thus cannot support important use cases. In this
paper we propose an extension of SAREF4ENER that integrates full support for
the complete FlexOffer model, including advanced use cases, while maintaining
backward compatibility. This novel ontology module can accurately describe
flexibility for advanced devices such as electric vehicles, batteries, and heat
pumps. It can also capture the inherent uncertainty associated with many
flexible load types.
| [
{
"version": "v1",
"created": "Fri, 4 Apr 2025 17:02:14 GMT"
},
{
"version": "v2",
"created": "Fri, 18 Apr 2025 21:02:57 GMT"
}
] | 2025-04-22T00:00:00 | [
[
"Lilliu",
"Fabio",
""
],
[
"Laadhar",
"Amir",
""
],
[
"Thomsen",
"Christian",
""
],
[
"Recupero",
"Diego Reforgiato",
""
],
[
"Pedersen",
"Torben Bach",
""
]
] |
2504.03931 | Zixuan Ke | Zixuan Ke, Yifei Ming, Shafiq Joty | NAACL2025 Tutorial: Adaptation of Large Language Models | NAACL2025 Tutorial | null | null | null | cs.CL cs.AI | http://creativecommons.org/licenses/by-nc-sa/4.0/ | This tutorial on adaptation of LLMs is designed to address the growing demand
for models that go beyond the static capabilities of generic LLMs by providing
an overview of dynamic, domain-specific, and task-adaptive LLM adaptation
techniques. While general LLMs have demonstrated strong generalization across a
variety of tasks, they often struggle to perform well in specialized domains
such as finance, healthcare, and code generation for underrepresented
languages. Additionally, their static nature limits their ability to evolve
with the changing world, and they are often extremely large in size, making
them impractical and costly to deploy at scale. As a result, the adaptation of
LLMs has drawn much attention since the birth of LLMs and is of core
importance, both for industry, which focuses on serving its targeted users, and
academia, which can greatly benefit from small but powerful LLMs. To address
this gap, this tutorial aims to provide an overview of the LLM adaptation
techniques. We start with an introduction to LLM adaptation, from both the data
perspective and the model perspective. We then emphasize how the evaluation
metrics and benchmarks are different from other techniques. After establishing
the problems, we explore various adaptation techniques. We categorize
adaptation techniques into two main families. The first is parametric knowledge
adaptation, which focuses on updating the parametric knowledge within LLMs.
Additionally, we will discuss real-time adaptation techniques, including model
editing, which allows LLMs to be updated dynamically in production
environments. The second kind of adaptation is semi-parametric knowledge
adaptation, where the goal is to update LLM parameters to better leverage
external knowledge or tools through techniques like retrieval-augmented
generation (RAG) and agent-based systems.
| [
{
"version": "v1",
"created": "Fri, 4 Apr 2025 20:57:41 GMT"
},
{
"version": "v2",
"created": "Mon, 21 Apr 2025 04:07:33 GMT"
}
] | 2025-04-22T00:00:00 | [
[
"Ke",
"Zixuan",
""
],
[
"Ming",
"Yifei",
""
],
[
"Joty",
"Shafiq",
""
]
] |
2504.03995 | Steven Schaefer | Steven Schaefer, Nathan Varner, Pedro H. Azevedo de Amorim, Max S. New | Intrinsic Verification of Parsers and Formal Grammar Theory in Dependent
Lambek Calculus (Extended Version) | 37 pages, 24 figures; typos corrected, code syntax highlighting,
added reference to artifact | null | null | null | cs.PL cs.FL | http://creativecommons.org/licenses/by/4.0/ | We present Dependent Lambek Calculus, a domain-specific dependent type theory
for verified parsing and formal grammar theory. In Dependent Lambek Calculus,
linear types are used as a syntax for formal grammars, and parsers can be
written as linear terms. The linear typing restriction provides a form of
intrinsic verification that a parser yields only valid parse trees for the
input string. We demonstrate the expressivity of this system by showing that
the combination of inductive linear types and dependency on non-linear data can
be used to encode commonly used grammar formalisms such as regular and
context-free grammars as well as traces of various types of automata. Using
these encodings, we define parsers for regular expressions using deterministic
automata, as well as examples of verified parsers of context-free grammars.
We present a denotational semantics of our type theory that interprets the
types as a mathematical notion of formal grammars. Based on this denotational
semantics, we have made a prototype implementation of Dependent Lambek Calculus
using a shallow embedding in the Agda proof assistant. All of our examples
parsers have been implemented in this prototype implementation.
| [
{
"version": "v1",
"created": "Fri, 4 Apr 2025 23:39:38 GMT"
},
{
"version": "v2",
"created": "Fri, 18 Apr 2025 21:02:08 GMT"
}
] | 2025-04-22T00:00:00 | [
[
"Schaefer",
"Steven",
""
],
[
"Varner",
"Nathan",
""
],
[
"de Amorim",
"Pedro H. Azevedo",
""
],
[
"New",
"Max S.",
""
]
] |
2504.04323 | Yiming Shi | Yiming Shi, Shaoshuai Yang, Xun Zhu, Haoyu Wang, Miao Li, Ji Wu | MedM-VL: What Makes a Good Medical LVLM? | null | null | null | null | cs.CV | http://arxiv.org/licenses/nonexclusive-distrib/1.0/ | Medical image analysis is essential in modern healthcare. Deep learning has
redirected research focus toward complex medical multimodal tasks, including
report generation and visual question answering. Traditional task-specific
models often fall short in handling these challenges. Large vision-language
models (LVLMs) offer new solutions for solving such tasks. In this study, we
build on the popular LLaVA framework to systematically explore model
architectures and training strategies for both 2D and 3D medical LVLMs. We
present extensive empirical findings and practical guidance. To support
reproducibility and future research, we release a modular codebase, MedM-VL,
and two pre-trained models: MedM-VL-2D for 2D medical image analysis and
MedM-VL-CT-Chest for 3D CT-based applications. The code and models are
available at: https://github.com/MSIIP/MedM-VL
| [
{
"version": "v1",
"created": "Sun, 6 Apr 2025 01:44:46 GMT"
},
{
"version": "v2",
"created": "Sun, 20 Apr 2025 15:56:56 GMT"
}
] | 2025-04-22T00:00:00 | [
[
"Shi",
"Yiming",
""
],
[
"Yang",
"Shaoshuai",
""
],
[
"Zhu",
"Xun",
""
],
[
"Wang",
"Haoyu",
""
],
[
"Li",
"Miao",
""
],
[
"Wu",
"Ji",
""
]
] |
2504.04367 | Vinod P | Sameera K. M., Vinod P., Anderson Rocha, Rafidha Rehiman K. A., Mauro
Conti | WeiDetect: Weibull Distribution-Based Defense against Poisoning Attacks
in Federated Learning for Network Intrusion Detection Systems | null | null | null | null | cs.CR cs.AI | http://arxiv.org/licenses/nonexclusive-distrib/1.0/ | In the era of data expansion, ensuring data privacy has become increasingly
critical, posing significant challenges to traditional AI-based applications.
In addition, the increasing adoption of IoT devices has introduced significant
cybersecurity challenges, making traditional Network Intrusion Detection
Systems (NIDS) less effective against evolving threats, and privacy concerns
and regulatory restrictions limit their deployment. Federated Learning (FL) has
emerged as a promising solution, allowing decentralized model training while
maintaining data privacy to solve these issues. However, despite implementing
privacy-preserving technologies, FL systems remain vulnerable to adversarial
attacks. Furthermore, data distribution among clients is not heterogeneous in
the FL scenario. We propose WeiDetect, a two-phase, server-side defense
mechanism for FL-based NIDS that detects malicious participants to address
these challenges. In the first phase, local models are evaluated using a
validation dataset to generate validation scores. These scores are then
analyzed using a Weibull distribution, identifying and removing malicious
models. We conducted experiments to evaluate the effectiveness of our approach
in diverse attack settings. Our evaluation included two popular datasets,
CIC-Darknet2020 and CSE-CIC-IDS2018, tested under non-IID data distributions.
Our findings highlight that WeiDetect outperforms state-of-the-art defense
approaches, improving higher target class recall up to 70% and enhancing the
global model's F1 score by 1% to 14%.
| [
{
"version": "v1",
"created": "Sun, 6 Apr 2025 05:31:24 GMT"
},
{
"version": "v2",
"created": "Sat, 19 Apr 2025 15:07:05 GMT"
}
] | 2025-04-22T00:00:00 | [
[
"M.",
"Sameera K.",
""
],
[
"P.",
"Vinod",
""
],
[
"Rocha",
"Anderson",
""
],
[
"A.",
"Rafidha Rehiman K.",
""
],
[
"Conti",
"Mauro",
""
]
] |
2504.04417 | Eoin \'O Colg\'ain | Eoin \'O Colg\'ain, Saeed Pourojaghi, M.M. Sheikh-Jabbari, Lu Yin | How much has DESI dark energy evolved since DR1? | 6 Pages, 5 figures, 3 Tables; v2, references added, typos corrected,
further comments welcome | null | null | null | astro-ph.CO hep-ph hep-th | http://arxiv.org/licenses/nonexclusive-distrib/1.0/ | DESI has reported a dynamical dark energy (DE) signal based on the $w_0
w_a$CDM model that is in conflict with Hubble tension. Recalling that the
combination of DESI DR1 BAO and DR1 full-shape (FS) modeling are consistent
with $\Lambda$CDM, in this letter we comment on the status of fluctuations in
DR1 BAO documented in \cite{DESI:2024mwx, Colgain:2024xqj} in the DR2 update.
In particular, we note that neither DR1 BAO nor DR2 BAO nor DR2 BAO+CMB
confronted to the $w_0 w_a$CDM model with relaxed model parameter priors
confirm late-time accelerated expansion today. Translating DESI BAO constraints
into flat $\Lambda$CDM constraints, we observe that the LRG1 constraint remains
the most prominent outlier, a distinction now held jointly with ELG1, LRG2
switches from smaller to larger $\Omega_m$ values relative to
Planck-$\Lambda$CDM, and ELG data drive the relatively low $\Omega_m$ in the
full DR2 BAO. We observe that one cannot restore $w_0 = -1$ within one $1
\sigma$ by removing either LRG1 or ELG1 or LRG2, but LRG2 in DR2, in contrast
to LRG1 in DR1, now has the greatest bearing on $w_0 > -1$. We conclude that
BAO has yet to stabilise, but the general trend is towards greater consistency
with DESI DR1 FS modeling results, where there may be no dynamical DE signal in
DESI data alone.
| [
{
"version": "v1",
"created": "Sun, 6 Apr 2025 08:58:41 GMT"
},
{
"version": "v2",
"created": "Sat, 19 Apr 2025 09:16:42 GMT"
}
] | 2025-04-22T00:00:00 | [
[
"Colgáin",
"Eoin Ó",
""
],
[
"Pourojaghi",
"Saeed",
""
],
[
"Sheikh-Jabbari",
"M. M.",
""
],
[
"Yin",
"Lu",
""
]
] |
Subsets and Splits
No saved queries yet
Save your SQL queries to embed, download, and access them later. Queries will appear here once saved.