id
string | submitter
string | authors
string | title
string | comments
string | journal-ref
string | doi
string | report-no
string | categories
string | license
string | abstract
string | versions
list | update_date
timestamp[s] | authors_parsed
sequence |
|---|---|---|---|---|---|---|---|---|---|---|---|---|---|
2503.17436 | Cristian Cioflan | Lars Kr\"oger, Cristian Cioflan, Victor Kartsch and Luca Benini | On-Device Federated Continual Learning on RISC-V-based Ultra-Low-Power
SoC for Intelligent Nano-Drone Swarms | 2 pages, 2 tables, 1 figure. Accepted as a poster at the RISC-V
Summit Europe 2025 | null | null | null | cs.LG cs.CV cs.MA | http://arxiv.org/licenses/nonexclusive-distrib/1.0/ | RISC-V-based architectures are paving the way for efficient On-Device
Learning (ODL) in smart edge devices. When applied across multiple nodes, ODL
enables the creation of intelligent sensor networks that preserve data privacy.
However, developing ODL-capable, battery-operated embedded platforms presents
significant challenges due to constrained computational resources and limited
device lifetime, besides intrinsic learning issues such as catastrophic
forgetting. We face these challenges by proposing a regularization-based
On-Device Federated Continual Learning algorithm tailored for multiple
nano-drones performing face recognition tasks. We demonstrate our approach on a
RISC-V-based 10-core ultra-low-power SoC, optimizing the ODL computational
requirements. We improve the classification accuracy by 24% over naive
fine-tuning, requiring 178 ms per local epoch and 10.5 s per global epoch,
demonstrating the effectiveness of the architecture for this task.
| [
{
"version": "v1",
"created": "Fri, 21 Mar 2025 15:53:57 GMT"
},
{
"version": "v2",
"created": "Mon, 21 Apr 2025 23:02:42 GMT"
}
] | 2025-04-23T00:00:00 | [
[
"Kröger",
"Lars",
""
],
[
"Cioflan",
"Cristian",
""
],
[
"Kartsch",
"Victor",
""
],
[
"Benini",
"Luca",
""
]
] |
2503.17614 | Jung Keun Ahn | Woo Seung Jung, Yudai Ichikawa, Byung Min Kang, Jung Keun Ahn, Sung
Wook Choi, Manami Fujita, Takeshi Harada, Shoichi Hasegawa, Shuhei Hayakawa,
Sang Hoon Hwang, Kenneth Hicks, Ken'ichi Imai, Yuji Ishikawa, Shunsuke
Kajikawa, Kento Kamada, Shin Hyung Kim, Tomomasa Kitaoka, Jaeyong Lee, Jong
Won Lee, Koji Miwa, Taito Morino, Fumiya Oura, Hiroyuki Sako, Tamao Sakao,
Masayoshi Saito, Susumu Sato, Hirokazu Tamura, Kiyoshi Tanida, Toshiyuki
Takahashi, Mifuyu Ukai, Shunsuke Wada, Takeshi O. Yamamoto, Seongbae Yang | Cross section Measurements for $^{12}$C$(K^-, K^+\Xi^-)$ and
$^{12}$C$(K^-, K^+\Lambda\Lambda)$ Reactions at 1.8 GeV$/c$ | null | null | null | null | nucl-ex hep-ph | http://arxiv.org/licenses/nonexclusive-distrib/1.0/ | We present a measurement of the production of $\Xi^-$ and $\Lambda\Lambda$ in
the $^{12}$C$(K^-, K^+)$ reaction at an incident beam momentum of 1.8
GeV/$\mathit{c}$, based on high-statistics data from J-PARC E42. The cross
section for the $^{12}$C$(K^-, K^+\Xi^-)$ reaction, compared to the inclusive
$^{12}$C$(K^-, K^+)$ reaction cross section, indicates that the $\Xi^-$
escaping probability peaks at 70\% in the energy region of $E_\Xi=$100 to 150
MeV above the $\Xi^-$ emission threshold. A classical approach using eikonal
approximation shows that the total cross sections for $\Xi^-$ inelastic
scattering ranges between 42 mb and 23 mb in the $\Xi^-$ momentum range from
0.4 to 0.6 GeV/c. Furthermore, based on the relative cross section for the
$^{12}$C$(K^-, K^+\Lambda\Lambda)$ reaction, the total cross section for
$\Xi^-p\to\Lambda\Lambda$ is estimated in the same approach to vary between 2.2
mb and 1.0 mb in the momentum range of 0.40 to 0.65 GeV/c. Specifically, a
cross section of 1.0 mb in the momentum range of 0.5 to 0.6 GeV/c imposes a
constraint on the upper bound of the decay width of the $\Xi^-$ particle in
infinite nuclear matter, revealing $\Gamma_\Xi< \sim 0.6$ MeV.
| [
{
"version": "v1",
"created": "Sat, 22 Mar 2025 02:20:23 GMT"
},
{
"version": "v2",
"created": "Tue, 22 Apr 2025 02:33:57 GMT"
}
] | 2025-04-23T00:00:00 | [
[
"Jung",
"Woo Seung",
""
],
[
"Ichikawa",
"Yudai",
""
],
[
"Kang",
"Byung Min",
""
],
[
"Ahn",
"Jung Keun",
""
],
[
"Choi",
"Sung Wook",
""
],
[
"Fujita",
"Manami",
""
],
[
"Harada",
"Takeshi",
""
],
[
"Hasegawa",
"Shoichi",
""
],
[
"Hayakawa",
"Shuhei",
""
],
[
"Hwang",
"Sang Hoon",
""
],
[
"Hicks",
"Kenneth",
""
],
[
"Imai",
"Ken'ichi",
""
],
[
"Ishikawa",
"Yuji",
""
],
[
"Kajikawa",
"Shunsuke",
""
],
[
"Kamada",
"Kento",
""
],
[
"Kim",
"Shin Hyung",
""
],
[
"Kitaoka",
"Tomomasa",
""
],
[
"Lee",
"Jaeyong",
""
],
[
"Lee",
"Jong Won",
""
],
[
"Miwa",
"Koji",
""
],
[
"Morino",
"Taito",
""
],
[
"Oura",
"Fumiya",
""
],
[
"Sako",
"Hiroyuki",
""
],
[
"Sakao",
"Tamao",
""
],
[
"Saito",
"Masayoshi",
""
],
[
"Sato",
"Susumu",
""
],
[
"Tamura",
"Hirokazu",
""
],
[
"Tanida",
"Kiyoshi",
""
],
[
"Takahashi",
"Toshiyuki",
""
],
[
"Ukai",
"Mifuyu",
""
],
[
"Wada",
"Shunsuke",
""
],
[
"Yamamoto",
"Takeshi O.",
""
],
[
"Yang",
"Seongbae",
""
]
] |
2503.17891 | Fatma Nisa Bostanc{\i} | F. Nisa Bostanc{\i}, O\u{g}uzhan Canpolat, Ataberk Olgun, \.Ismail
Emir Y\"uksel, Konstantinos Kanellopoulos, Mohammad Sadrosadati, A. Giray
Ya\u{g}l{\i}k\c{c}{\i}, Onur Mutlu | Understanding and Mitigating Side and Covert Channel Vulnerabilities
Introduced by RowHammer Defenses | This work was submitted to ISCA 2025 on November 22, 2024 | null | null | null | cs.CR cs.AR | http://creativecommons.org/licenses/by/4.0/ | DRAM chips are vulnerable to read disturbance phenomena (e.g., RowHammer and
RowPress), where repeatedly accessing or keeping open a DRAM row causes
bitflips in nearby rows. Attackers leverage RowHammer bitflips in real systems
to take over systems and leak data. Consequently, many prior works propose
mitigations, including recent DDR specifications introducing new mitigations
(e.g., PRAC and RFM). For robust operation, it is critical to analyze other
security implications of RowHammer mitigations. Unfortunately, no prior work
analyzes the timing covert and side channel vulnerabilities introduced by
RowHammer mitigations.
This paper presents the first analysis and evaluation of timing covert and
side channel vulnerabilities introduced by state-of-the-art RowHammer
mitigations. We demonstrate that RowHammer mitigations' preventive actions have
two fundamental features that enable timing channels. First, preventive actions
reduce DRAM bandwidth availability, resulting in longer memory latencies.
Second, preventive actions can be triggered on demand depending on memory
access patterns.
We introduce LeakyHammer, a new class of attacks that leverage the RowHammer
mitigation-induced memory latency differences to establish communication
channels and leak secrets. First, we build two covert channel attacks
exploiting two state-of-the-art RowHammer mitigations, achieving 38.6 Kbps and
48.6 Kbps channel capacity. Second, we demonstrate a website fingerprinting
attack that identifies visited websites based on the RowHammer-preventive
actions they cause. We propose and evaluate two countermeasures against
LeakyHammer and show that fundamentally mitigating LeakyHammer induces large
overheads in highly RowHammer-vulnerable systems. We believe and hope our work
can enable and aid future work on designing robust systems against RowHammer
mitigation-based side and covert channels.
| [
{
"version": "v1",
"created": "Sun, 23 Mar 2025 00:26:47 GMT"
},
{
"version": "v2",
"created": "Tue, 22 Apr 2025 16:34:57 GMT"
}
] | 2025-04-23T00:00:00 | [
[
"Bostancı",
"F. Nisa",
""
],
[
"Canpolat",
"Oğuzhan",
""
],
[
"Olgun",
"Ataberk",
""
],
[
"Yüksel",
"İsmail Emir",
""
],
[
"Kanellopoulos",
"Konstantinos",
""
],
[
"Sadrosadati",
"Mohammad",
""
],
[
"Yağlıkçı",
"A. Giray",
""
],
[
"Mutlu",
"Onur",
""
]
] |
2503.18869 | Rui Xie | Rui Xie, Asad Ul Haq, Linsen Ma, Yunhua Fang, Zirak Burzin Engineer,
Liu Liu, Tong Zhang | Reimagining Memory Access for LLM Inference: Compression-Aware Memory
Controller Design | 9 pages, 11 figures | null | null | null | cs.AR | http://arxiv.org/licenses/nonexclusive-distrib/1.0/ | The efficiency of Large Language Model~(LLM) inference is often constrained
by substantial memory bandwidth and capacity demands. Existing techniques, such
as pruning, quantization, and mixture of experts/depth, reduce memory capacity
and/or bandwidth consumption at the cost of slight degradation in inference
quality. This paper introduces a design solution that further alleviates memory
bottlenecks by enhancing the on-chip memory controller in AI accelerators to
achieve two main objectives: (1) significantly reducing memory capacity and
bandwidth usage through lossless block compression~(e.g., LZ4 and ZSTD) of
model weights and key-value (KV) cache without compromising inference quality,
and (2) enabling memory bandwidth and energy consumption to scale
proportionally with context-dependent dynamic quantization. These goals are
accomplished by equipping the on-chip memory controller with mechanisms to
improve fine-grained bit-level accessibility and compressibility of weights and
KV cache through LLM-aware configuration of in-memory placement and
representation. Experimental results on publicly available LLMs demonstrate the
effectiveness of this approach, showing memory footprint reductions of 25.2\%
for model weights and 46.9\% for KV cache. In addition, our hardware prototype
at 4\,GHz and 32 lanes (7\,nm) achieves 8\,TB/s throughput with a modest area
overhead (under 3.8\,mm\(^2\)), which underscores the viability of LLM-aware
memory control as a key to efficient large-scale inference.
| [
{
"version": "v1",
"created": "Mon, 24 Mar 2025 16:44:32 GMT"
},
{
"version": "v2",
"created": "Thu, 27 Mar 2025 17:48:14 GMT"
},
{
"version": "v3",
"created": "Mon, 21 Apr 2025 22:13:07 GMT"
}
] | 2025-04-23T00:00:00 | [
[
"Xie",
"Rui",
""
],
[
"Haq",
"Asad Ul",
""
],
[
"Ma",
"Linsen",
""
],
[
"Fang",
"Yunhua",
""
],
[
"Engineer",
"Zirak Burzin",
""
],
[
"Liu",
"Liu",
""
],
[
"Zhang",
"Tong",
""
]
] |
2503.19250 | Charlie Wu | Charlie Wu | Minimal Energy Local Systems on Curves | fixed typos and added some references, results are unchanged; 33
pages, comments welcome! | null | null | null | math.AG math.GT | http://creativecommons.org/licenses/by/4.0/ | Let $\Sigma_{g,d}$ an orientable topological surface of genus $g$ with $d$
punctures. When $g = 0$, Deroin and Tholozan studied the class of supra-maximal
representations $\pi_1(\Sigma_{0,d})\to \mathrm{PSL}_2(\mathbb{R})$, and they
showed that the supra-maximal representations form a compact component of a
real relative character variety. We study a collection of rank $n$ local
systems on $\Sigma_{g,d}$ which we call of minimal energy. These are
generalizations of supra-maximal representations, and underlie polarizable
complex variations of Hodge structure for any choice of complex structure on
$\Sigma_{g,d}$. Like the supra-maximal representations, the minimal energy
local systems form a compact connected component of a real relative character
variety.
We show that when the local monodromy data around the punctures is chosen to
be unitary and generic, and the relative character variety is nonempty, these
minimal energy local systems always exist. When $g > 0$ we show that the
minimal energy local systems come from unitary representations of
$\pi_1(\Sigma_{g,d})$. If $g = 0$ we show that they do not always come from
unitary representations, and we study their structure in general.
| [
{
"version": "v1",
"created": "Tue, 25 Mar 2025 01:20:58 GMT"
},
{
"version": "v2",
"created": "Mon, 21 Apr 2025 20:04:00 GMT"
}
] | 2025-04-23T00:00:00 | [
[
"Wu",
"Charlie",
""
]
] |
2503.20003 | Diana Prado Lopes Aude Craik | Diana P. L. Aude Craik | An entanglement protocol to measure atomic parity violation at sub 0.1%
precision | Text and Fig. 2 updated for clarity and to match version submitted to
journal | null | null | null | quant-ph hep-ex physics.atom-ph | http://creativecommons.org/licenses/by/4.0/ | This paper proposes a scheme to measure atomic parity violation (APV) in
barium ions at <0.1% precision. The scheme is based on using multi-ion
entangled states to common-mode reject parity-conserving systematic shifts and
selectively detect a parity-violating vector light shift. This measurement
protocol eliminates the need to suppress a leading systematic by 11 orders of
magnitude, as is required in the single-ion measurement scheme [N. Fortson,
Phys. Rev. Lett. 70, 2383 (1993)]. Furthermore, the protocol can be combined
with the use of integrated photonic waveguides in an architecture that is
scalable to large ion numbers, with a proportional increase in measurement
precision.
| [
{
"version": "v1",
"created": "Tue, 25 Mar 2025 18:43:16 GMT"
},
{
"version": "v2",
"created": "Tue, 22 Apr 2025 15:53:14 GMT"
}
] | 2025-04-23T00:00:00 | [
[
"Craik",
"Diana P. L. Aude",
""
]
] |
2503.20140 | Isaac Goldbring | Nicol\'as Cuervo Ovalle, Isaac Goldbring, and Netanel Levi | The Schr\"oder-Bernstein property for operators on Hilbert spaces | 8 pages; third draft. The main theorem was improved to cover
operators on nonseparable Hilbert spaces. Comments still welcome! | null | null | null | math.LO math.FA math.SP | http://creativecommons.org/licenses/by/4.0/ | We establish that the complete theory of a Hilbert space equipped with a
normal operator has the Schr\"oder-Bernstein property. This answers a question
of Argoty, Berenstein, and the first-named author. We also prove an analogous
statement for unbounded self-adjoint operators.
| [
{
"version": "v1",
"created": "Wed, 26 Mar 2025 01:11:43 GMT"
},
{
"version": "v2",
"created": "Sat, 29 Mar 2025 00:22:08 GMT"
},
{
"version": "v3",
"created": "Mon, 21 Apr 2025 21:53:48 GMT"
}
] | 2025-04-23T00:00:00 | [
[
"Ovalle",
"Nicolás Cuervo",
""
],
[
"Goldbring",
"Isaac",
""
],
[
"Levi",
"Netanel",
""
]
] |
2503.20443 | Antonio De Benedittis | Antonio De Benedittis, Pasquale Migliozzi, Carlos Maximiliano Mollo,
Andreino Simonelli, Daniele Vivolo | Light exposure and temperature effects on quantum efficiency of bialkali
metal photomultipier tubes | 14 pages, 17 figures | null | null | null | physics.ins-det | http://creativecommons.org/licenses/by-nc-sa/4.0/ | In astroparticle experiments, photomultiplier tubes are crucial for detecting
Cherenkov radiation emitted by charged particles, owing to their exceptional
sensitivity to low-intensity light, which is essential for studying high-energy
phenomena associated with astrophysical neutrinos. However, due to their high
sensitivity, PMTs are vulnerable to significant damage to their photocathode
coating when exposed to intense light and/or high temperatures. Although this
scenario is rare under controlled conditions, it can become relevant in
experiments with prolonged exposure to sunlight and elevated temperatures.
This study presents an analysis of the damage threshold and recovery time of
photomultipliers with bialkali metal coatings. The investigation involved
measuring the quantum efficiency of the PMTs before and after exposure to a
xenon lamp for varying durations, simulating sunlight exposure over several
days. Additionally, quantum efficiency was assessed before and after the PMTs
were subjected to thermal stress, providing an evaluation of their performance
under different thermal conditions.
| [
{
"version": "v1",
"created": "Wed, 26 Mar 2025 11:19:33 GMT"
},
{
"version": "v2",
"created": "Tue, 22 Apr 2025 10:15:41 GMT"
}
] | 2025-04-23T00:00:00 | [
[
"De Benedittis",
"Antonio",
""
],
[
"Migliozzi",
"Pasquale",
""
],
[
"Mollo",
"Carlos Maximiliano",
""
],
[
"Simonelli",
"Andreino",
""
],
[
"Vivolo",
"Daniele",
""
]
] |
2503.20507 | Rakesh Nadig | Rakesh Nadig, Vamanan Arulchelvan, Rahul Bera, Taha Shahroodi,
Gagandeep Singh, Andreas Kakolyris, Mohammad Sadrosadati, Jisung Park, Onur
Mutlu | Harmonia: A Multi-Agent Reinforcement Learning Approach to Data
Placement and Migration in Hybrid Storage Systems | null | null | null | null | cs.AR cs.DC cs.LG | http://creativecommons.org/licenses/by/4.0/ | Hybrid storage systems (HSS) combine multiple storage devices with diverse
characteristics to achieve high performance and capacity at low cost. The
performance of an HSS highly depends on the effectiveness of two key policies:
(1) the data-placement policy, which determines the best-fit storage device for
incoming data, and (2) the data-migration policy, which rearranges stored data
across the devices to sustain high HSS performance. Prior works focus on
improving only data placement or only data migration in HSS, which leads to
relatively low HSS performance. Unfortunately, no prior work tries to optimize
both policies together. Our goal is to design a holistic data-management
technique that optimizes both data-placement and data-migration policies to
fully exploit the potential of an HSS, and thus significantly improve system
performance. We demonstrate the need for multiple reinforcement learning (RL)
agents to accomplish our goal. We propose Harmonia, a multi-agent RL-based
data-management technique that employs two lightweight autonomous RL agents, a
data-placement agent and a data-migration agent, which adapt their policies for
the current workload and HSS configuration, and coordinate with each other to
improve overall HSS performance. We evaluate Harmonia on a real HSS with up to
four heterogeneous and diverse storage devices. Our evaluation using 17
data-intensive workloads on performance-optimized (cost-optimized) HSS with two
storage devices shows that, on average, Harmonia outperforms the
best-performing prior approach by 49.5% (31.7%). On an HSS with three (four)
devices, Harmonia outperforms the best-performing prior work by 37.0% (42.0%).
Harmonia's performance benefits come with low latency (240ns for inference) and
storage overheads (206 KiB in DRAM for both RL agents together). We will
open-source Harmonia's implementation to aid future research on HSS.
| [
{
"version": "v1",
"created": "Wed, 26 Mar 2025 12:47:52 GMT"
},
{
"version": "v2",
"created": "Tue, 22 Apr 2025 16:55:45 GMT"
}
] | 2025-04-23T00:00:00 | [
[
"Nadig",
"Rakesh",
""
],
[
"Arulchelvan",
"Vamanan",
""
],
[
"Bera",
"Rahul",
""
],
[
"Shahroodi",
"Taha",
""
],
[
"Singh",
"Gagandeep",
""
],
[
"Kakolyris",
"Andreas",
""
],
[
"Sadrosadati",
"Mohammad",
""
],
[
"Park",
"Jisung",
""
],
[
"Mutlu",
"Onur",
""
]
] |
2503.21979 | Size Wu | Size Wu, Wenwei Zhang, Lumin Xu, Sheng Jin, Zhonghua Wu, Qingyi Tao,
Wentao Liu, Wei Li, Chen Change Loy | Harmonizing Visual Representations for Unified Multimodal Understanding
and Generation | null | null | null | null | cs.CV | http://creativecommons.org/licenses/by-sa/4.0/ | Unifying visual understanding and generation within a single multimodal
framework remains a significant challenge, as the two inherently heterogeneous
tasks require representations at different levels of granularity. Current
approaches that utilize vector quantization (VQ) or variational autoencoders
(VAE) for unified visual representation prioritize intrinsic imagery features
over semantics, compromising understanding performance. In this work, we take
inspiration from masked image modelling (MIM) that learns rich semantics via a
mask-and-reconstruct pre-training and its successful extension to masked
autoregressive (MAR) image generation. A preliminary study on the MAR encoder's
representation reveals exceptional linear probing accuracy and precise feature
response to visual concepts, which indicates MAR's potential for visual
understanding tasks beyond its original generation role. Based on these
insights, we present \emph{Harmon}, a unified autoregressive framework that
harmonizes understanding and generation tasks with a shared MAR encoder.
Through a three-stage training procedure that progressively optimizes
understanding and generation capabilities, Harmon achieves state-of-the-art
image generation results on the GenEval, MJHQ30K and WISE benchmarks while
matching the performance of methods with dedicated semantic encoders (e.g.,
Janus) on image understanding benchmarks. Our code and models will be available
at https://github.com/wusize/Harmon.
| [
{
"version": "v1",
"created": "Thu, 27 Mar 2025 20:50:38 GMT"
},
{
"version": "v2",
"created": "Tue, 22 Apr 2025 04:26:58 GMT"
}
] | 2025-04-23T00:00:00 | [
[
"Wu",
"Size",
""
],
[
"Zhang",
"Wenwei",
""
],
[
"Xu",
"Lumin",
""
],
[
"Jin",
"Sheng",
""
],
[
"Wu",
"Zhonghua",
""
],
[
"Tao",
"Qingyi",
""
],
[
"Liu",
"Wentao",
""
],
[
"Li",
"Wei",
""
],
[
"Loy",
"Chen Change",
""
]
] |
2503.22160 | Yu-Fan Zheng | Yu-fan Zheng, Bin Chen | Structure of Carrollian (conformal) superalgebra | 43 pages, 5 figures, citations added | null | null | null | hep-th | http://creativecommons.org/licenses/by/4.0/ | In this work, we investigate possible supersymmetric extensions of the
Carrollian algebra and the Carrollian conformal algebra in both $d=4$ and
$d=3$. For the super-Carrollian algebra in $d=4$, we identify multiple
admissible structures, depending on the representations of the supercharges
with respect to the Carrollian rotation. Some of these structures can be
derived by taking the speed of light $c\to 0$ limit from super-Poincar\'e
algebra, but others are completely novel. In the conformal case, we demonstrate
that the nontrivial Carrollian superconformal algebras for $d=4$ and $d=3$ are
isomorphic to super-Poincar\'e algebra of $d=5$ and $d=4$ respectively.
Remarkably, neither of these constructions requires R-symmetry to ensure the
algebraic closure. Furthermore, we discover two distinct classes of
super-BMS$_4$ algebras, i.e. one singlet super-BMS$_4$ algebra and two
multiplet chiral super-BMS$_4$ algebras. The singlet case arises from extending
the $3$D Carrollian superconformal algebra, whereas the multiplet cases do not
admit this pathology due to their finite-dimensional subalgebra containing
supercharges with conformal dimension $\Delta=\pm\frac{3}{2}$.
| [
{
"version": "v1",
"created": "Fri, 28 Mar 2025 05:53:26 GMT"
},
{
"version": "v2",
"created": "Tue, 22 Apr 2025 09:31:22 GMT"
}
] | 2025-04-23T00:00:00 | [
[
"Zheng",
"Yu-fan",
""
],
[
"Chen",
"Bin",
""
]
] |
2504.00021 | Arpita Vats | Rahul Raja, Arpita Vats | FUSE : A Ridge and Random Forest-Based Metric for Evaluating MT in
Indigenous Languages | NACCL 2025 | null | null | null | cs.CL | http://creativecommons.org/licenses/by/4.0/ | This paper presents the winning submission of the RaaVa team to the
AmericasNLP 2025 Shared Task 3 on Automatic Evaluation Metrics for Machine
Translation (MT) into Indigenous Languages of America, where our system ranked
first overall based on average Pearson correlation with the human annotations.
We introduce Feature-Union Scorer (FUSE) for Evaluation, FUSE integrates Ridge
regression and Gradient Boosting to model translation quality. In addition to
FUSE, we explore five alternative approaches leveraging different combinations
of linguistic similarity features and learning paradigms. FUSE Score highlights
the effectiveness of combining lexical, phonetic, semantic, and fuzzy token
similarity with learning-based modeling to improve MT evaluation for
morphologically rich and low-resource languages. MT into Indigenous languages
poses unique challenges due to polysynthesis, complex morphology, and
non-standardized orthography. Conventional automatic metrics such as BLEU, TER,
and ChrF often fail to capture deeper aspects like semantic adequacy and
fluency. Our proposed framework, formerly referred to as FUSE, incorporates
multilingual sentence embeddings and phonological encodings to better align
with human evaluation. We train supervised models on human-annotated
development sets and evaluate held-out test data. Results show that FUSE
consistently achieves higher Pearson and Spearman correlations with human
judgments, offering a robust and linguistically informed solution for MT
evaluation in low-resource settings.
| [
{
"version": "v1",
"created": "Fri, 28 Mar 2025 06:58:55 GMT"
},
{
"version": "v2",
"created": "Tue, 22 Apr 2025 05:42:09 GMT"
}
] | 2025-04-23T00:00:00 | [
[
"Raja",
"Rahul",
""
],
[
"Vats",
"Arpita",
""
]
] |
2504.01143 | Ariel A. P\'erez | Rodrigo Lecaros, Juan L\'opez-R\'ios, Ariel A. P\'erez | Lipschitz stability in inverse problems for semi-discrete parabolic
operators | null | null | null | null | math.AP | http://creativecommons.org/licenses/by/4.0/ | This work addresses an inverse problem for a semi-discrete parabolic
equation, which consists of identifying the right-hand side of the equation
based on solution measurements at an intermediate time and within a spatial
subdomain. This result can be applied to establish a stability estimate for the
spatially dependent potential function. Our approach relies on a novel
semi-discrete Carleman estimate, whose parameter is constrained by the mesh
size. As a consequence of the discrete terms arising in the Carleman
inequality, this method naturally introduces an error term related to the
solution's initial condition.
| [
{
"version": "v1",
"created": "Tue, 1 Apr 2025 19:27:10 GMT"
},
{
"version": "v2",
"created": "Tue, 22 Apr 2025 14:45:06 GMT"
}
] | 2025-04-23T00:00:00 | [
[
"Lecaros",
"Rodrigo",
""
],
[
"López-Ríos",
"Juan",
""
],
[
"Pérez",
"Ariel A.",
""
]
] |
2504.01500 | Gustaf Lorentzon | Elias Jarlebring, Gustaf Lorentzon | The Polynomial Set Associated with a Fixed Number of Matrix-Matrix
Multiplications | 23 pages, 1 figure, 1 table | null | null | null | math.NA cs.NA | http://creativecommons.org/licenses/by-nc-nd/4.0/ | We consider the problem of computing matrix polynomials $p(X)$, where $X$ is
a large dense matrix, with as few matrix-matrix multiplications as possible.
More precisely, let $\Pi_{2^{m}}^*$ represent the set of polynomials computable
with $m$ matrix-matrix multiplications, but with an arbitrary number of matrix
additions and scaling operations. We characterize this set through a tabular
parameterization. By deriving equivalence transformations of the tabular
representation, we establish new methods that can be used to construct elements
of $\Pi_{2^{m}}^*$ and determine general properties of the set. The
transformations allow us to eliminate variables and prove that the dimension is
bounded by $m^2$. Numerical simulations suggest that this is a sharp bound.
Consequently, we have identified a parameterization that, to the best of our
knowledge, is the first minimal parameterization. We also conduct a study using
computational tools from algebraic geometry to determine the largest degree $d$
such that all polynomials of that degree belong to $\Pi_{2^{m}}^*$, or its
closure. In many cases, the computational setup is constructive in the sense
that it can also be used to determine a specific evaluation scheme for a given
polynomial.
| [
{
"version": "v1",
"created": "Wed, 2 Apr 2025 08:51:58 GMT"
},
{
"version": "v2",
"created": "Mon, 21 Apr 2025 21:26:58 GMT"
}
] | 2025-04-23T00:00:00 | [
[
"Jarlebring",
"Elias",
""
],
[
"Lorentzon",
"Gustaf",
""
]
] |
2504.01528 | Izumi Seno | Izumi Seno and Shu-ichiro Inutsuka | Analytical and Numerical Linear Analyses of Convection Revisited | 16 pages, 22 figures, accepted for Physics of Fluids | Physics of Fluids 1 April 2025 | 10.1063/5.0263373 | 37 (4): 044118 | physics.flu-dyn astro-ph.GA | http://creativecommons.org/licenses/by/4.0/ | We conduct linear analyses of convection in domains larger than the
temperature scale height. We employ both analytical and numerical methods in
these analyses. In the case excluding all dissipation, the typical time scale
of convection is determined by the free fall time over the temperature scale
height. We quantitatively show the condition for the Boussinesq and
Wentzel-Kramers-Brillouin (WKB) approximations to be applicable. We provide a
reassessment of the critical Rayleigh number, a key indicator of convection,
and show that WKB approximation tends to underestimate the critical Rayleigh
number, particularly when the temperature scale height is comparable to or
smaller than the domain height. We show clear explanation why both thermal
conduction and viscosity are required for stabilizing negative entropy gradient
medium.
| [
{
"version": "v1",
"created": "Wed, 2 Apr 2025 09:13:31 GMT"
},
{
"version": "v2",
"created": "Mon, 7 Apr 2025 01:31:52 GMT"
}
] | 2025-04-23T00:00:00 | [
[
"Seno",
"Izumi",
""
],
[
"Inutsuka",
"Shu-ichiro",
""
]
] |
2504.01801 | Zhijun Wang | Zhijun Wang, Jiahuan Li, Hao Zhou, Rongxiang Weng, Jingang Wang, Xin
Huang, Xue Han, Junlan Feng, Chao Deng, Shujian Huang | Investigating and Scaling up Code-Switching for Multilingual Language
Model Pre-Training | null | null | null | null | cs.CL | http://creativecommons.org/licenses/by/4.0/ | Large language models (LLMs) exhibit remarkable multilingual capabilities
despite the extreme language imbalance in the pre-training data. In this paper,
we closely examine the reasons behind this phenomenon, focusing on the
pre-training corpus. We find that the existence of code-switching, alternating
between different languages within a context, is key to multilingual
capabilities. We conduct an analysis to investigate code-switching in the
pre-training corpus, examining its presence and categorizing it into four types
within two quadrants. We then assess its impact on multilingual performance.
These types of code-switching data are unbalanced in proportions and
demonstrate different effects on facilitating language transfer. To better
explore the power of code-switching for language alignment during pre-training,
we investigate the strategy of synthetic code-switching. We continuously scale
up the synthetic code-switching data and observe remarkable improvements in
both benchmarks and representation space. Extensive experiments indicate that
incorporating synthetic code-switching data enables better language alignment
and generalizes well to high, medium, and low-resource languages with
pre-training corpora of varying qualities.
| [
{
"version": "v1",
"created": "Wed, 2 Apr 2025 15:09:58 GMT"
},
{
"version": "v2",
"created": "Tue, 22 Apr 2025 06:19:26 GMT"
}
] | 2025-04-23T00:00:00 | [
[
"Wang",
"Zhijun",
""
],
[
"Li",
"Jiahuan",
""
],
[
"Zhou",
"Hao",
""
],
[
"Weng",
"Rongxiang",
""
],
[
"Wang",
"Jingang",
""
],
[
"Huang",
"Xin",
""
],
[
"Han",
"Xue",
""
],
[
"Feng",
"Junlan",
""
],
[
"Deng",
"Chao",
""
],
[
"Huang",
"Shujian",
""
]
] |
2504.01889 | Honghao Jing | Honghao Jing | On SYZ mirrors of Hirzebruch surfaces | 50 pages. v2: Added several references | null | null | null | math.SG math.AG | http://arxiv.org/licenses/nonexclusive-distrib/1.0/ | The Strominger-Yau-Zaslow (SYZ) approach to mirror symmetry constructs a
mirror space and a superpotential from the data of a Lagrangian torus fibration
on a K\"ahler manifold with effective first Chern class. For K\"ahler manifolds
whose first Chern class is not nef, the SYZ construction is further complicated
by the presence of additional holomorphic discs with non-positive Maslov index.
In this paper, we study SYZ mirror symmetry for two of the simplest toric
examples: the non-Fano Hirzebruch surfaces F_3 and F_4. For F_3, we determine
the SYZ mirror associated to generic perturbations of the complex structure,
and demonstrate that the SYZ mirror depends on the choice of perturbation. For
F_4, we determine the SYZ mirror for a specific perturbation of complex
structure, where the mirror superpotential is an explicit infinite Laurent
series. Finally, we relate this superpotential to those arising from other
perturbations of F_4 via a scattering diagram.
| [
{
"version": "v1",
"created": "Wed, 2 Apr 2025 16:49:05 GMT"
},
{
"version": "v2",
"created": "Tue, 22 Apr 2025 12:50:43 GMT"
}
] | 2025-04-23T00:00:00 | [
[
"Jing",
"Honghao",
""
]
] |
2504.02407 | Ruitong Xiao | Xiaohui Sun, Ruitong Xiao, Jianye Mo, Bowen Wu, Qun Yu, Baoxun Wang | F5R-TTS: Improving Flow-Matching based Text-to-Speech with Group
Relative Policy Optimization | null | null | null | null | cs.SD eess.AS | http://arxiv.org/licenses/nonexclusive-distrib/1.0/ | We present F5R-TTS, a novel text-to-speech (TTS) system that integrates Group
Relative Policy Optimization (GRPO) into a flow-matching based architecture. By
reformulating the deterministic outputs of flow-matching TTS into probabilistic
Gaussian distributions, our approach enables seamless integration of
reinforcement learning algorithms. During pretraining, we train a
probabilistically reformulated flow-matching based model which is derived from
F5-TTS with an open-source dataset. In the subsequent reinforcement learning
(RL) phase, we employ a GRPO-driven enhancement stage that leverages dual
reward metrics: word error rate (WER) computed via automatic speech recognition
and speaker similarity (SIM) assessed by verification models. Experimental
results on zero-shot voice cloning demonstrate that F5R-TTS achieves
significant improvements in both speech intelligibility (a 29.5% relative
reduction in WER) and speaker similarity (a 4.6% relative increase in SIM
score) compared to conventional flow-matching based TTS systems. Audio samples
are available at https://frontierlabs.github.io/F5R.
| [
{
"version": "v1",
"created": "Thu, 3 Apr 2025 08:57:15 GMT"
},
{
"version": "v2",
"created": "Wed, 9 Apr 2025 02:53:57 GMT"
},
{
"version": "v3",
"created": "Tue, 22 Apr 2025 10:49:31 GMT"
}
] | 2025-04-23T00:00:00 | [
[
"Sun",
"Xiaohui",
""
],
[
"Xiao",
"Ruitong",
""
],
[
"Mo",
"Jianye",
""
],
[
"Wu",
"Bowen",
""
],
[
"Yu",
"Qun",
""
],
[
"Wang",
"Baoxun",
""
]
] |
2504.03732 | Nika Mansouri Ghiasi | Nika Mansouri Ghiasi, Talu G\"uloglu, Harun Mustafa, Can Firtina,
Konstantina Koliogeorgi, Konstantinos Kanellopoulos, Haiyu Mao, Rakesh Nadig,
Mohammad Sadrosadati, Jisung Park, Onur Mutlu | SAGe: A Lightweight Algorithm-Architecture Co-Design for Mitigating the
Data Preparation Bottleneck in Large-Scale Genome Analysis | null | null | null | null | cs.AR cs.DC q-bio.GN | http://creativecommons.org/licenses/by/4.0/ | Given the exponentially growing volumes of genomic data, there are extensive
efforts to accelerate genome analysis. We demonstrate a major bottleneck that
greatly limits and diminishes the benefits of state-of-the-art genome analysis
accelerators: the data preparation bottleneck, where genomic data is stored in
compressed form and needs to be decompressed and formatted first before an
accelerator can operate on it. To mitigate this bottleneck, we propose SAGe, an
algorithm-architecture co-design for highly-compressed storage and
high-performance access of large-scale genomic data. SAGe overcomes the
challenges of mitigating the data preparation bottleneck while maintaining high
compression ratios (comparable to genomic-specific compression algorithms) at
low hardware cost. This is enabled by leveraging key features of genomic
datasets to co-design (i) a new (de)compression algorithm, (ii) hardware, (iii)
storage data layout, and (iv) interface commands to access storage. SAGe stores
data in structures that can be rapidly interpreted and decompressed by
efficient streaming accesses and lightweight hardware. To achieve high
compression ratios using only these lightweight structures, SAGe exploits
unique features of genomic data. We show that SAGe can be seamlessly integrated
with a broad range of genome analysis hardware accelerators to mitigate their
data preparation bottlenecks. Our results demonstrate that SAGe improves the
average end-to-end performance and energy efficiency of two state-of-the-art
genome analysis accelerators by 3.0x-32.1x and 18.8x-49.6x, respectively,
compared to when the accelerators rely on state-of-the-art decompression tools.
| [
{
"version": "v1",
"created": "Mon, 31 Mar 2025 23:36:26 GMT"
},
{
"version": "v2",
"created": "Mon, 21 Apr 2025 22:43:40 GMT"
}
] | 2025-04-23T00:00:00 | [
[
"Ghiasi",
"Nika Mansouri",
""
],
[
"Güloglu",
"Talu",
""
],
[
"Mustafa",
"Harun",
""
],
[
"Firtina",
"Can",
""
],
[
"Koliogeorgi",
"Konstantina",
""
],
[
"Kanellopoulos",
"Konstantinos",
""
],
[
"Mao",
"Haiyu",
""
],
[
"Nadig",
"Rakesh",
""
],
[
"Sadrosadati",
"Mohammad",
""
],
[
"Park",
"Jisung",
""
],
[
"Mutlu",
"Onur",
""
]
] |
2504.04060 | Yuhao Wang | Yuhao Wang, Heyang Liu, Ziyang Cheng, Ronghua Wu, Qunshan Gu, Yanfeng
Wang, Yu Wang | VocalNet: Speech LLM with Multi-Token Prediction for Faster and
High-Quality Generation | null | null | null | null | cs.CL cs.AI cs.SD eess.AS | http://arxiv.org/licenses/nonexclusive-distrib/1.0/ | Speech large language models (LLMs) have emerged as a prominent research
focus in speech processing. We introduce VocalNet-1B and VocalNet-8B, a series
of high-performance, low-latency speech LLMs enabled by a scalable and
model-agnostic training framework designed for real-time voice interaction.
Central to our contribution is the first application of multi-token prediction
(MTP) to speech LLMs. This approach represents a paradigm shift from standard
next-token prediction (NTP), offering simultaneous improvements in generation
speed and quality. Informed by analysis of MTP's effect on speech generation
and experimental comparisons, we designed a straightforward and highly
effective MTP implementation. Experiments demonstrate that VocalNet performs on
par with mainstream Omni LLMs even with limited training data, and
significantly surpasses existing open-source speech LLMs. To foster
reproducibility and community advancement, all model weights, inference code,
training data, and framework implementations have been made publicly available
at https://github.com/SJTU-OmniAgent/VocalNet
| [
{
"version": "v1",
"created": "Sat, 5 Apr 2025 04:57:12 GMT"
},
{
"version": "v2",
"created": "Tue, 22 Apr 2025 07:59:31 GMT"
}
] | 2025-04-23T00:00:00 | [
[
"Wang",
"Yuhao",
""
],
[
"Liu",
"Heyang",
""
],
[
"Cheng",
"Ziyang",
""
],
[
"Wu",
"Ronghua",
""
],
[
"Gu",
"Qunshan",
""
],
[
"Wang",
"Yanfeng",
""
],
[
"Wang",
"Yu",
""
]
] |
2504.04251 | Alberto Martin-Lopez | Davide Molinelli, Alberto Martin-Lopez, Elliott Zackrone, Beyza Eken,
Michael D. Ernst, Mauro Pezz\`e | Tratto: A Neuro-Symbolic Approach to Deriving Axiomatic Test Oracles | Accepted for publication at ISSTA 2025 | null | null | null | cs.SE | http://arxiv.org/licenses/nonexclusive-distrib/1.0/ | This paper presents Tratto, a neuro-symbolic approach that generates
assertions (boolean expressions) that can serve as axiomatic oracles, from
source code and documentation. The symbolic module of Tratto takes advantage of
the grammar of the programming language, the unit under test, and the context
of the unit (its class and available APIs) to restrict the search space of the
tokens that can be successfully used to generate valid oracles. The neural
module of Tratto uses transformers fine-tuned for both deciding whether to
output an oracle or not and selecting the next lexical token to incrementally
build the oracle from the set of tokens returned by the symbolic module. Our
experiments show that Tratto outperforms the state-of-the-art axiomatic oracle
generation approaches, with 73% accuracy, 72% precision, and 61% F1-score,
largely higher than the best results of the symbolic and neural approaches
considered in our study (61%, 62%, and 37%, respectively). Tratto can generate
three times more axiomatic oracles than current symbolic approaches, while
generating 10 times less false positives than GPT4 complemented with few-shot
learning and Chain-of-Thought prompting.
| [
{
"version": "v1",
"created": "Sat, 5 Apr 2025 19:14:35 GMT"
},
{
"version": "v2",
"created": "Tue, 22 Apr 2025 09:50:44 GMT"
}
] | 2025-04-23T00:00:00 | [
[
"Molinelli",
"Davide",
""
],
[
"Martin-Lopez",
"Alberto",
""
],
[
"Zackrone",
"Elliott",
""
],
[
"Eken",
"Beyza",
""
],
[
"Ernst",
"Michael D.",
""
],
[
"Pezzè",
"Mauro",
""
]
] |
2504.04360 | Jie Zhu | Jie Zhu, Yujun Zhu, Ju Ming, Max D. Gunzburger | Splitting Method for Stochastic Navier-Stokes Equations | pages 30 | null | null | null | math.NA cs.NA | http://arxiv.org/licenses/nonexclusive-distrib/1.0/ | This paper investigates the two-dimensional stochastic steady-state
Navier-Stokes(NS) equations with additive random noise. We introduce an
innovative splitting method that decomposes the stochastic NS equations into a
deterministic NS component and a stochastic equation. We rigorously analyze the
proposed splitting method from the perspectives of equivalence, stability,
existence and uniqueness of the solution. We also propose a modified splitting
scheme, which simplified the stochastic equation by omitting its nonlinear
terms. A detailed analysis of the solution properties for this modified
approach is provided. Additionally, we discuss the statistical errors with both
the original splitting format and the modified scheme. Our theoretical and
numerical studies demonstrate that the equivalent splitting scheme exhibits
significantly enhanced stability compared to the original stochastic NS
equations, enabling more effective handling of nonlinear characteristics.
Several numerical experiments were performed to compare the statistical errors
of the splitting method and the modified splitting method. Notably, the
deterministic NS equation in the splitting method does not require repeated
solving, and the stochastic equation in the modified scheme is free of
nonlinear terms. These features make the modified splitting method particularly
advantageous for large-scale computations, as it significantly improves
computational efficiency without compromising accuracy.
| [
{
"version": "v1",
"created": "Sun, 6 Apr 2025 05:15:40 GMT"
},
{
"version": "v2",
"created": "Sat, 12 Apr 2025 12:24:31 GMT"
},
{
"version": "v3",
"created": "Tue, 22 Apr 2025 17:06:46 GMT"
}
] | 2025-04-23T00:00:00 | [
[
"Zhu",
"Jie",
""
],
[
"Zhu",
"Yujun",
""
],
[
"Ming",
"Ju",
""
],
[
"Gunzburger",
"Max D.",
""
]
] |
2504.05138 | Haoran Zhang | Haoran Zhang, Zejun Gong, Zekai Li, Marie Siew, Carlee Joe-Wong,
Rachid El-Azouzi | Towards Optimal Heterogeneous Client Sampling in Multi-Model Federated
Learning | 29 pages with full proofs | null | null | null | cs.LG cs.DC | http://creativecommons.org/licenses/by/4.0/ | Federated learning (FL) allows edge devices to collaboratively train models
without sharing local data. As FL gains popularity, clients may need to train
multiple unrelated FL models, but communication constraints limit their ability
to train all models simultaneously. While clients could train FL models
sequentially, opportunistically having FL clients concurrently train different
models -- termed multi-model federated learning (MMFL) -- can reduce the
overall training time. Prior work uses simple client-to-model assignments that
do not optimize the contribution of each client to each model over the course
of its training. Prior work on single-model FL shows that intelligent client
selection can greatly accelerate convergence, but na\"ive extensions to MMFL
can violate heterogeneous resource constraints at both the server and the
clients. In this work, we develop a novel convergence analysis of MMFL with
arbitrary client sampling methods, theoretically demonstrating the strengths
and limitations of previous well-established gradient-based methods. Motivated
by this analysis, we propose MMFL-LVR, a loss-based sampling method that
minimizes training variance while explicitly respecting communication limits at
the server and reducing computational costs at the clients. We extend this to
MMFL-StaleVR, which incorporates stale updates for improved efficiency and
stability, and MMFL-StaleVRE, a lightweight variant suitable for low-overhead
deployment. Experiments show our methods improve average accuracy by up to
19.1% over random sampling, with only a 5.4% gap from the theoretical optimum
(full client participation).
| [
{
"version": "v1",
"created": "Mon, 7 Apr 2025 14:43:17 GMT"
},
{
"version": "v2",
"created": "Tue, 8 Apr 2025 03:29:49 GMT"
},
{
"version": "v3",
"created": "Mon, 21 Apr 2025 18:13:21 GMT"
}
] | 2025-04-23T00:00:00 | [
[
"Zhang",
"Haoran",
""
],
[
"Gong",
"Zejun",
""
],
[
"Li",
"Zekai",
""
],
[
"Siew",
"Marie",
""
],
[
"Joe-Wong",
"Carlee",
""
],
[
"El-Azouzi",
"Rachid",
""
]
] |
2504.05223 | Henrik Ebel | Torben Schiz, Henrik Ebel | Reducing the Communication of Distributed Model Predictive Control:
Autoencoders and Formation Control | 24 pages, 15 figures | null | null | null | eess.SY cs.DC cs.RO cs.SY | http://creativecommons.org/licenses/by/4.0/ | Communication remains a key factor limiting the applicability of distributed
model predictive control (DMPC) in realistic settings, despite advances in
wireless communication. DMPC schemes can require an overwhelming amount of
information exchange between agents as the amount of data depends on the length
of the predication horizon, for which some applications require a significant
length to formally guarantee nominal asymptotic stability. This work aims to
provide an approach to reduce the communication effort of DMPC by reducing the
size of the communicated data between agents. Using an autoencoder, the
communicated data is reduced by the encoder part of the autoencoder prior to
communication and reconstructed by the decoder part upon reception within the
distributed optimization algorithm that constitutes the DMPC scheme. The choice
of a learning-based reduction method is motivated by structure inherent to the
data, which results from the data's connection to solutions of optimal control
problems. The approach is implemented and tested at the example of formation
control of differential-drive robots, which is challenging for
optimization-based control due to the robots' nonholonomic constraints, and
which is interesting due to the practical importance of mobile robotics. The
applicability of the proposed approach is presented first in form of a
simulative analysis showing that the resulting control performance yields a
satisfactory accuracy. In particular, the proposed approach outperforms the
canonical naive way to reduce communication by reducing the length of the
prediction horizon. Moreover, it is shown that numerical experiments conducted
on embedded computation hardware, with real distributed computation and
wireless communication, work well with the proposed way of reducing
communication even in practical scenarios in which full communication fails.
| [
{
"version": "v1",
"created": "Mon, 7 Apr 2025 16:10:52 GMT"
},
{
"version": "v2",
"created": "Tue, 22 Apr 2025 17:47:12 GMT"
}
] | 2025-04-23T00:00:00 | [
[
"Schiz",
"Torben",
""
],
[
"Ebel",
"Henrik",
""
]
] |
2504.05490 | Sasan Vakili | Sasan Vakili, Manuel Mazo Jr., and Peyman Mohajerin Esfahani | Optimal Bayesian Affine Estimator and Active Learning for the Wiener
Model | 23 pages, 4 figures | null | null | null | cs.LG cs.SY eess.SY | http://arxiv.org/licenses/nonexclusive-distrib/1.0/ | This paper presents a Bayesian estimation framework for Wiener models,
focusing on learning nonlinear output functions under known linear state
dynamics. We derive a closed-form optimal affine estimator for the unknown
parameters, characterized by the so-called "dynamic basis statistics" (DBS).
Several features of the proposed estimator are studied, including Bayesian
unbiasedness, closed-form posterior statistics, error monotonicity in
trajectory length, and consistency condition (also known as persistent
excitation). In the special case of Fourier basis functions, we demonstrate
that the closed-form description is computationally available, as the Fourier
DBS enjoys explicit expressions. Furthermore, we identify an inherent
inconsistency in the Fourier bases for single-trajectory measurements,
regardless of the input excitation. Leveraging the closed-form estimation
error, we develop an active learning algorithm synthesizing input signals to
minimize estimation error. Numerical experiments validate the efficacy of our
approach, showing significant improvements over traditional regularized
least-squares methods.
| [
{
"version": "v1",
"created": "Mon, 7 Apr 2025 20:36:06 GMT"
},
{
"version": "v2",
"created": "Mon, 21 Apr 2025 21:58:45 GMT"
}
] | 2025-04-23T00:00:00 | [
[
"Vakili",
"Sasan",
""
],
[
"Mazo",
"Manuel",
"Jr."
],
[
"Esfahani",
"Peyman Mohajerin",
""
]
] |
2504.06042 | Xu Shi | Xu Shi, Rufeng Xiao, Rujun Jiang | An Adaptive Algorithm for Bilevel Optimization on Riemannian Manifolds | null | null | null | null | math.OC | http://arxiv.org/licenses/nonexclusive-distrib/1.0/ | Existing methods for solving Riemannian bilevel optimization (RBO) problems
require prior knowledge of the problem's first- and second-order information
and curvature parameter of the Riemannian manifold to determine step sizes,
which poses practical limitations when these parameters are unknown or
computationally infeasible to obtain. In this paper, we introduce the Adaptive
Riemannian Hypergradient Descent (AdaRHD) algorithm for solving RBO problems.
To our knowledge, AdaRHD is the first method to incorporate a fully adaptive
step size strategy that eliminates the need for problem-specific parameters in
RBO problem resolution. We prove that AdaRHD achieves an
$\mathcal{O}(1/\epsilon)$ iteration complexity for finding an
$\epsilon$-stationary point, thus matching the complexity of existing
non-adaptive methods. Furthermore, we demonstrate that substituting exponential
mappings with retraction mappings maintains the same complexity bound.
Experiments demonstrate that AdaRHD achieves comparable performance to existing
non-adaptive approaches while exhibiting greater robustness.
| [
{
"version": "v1",
"created": "Tue, 8 Apr 2025 13:42:18 GMT"
},
{
"version": "v2",
"created": "Tue, 22 Apr 2025 13:29:25 GMT"
}
] | 2025-04-23T00:00:00 | [
[
"Shi",
"Xu",
""
],
[
"Xiao",
"Rufeng",
""
],
[
"Jiang",
"Rujun",
""
]
] |
2504.06116 | Davide Sferrazza | Davide Sferrazza, Gabriele Berton, Gabriele Trivigno, Carlo Masone | To Match or Not to Match: Revisiting Image Matching for Reliable Visual
Place Recognition | CVPRW 2025 | null | null | null | cs.CV | http://arxiv.org/licenses/nonexclusive-distrib/1.0/ | Visual Place Recognition (VPR) is a critical task in computer vision,
traditionally enhanced by re-ranking retrieval results with image matching.
However, recent advancements in VPR methods have significantly improved
performance, challenging the necessity of re-ranking. In this work, we show
that modern retrieval systems often reach a point where re-ranking can degrade
results, as current VPR datasets are largely saturated. We propose using image
matching as a verification step to assess retrieval confidence, demonstrating
that inlier counts can reliably predict when re-ranking is beneficial. Our
findings shift the paradigm of retrieval pipelines, offering insights for more
robust and adaptive VPR systems. The code is available at
https://github.com/FarInHeight/To-Match-or-Not-to-Match.
| [
{
"version": "v1",
"created": "Tue, 8 Apr 2025 15:10:10 GMT"
},
{
"version": "v2",
"created": "Tue, 22 Apr 2025 07:44:36 GMT"
}
] | 2025-04-23T00:00:00 | [
[
"Sferrazza",
"Davide",
""
],
[
"Berton",
"Gabriele",
""
],
[
"Trivigno",
"Gabriele",
""
],
[
"Masone",
"Carlo",
""
]
] |
2504.06537 | Shihang Lu | Shihang Lu, Fan Liu, Yifeng Xiong, Zhen Du, Yuanhao Cui, Shuangyang
Li, Weijie Yuan, Jie Yang, and Shi Jin | Sensing With Random Communication Signals | 8 pages, 5 figures, submitted to an IEEE Journal | null | 10.1109/MNET.2025.3562144 | null | eess.SP | http://creativecommons.org/licenses/by-sa/4.0/ | Communication-centric Integrated Sensing and Communication (ISAC) has been
recognized as a promising methodology to implement wireless sensing
functionality over existing network architectures, due to its
cost-effectiveness and backward compatibility to legacy cellular systems.
However, the inherent randomness of the communication signal may incur huge
fluctuations in sensing capabilities, leading to unfavorable detection and
estimation performance. To address this issue, we elaborate on random ISAC
signal processing methods in this article, aiming at improving the sensing
performance without unduly deteriorating the communication functionality.
Specifically, we commence by discussing the fundamentals of sensing with random
communication signals, including the performance metrics and optimal ranging
waveforms. Building on these concepts, we then present a general framework for
random ISAC signal transmission, followed by an in-depth exploration of
time-domain pulse shaping, frequency-domain constellation shaping, and
spatial-domain precoding methods. We provide a comprehensive overview of each
of these topics, including models, results, and design guidelines. Finally, we
conclude this article by identifying several promising research directions for
random ISAC signal transmission.
| [
{
"version": "v1",
"created": "Wed, 9 Apr 2025 02:27:18 GMT"
}
] | 2025-04-23T00:00:00 | [
[
"Lu",
"Shihang",
""
],
[
"Liu",
"Fan",
""
],
[
"Xiong",
"Yifeng",
""
],
[
"Du",
"Zhen",
""
],
[
"Cui",
"Yuanhao",
""
],
[
"Li",
"Shuangyang",
""
],
[
"Yuan",
"Weijie",
""
],
[
"Yang",
"Jie",
""
],
[
"Jin",
"Shi",
""
]
] |
2504.06621 | Haoran Ma | Gang Bao, Jun Lai and Haoran Ma | Computation of shape Taylor expansions | null | null | null | null | math.NA cs.NA | http://arxiv.org/licenses/nonexclusive-distrib/1.0/ | Shape derivative is an important analytical tool for studying scattering
problems involving perturbations in scatterers. Many applications, including
inverse scattering, optimal design, and uncertainty quantification, are based
on shape derivatives. However, computing high order shape derivatives is
challenging due to the complexity of shape calculus. This work introduces a
comprehensive method for computing shape Taylor expansions in two dimensions
using recurrence formulas. The approach is developed under sound-soft,
sound-hard, impedance, and transmission boundary conditions. Additionally, we
apply the shape Taylor expansion to uncertainty quantification in wave
scattering, enabling high order moment estimation for the scattered field under
random boundary perturbations. Numerical examples are provided to illustrate
the effectiveness of the shape Taylor expansion in achieving high order
approximations.
| [
{
"version": "v1",
"created": "Wed, 9 Apr 2025 06:42:21 GMT"
},
{
"version": "v2",
"created": "Tue, 22 Apr 2025 02:49:20 GMT"
}
] | 2025-04-23T00:00:00 | [
[
"Bao",
"Gang",
""
],
[
"Lai",
"Jun",
""
],
[
"Ma",
"Haoran",
""
]
] |
2504.06814 | Yunlu Shu | Yunlu Shu, Jiaxin Jiang, Lei Shi, Tianyu Wang | Revisit Gradient Descent for Geodesically Convex Optimization | null | null | null | null | math.OC | http://arxiv.org/licenses/nonexclusive-distrib/1.0/ | In a seminal work of Zhang and Sra, gradient descent methods for geodesically
convex optimization were comprehensively studied. In particular, based on a
refined use of the triangle comparison theorem of Toponogov, Zhang and Sra
derived a comparison inequality that relates the current iterate, the next
iterate and the optimum point. Since their seminal work, numerous follow-ups
have studied different downstream usages of their comparison lemma. However,
all results along this line relies on strong assumptions, such as bounded
domain assumption or curvature bounded below assumption.
In this work, we introduce the concept of quasilinearization to optimization,
presenting a novel framework for analyzing geodesically convex optimization. By
leveraging this technique, we establish state-of-the-art convergence rates --
for both deterministic and stochastic settings -- under substantially weaker
assumptions than previously required.
| [
{
"version": "v1",
"created": "Wed, 9 Apr 2025 12:08:43 GMT"
},
{
"version": "v2",
"created": "Tue, 22 Apr 2025 06:01:09 GMT"
}
] | 2025-04-23T00:00:00 | [
[
"Shu",
"Yunlu",
""
],
[
"Jiang",
"Jiaxin",
""
],
[
"Shi",
"Lei",
""
],
[
"Wang",
"Tianyu",
""
]
] |
2504.06948 | Dekuan Dong | Dekuan Dong, Yingzhou Li, and Jungong Xue | An improved quantum algorithm for linear autonomous differential
equations via Pad\'e approximation | 50 pages, 15 figures | null | null | null | quant-ph | http://creativecommons.org/licenses/by/4.0/ | We propose a novel quantum algorithm for solving linear autonomous ordinary
differential equations (ODEs) using the Pad\'e approximation. For linear
autonomous ODEs, the discretized solution can be represented by a product of
matrix exponentials. The proposed algorithm approximates the matrix exponential
by the diagonal Pad\'e approximation, which is then encoded into a large,
block-sparse linear system and solved via quantum linear system algorithms
(QLSA). The detailed quantum circuit is given based on quantum oracle access to
the matrix, the inhomogeneous term, and the initial state. The complexity of
the proposed algorithm is analyzed. Compared to the method based on Taylor
approximation, which approximates the matrix exponential using a $k$-th order
Taylor series, the proposed algorithm improves the approximation order $k$ from
two perspectives: 1) the explicit complexity dependency on $k$ is improved, and
2) a smaller $k$ suffices for the same precision. Numerical experiments
demonstrate the advantages of the proposed algorithm comparing to other related
algorithms.
| [
{
"version": "v1",
"created": "Wed, 9 Apr 2025 14:54:27 GMT"
},
{
"version": "v2",
"created": "Tue, 22 Apr 2025 03:24:32 GMT"
}
] | 2025-04-23T00:00:00 | [
[
"Dong",
"Dekuan",
""
],
[
"Li",
"Yingzhou",
""
],
[
"Xue",
"Jungong",
""
]
] |
2504.07003 | Konstantinos Tzirakis | Georgia Karali, Konstantinos Tzirakis, Israel Michael Sigal | The FitzHugh-Nagumo system on undulated cylinders: spontaneous
symmetrization and effective system | 31 pages | Journal of Nonlinear Science, Volume 35, article number 64, (2025) | 10.1007/s00332-025-10152-9 | null | math.AP | http://creativecommons.org/licenses/by-nc-nd/4.0/ | We consider the FitzHugh-Nagumo system on undulated cylindrical surfaces
modeling nerve axons. We show that for sufficiently small radii and for initial
conditions close to radially symmetrical ones, (i) the solutions converge to
their radial averages, and (ii) the latter averages can be approximated by
solutions of a 1+1 dimensional ('radial') system (the effective system)
involving the surface radius function in its coefficients. This perhaps
explains why solutions of the original 1+1 dimensional FitzHugh-Nagumo system
agree so well with experimental data on electrical impulse propagation.
| [
{
"version": "v1",
"created": "Wed, 9 Apr 2025 16:22:58 GMT"
}
] | 2025-04-23T00:00:00 | [
[
"Karali",
"Georgia",
""
],
[
"Tzirakis",
"Konstantinos",
""
],
[
"Sigal",
"Israel Michael",
""
]
] |
2504.07307 | Jingxin Zhan | Jingxin Zhan, Yuchen Xin, Zhihua Zhang | Follow-the-Perturbed-Leader Approaches Best-of-Both-Worlds for the m-Set
Semi-Bandit Problems | null | null | null | null | cs.LG stat.ML | http://arxiv.org/licenses/nonexclusive-distrib/1.0/ | We consider a common case of the combinatorial semi-bandit problem, the
$m$-set semi-bandit, where the learner exactly selects $m$ arms from the total
$d$ arms. In the adversarial setting, the best regret bound, known to be
$\mathcal{O}(\sqrt{nmd})$ for time horizon $n$, is achieved by the well-known
Follow-the-Regularized-Leader (FTRL) policy. However, this requires to
explicitly compute the arm-selection probabilities via optimizing problems at
each time step and sample according to them. This problem can be avoided by the
Follow-the-Perturbed-Leader (FTPL) policy, which simply pulls the $m$ arms that
rank among the $m$ smallest (estimated) loss with random perturbation. In this
paper, we show that FTPL with a Fr\'echet perturbation also enjoys the near
optimal regret bound $\mathcal{O}(\sqrt{nmd\log(d)})$ in the adversarial
setting and approaches best-of-both-world regret bounds, i.e., achieves a
logarithmic regret for the stochastic setting.
| [
{
"version": "v1",
"created": "Wed, 9 Apr 2025 22:07:01 GMT"
},
{
"version": "v2",
"created": "Tue, 22 Apr 2025 15:16:03 GMT"
}
] | 2025-04-23T00:00:00 | [
[
"Zhan",
"Jingxin",
""
],
[
"Xin",
"Yuchen",
""
],
[
"Zhang",
"Zhihua",
""
]
] |
2504.07344 | Gunhee Park | Gunhee Park, Johnnie Gray, Garnet Kin-Lic Chan | Simulating quantum dynamics in two-dimensional lattices with tensor
network influence functional belief propagation | 16 pages, 18 figures | null | null | null | quant-ph cond-mat.str-el physics.comp-ph | http://arxiv.org/licenses/nonexclusive-distrib/1.0/ | Describing nonequilibrium quantum dynamics remains a significant
computational challenge due to the growth of spatial entanglement. The tensor
network influence functional (TN-IF) approach mitigates this problem for
computing the time evolution of local observables by encoding the subsystem's
influence functional path integral as a matrix product state (MPS), thereby
shifting the resource governing computational cost from spatial entanglement to
temporal entanglement. We extend the applicability of the TN-IF method to
two-dimensional lattices by demonstrating its construction on tree lattices and
proposing a belief propagation (BP) algorithm for the TN-IF, termed influence
functional BP (IF-BP), to simulate local observable dynamics on arbitrary
graphs. Even though the BP algorithm introduces uncontrolled approximation
errors on arbitrary graphs, it provides an accurate description for locally
tree-like lattices. Numerical simulations of the kicked Ising model on a
heavy-hex lattice, motivated by a recent quantum experiment, highlight the
effectiveness of the IF-BP method, which demonstrates superior performance in
capturing long-time dynamics where traditional tensor network state-based
methods struggle. Our results further reveal that the temporal entanglement
entropy (TEE) only grows logarithmically with time for this model, resulting in
a polynomial computational cost for the whole method. We further construct a
cluster expansion of IF-BP to introduce loop correlations beyond the BP
approximation, providing a systematic correction to the IF-BP estimate. We
demonstrate the power of the cluster expansion of the IF-BP in simulating the
quantum quench dynamics of the 2D transverse field Ising model, obtaining
numerical results that improve on the state-of-the-art.
| [
{
"version": "v1",
"created": "Thu, 10 Apr 2025 00:04:32 GMT"
},
{
"version": "v2",
"created": "Sat, 19 Apr 2025 20:59:56 GMT"
}
] | 2025-04-23T00:00:00 | [
[
"Park",
"Gunhee",
""
],
[
"Gray",
"Johnnie",
""
],
[
"Chan",
"Garnet Kin-Lic",
""
]
] |
2504.07481 | Tian Xie | Tian Xie, Menghui Jiang, Huanfeng Shen, Huifang Li, Chao Zeng, Jun Ma,
Guanhao Zhang, Liangpei Zhang | A Mechanism-Learning Deeply Coupled Model for Remote Sensing Retrieval
of Global Land Surface Temperature | null | null | null | null | physics.ao-ph cs.LG | http://arxiv.org/licenses/nonexclusive-distrib/1.0/ | Land surface temperature (LST) retrieval from remote sensing data is pivotal
for analyzing climate processes and surface energy budgets. However, LST
retrieval is an ill-posed inverse problem, which becomes particularly severe
when only a single band is available. In this paper, we propose a deeply
coupled framework integrating mechanistic modeling and machine learning to
enhance the accuracy and generalizability of single-channel LST retrieval.
Training samples are generated using a physically-based radiative transfer
model and a global collection of 5810 atmospheric profiles. A physics-informed
machine learning framework is proposed to systematically incorporate the first
principles from classical physical inversion models into the learning workflow,
with optimization constrained by radiative transfer equations. Global
validation demonstrated a 30% reduction in root-mean-square error versus
standalone methods. Under extreme humidity, the mean absolute error decreased
from 4.87 K to 2.29 K (53% improvement). Continental-scale tests across five
continents confirmed the superior generalizability of this model.
| [
{
"version": "v1",
"created": "Thu, 10 Apr 2025 06:19:01 GMT"
},
{
"version": "v2",
"created": "Fri, 11 Apr 2025 01:42:47 GMT"
},
{
"version": "v3",
"created": "Tue, 22 Apr 2025 13:51:47 GMT"
}
] | 2025-04-23T00:00:00 | [
[
"Xie",
"Tian",
""
],
[
"Jiang",
"Menghui",
""
],
[
"Shen",
"Huanfeng",
""
],
[
"Li",
"Huifang",
""
],
[
"Zeng",
"Chao",
""
],
[
"Ma",
"Jun",
""
],
[
"Zhang",
"Guanhao",
""
],
[
"Zhang",
"Liangpei",
""
]
] |
2504.07483 | Tianyu Li | Bo Wang, Tianyu Li, Ruishi Li, Umang Mathur, Prateek Saxena | Program Skeletons for Automated Program Translation | Accepted by PLDI 2025 (46th ACM SIGPLAN Conference on Programming
Language Design and Implementation) | null | 10.1145/3729287 | null | cs.PL cs.SE | http://arxiv.org/licenses/nonexclusive-distrib/1.0/ | Translating software between programming languages is a challenging task, for
which automated techniques have been elusive and hard to scale up to larger
programs. A key difficulty in cross-language translation is that one has to
re-express the intended behavior of the source program into idiomatic
constructs of a different target language. This task needs abstracting away
from the source language-specific details, while keeping the overall
functionality the same. In this work, we propose a novel and systematic
approach for making such translation amenable to automation based on a
framework we call program skeletons. A program skeleton retains the high-level
structure of the source program by abstracting away and effectively summarizing
lower-level concrete code fragments, which can be mechanically translated to
the target programming language. A skeleton, by design, permits many different
ways of filling in the concrete implementation for fragments, which can work in
conjunction with existing data-driven code synthesizers. Most importantly,
skeletons can conceptually enable sound decomposition, i.e., if each individual
fragment is correctly translated, taken together with the mechanically
translated skeleton, the final translated program is deemed to be correct as a
whole. We present a prototype system called Skel embodying the idea of
skeleton-based translation from Python to JavaScript. Our results show
promising scalability compared to prior works. For 9 real-world Python
programs, some with more than about 1k lines of code, 95% of their code
fragments can be automatically translated, while about 5% require manual
effort. All the final translations are correct with respect to whole-program
test suites.
| [
{
"version": "v1",
"created": "Thu, 10 Apr 2025 06:25:17 GMT"
},
{
"version": "v2",
"created": "Tue, 22 Apr 2025 12:49:11 GMT"
}
] | 2025-04-23T00:00:00 | [
[
"Wang",
"Bo",
""
],
[
"Li",
"Tianyu",
""
],
[
"Li",
"Ruishi",
""
],
[
"Mathur",
"Umang",
""
],
[
"Saxena",
"Prateek",
""
]
] |
2504.07559 | Atsushi Nishizawa | Atsushi Nishizawa, Atsushi Taruya, Yoshiaki Himemoto | Axion dark matter search from terrestrial magnetic fields at extremely
low frequencies | 19 pages, 12 figures, minor changes | null | null | YITP-25-55 | hep-ph astro-ph.CO | http://arxiv.org/licenses/nonexclusive-distrib/1.0/ | The natural environment of the Earth can act as a sensitive detector for dark
matter in ultralight axions. When axions with masses between
$1\times10^{-15}\,{\rm eV}$ and $1\times10^{-13}\,{\rm eV}$ pass through the
Earth, they interact with the global geomagnetic field, generating
electromagnetic (EM) waves in the extremely low-frequency range
($0.3$--$30\,{\rm Hz}$) through axion-photon coupling. This paper is one of a
series of companion papers for~\cite{Taruya:2025zql}, focusing on the data
analysis method and search results for an axion signal. Utilizing the
theoretical predictions of axion-induced EM spectra from a companion study, we
analyzed long-term observational data of terrestrial magnetic fields in this
frequency band to search for axion-induced signals. Our analysis identified 65
persistent signal candidates with a signal-to-noise ratio (SNR) greater than 3.
Aside from these candidates, we placed a new upper bound on the axion-photon
coupling parameter, significantly refining the previous constraint from CAST by
at most two orders of magnitude down to $g_{a\gamma} \lesssim 4\times10^{-13}
\,{\rm GeV}^{-1}$ for the axion mass around $3 \times 10^{-14}\,{\rm eV}$.
| [
{
"version": "v1",
"created": "Thu, 10 Apr 2025 08:40:39 GMT"
},
{
"version": "v2",
"created": "Mon, 21 Apr 2025 01:22:18 GMT"
}
] | 2025-04-23T00:00:00 | [
[
"Nishizawa",
"Atsushi",
""
],
[
"Taruya",
"Atsushi",
""
],
[
"Himemoto",
"Yoshiaki",
""
]
] |
2504.07677 | Jiyong Oh Dr. | Hye-Min Won, Jieun Lee, Jiyong Oh | Localization Meets Uncertainty: Uncertainty-Aware Multi-Modal
Localization | 13 pages, 6 figures | null | null | null | cs.RO cs.CV | http://creativecommons.org/licenses/by-nc-nd/4.0/ | Reliable localization is critical for robot navigation in complex indoor
environments. In this paper, we propose an uncertainty-aware localization
method that enhances the reliability of localization outputs without modifying
the prediction model itself. This study introduces a percentile-based rejection
strategy that filters out unreliable 3-DoF pose predictions based on aleatoric
and epistemic uncertainties the network estimates. We apply this approach to a
multi-modal end-to-end localization that fuses RGB images and 2D LiDAR data,
and we evaluate it across three real-world datasets collected using a
commercialized serving robot. Experimental results show that applying stricter
uncertainty thresholds consistently improves pose accuracy. Specifically, the
mean position error is reduced by 41.0%, 56.7%, and 69.4%, and the mean
orientation error by 55.6%, 65.7%, and 73.3%, when applying 90%, 80%, and 70%
thresholds, respectively. Furthermore, the rejection strategy effectively
removes extreme outliers, resulting in better alignment with ground truth
trajectories. To the best of our knowledge, this is the first study to
quantitatively demonstrate the benefits of percentile-based uncertainty
rejection in multi-modal end-to-end localization tasks. Our approach provides a
practical means to enhance the reliability and accuracy of localization systems
in real-world deployments.
| [
{
"version": "v1",
"created": "Thu, 10 Apr 2025 12:07:24 GMT"
},
{
"version": "v2",
"created": "Tue, 22 Apr 2025 11:34:10 GMT"
}
] | 2025-04-23T00:00:00 | [
[
"Won",
"Hye-Min",
""
],
[
"Lee",
"Jieun",
""
],
[
"Oh",
"Jiyong",
""
]
] |
2504.07690 | Michele Andreoli | Michele Andreoli | Perihelion precession in non-Newtonian central potentials | Accepted for publication in Astrophysics and Space Science,
SpringerNature | null | null | null | astro-ph.EP gr-qc | http://creativecommons.org/licenses/by-sa/4.0/ | High order corrections to the perihelion precession are obtained in
non-Newtonian central potentials, via complex analysis techniques. The result
is an exact series expansion whose terms, for a perturbation of the form
$\delta V=\frac{\gamma}{r^{s}}$, are calculated in closed form. To validate the
method, the series is applied to the specific case of s=3, and the results are
compared with those presented in literature, which are relate to the
Schwarzschild metric. As a further test, a numerical simulation was carried out
for the case where s=4. The algebraic calculations and numerical simulations
were carried out via software with symbolic capabilities.
| [
{
"version": "v1",
"created": "Thu, 10 Apr 2025 12:22:34 GMT"
},
{
"version": "v2",
"created": "Sun, 20 Apr 2025 15:54:47 GMT"
},
{
"version": "v3",
"created": "Tue, 22 Apr 2025 06:11:51 GMT"
}
] | 2025-04-23T00:00:00 | [
[
"Andreoli",
"Michele",
""
]
] |
2504.07758 | Shuangfan Zhou | Shuangfan Zhou, Chu Zhou, Youwei Lyu, Heng Guo, Zhanyu Ma, Boxin Shi
and Imari Sato | PIDSR: Complementary Polarized Image Demosaicing and Super-Resolution | CVPR 2025 | null | null | null | cs.CV eess.IV | http://creativecommons.org/licenses/by/4.0/ | Polarization cameras can capture multiple polarized images with different
polarizer angles in a single shot, bringing convenience to polarization-based
downstream tasks. However, their direct outputs are color-polarization filter
array (CPFA) raw images, requiring demosaicing to reconstruct full-resolution,
full-color polarized images; unfortunately, this necessary step introduces
artifacts that make polarization-related parameters such as the degree of
polarization (DoP) and angle of polarization (AoP) prone to error. Besides,
limited by the hardware design, the resolution of a polarization camera is
often much lower than that of a conventional RGB camera. Existing polarized
image demosaicing (PID) methods are limited in that they cannot enhance
resolution, while polarized image super-resolution (PISR) methods, though
designed to obtain high-resolution (HR) polarized images from the demosaicing
results, tend to retain or even amplify errors in the DoP and AoP introduced by
demosaicing artifacts. In this paper, we propose PIDSR, a joint framework that
performs complementary Polarized Image Demosaicing and Super-Resolution,
showing the ability to robustly obtain high-quality HR polarized images with
more accurate DoP and AoP from a CPFA raw image in a direct manner. Experiments
show our PIDSR not only achieves state-of-the-art performance on both synthetic
and real data, but also facilitates downstream tasks.
| [
{
"version": "v1",
"created": "Thu, 10 Apr 2025 13:56:33 GMT"
},
{
"version": "v2",
"created": "Tue, 22 Apr 2025 13:13:39 GMT"
}
] | 2025-04-23T00:00:00 | [
[
"Zhou",
"Shuangfan",
""
],
[
"Zhou",
"Chu",
""
],
[
"Lyu",
"Youwei",
""
],
[
"Guo",
"Heng",
""
],
[
"Ma",
"Zhanyu",
""
],
[
"Shi",
"Boxin",
""
],
[
"Sato",
"Imari",
""
]
] |
2504.07796 | Julius Fergy Rabago | Elmehdi Cherrat and Lekbir Afraites and Julius Fergy Tiongson Rabago | Numerical solution by shape optimization method to an inverse shape
problem in multi-dimensional advection-diffusion problem with space dependent
coefficients | null | null | null | null | math.OC cs.NA math.NA | http://creativecommons.org/licenses/by/4.0/ | This work focuses on numerically solving a shape identification problem
related to advection-diffusion processes with space-dependent coefficients
using shape optimization techniques. Two boundary-type cost functionals are
considered, and their corresponding variations with respect to shapes are
derived using the adjoint method, employing the chain rule approach. This
involves firstly utilizing the material derivative of the state system and
secondly using its shape derivative. Subsequently, an alternating direction
method of multipliers (ADMM) combined with the Sobolev-gradient-descent
algorithm is applied to stably solve the shape reconstruction problem.
Numerical experiments in two and three dimensions are conducted to demonstrate
the feasibility of the methods.
| [
{
"version": "v1",
"created": "Thu, 10 Apr 2025 14:36:56 GMT"
},
{
"version": "v2",
"created": "Sat, 12 Apr 2025 09:00:43 GMT"
},
{
"version": "v3",
"created": "Tue, 22 Apr 2025 13:19:29 GMT"
}
] | 2025-04-23T00:00:00 | [
[
"Cherrat",
"Elmehdi",
""
],
[
"Afraites",
"Lekbir",
""
],
[
"Rabago",
"Julius Fergy Tiongson",
""
]
] |
2504.07883 | Stefan Enzner | Stefan Enzner and Jan Berges and Arne Schobert and Dongjin Oh and
Mingu Kang and Riccardo Comin and Ronny Thomale and Tim O. Wehling and
Domenico Di Sante and Giorgio Sangiovanni | Phonon fluctuation diagnostics: Origin of charge order in AV$_3$Sb$_5$
kagome metals | 13 pages, 11 figures | null | null | null | cond-mat.str-el cond-mat.mtrl-sci cond-mat.supr-con | http://creativecommons.org/licenses/by/4.0/ | The microsopic origin of the charge-density wave (CDW) in AV$_3$Sb$_5$ (A =
K, Rb, Cs) kagome metals remains a longstanding question, often revolving
around electron-phonon coupling and purely electronic mechanisms involving Van
Hove scenarios, nesting, and sublattice interference. To reveal the processes
driving the CDW transition, we combine ab-initio calculations analysis of the
phonon self-energy and angle-resolved photoemission spectroscopy (ARPES). Our
momentum-resolved study, supported by ARPES data, reveals that lattice
instabilities in the V-135 family of kagome metals appear to also be driven by
electronic states far from high-symmetry points, where these states exhibit the
strongest coupling with the phonon modes responsible for the CDW distortion.
Footing on an interpretation scheme based on phonon fluctuation diagnostics,
our work challenges and revises theories that so far have exclusively
attributed CDW formation to nesting effects close to the Fermi level.
| [
{
"version": "v1",
"created": "Thu, 10 Apr 2025 15:55:41 GMT"
},
{
"version": "v2",
"created": "Tue, 22 Apr 2025 15:52:03 GMT"
}
] | 2025-04-23T00:00:00 | [
[
"Enzner",
"Stefan",
""
],
[
"Berges",
"Jan",
""
],
[
"Schobert",
"Arne",
""
],
[
"Oh",
"Dongjin",
""
],
[
"Kang",
"Mingu",
""
],
[
"Comin",
"Riccardo",
""
],
[
"Thomale",
"Ronny",
""
],
[
"Wehling",
"Tim O.",
""
],
[
"Di Sante",
"Domenico",
""
],
[
"Sangiovanni",
"Giorgio",
""
]
] |
2504.07902 | Sabir Ramazanov Dr. | E. Babichev, I. Dankovsky, D. Gorbunov, S. Ramazanov, A. Vikman | Biased domain walls: faster annihilation, weaker gravitational waves | 24 pages, 11 figures; more results of numerical simulations added,
fits to GW spectral shapes updated; conclusions unchanged | null | null | null | hep-ph astro-ph.CO hep-th | http://arxiv.org/licenses/nonexclusive-distrib/1.0/ | We study the evolution of domain wall networks and their phenomenological
implications in a model of a real scalar $\chi$, where a $Z_2$-symmetry is
slightly broken by a potential bias $V_{bias}$. It is demonstrated that the
latter triggers domain wall annihilation considerably earlier than previously
thought. Namely, we observe that the scaling relation $t_{ann} \propto
1/V^{2/3}_{bias}$ for the annihilation time $t_{ann}$ fits to the simulation
data better than a commonly assumed $t_{ann} \propto 1/V_{bias}$. As a result,
the energy density of gravitational waves produced by the network of biased
domain walls, for a given tiny $V_{bias}$, is suppressed compared to naive
expectations. The spectral shape of gravitational waves is similar to that
resulting from unbiased domain walls, but with more power in the
close-to-maximum ultraviolet part. In the far ultraviolet region, the spectrum
of gravitational waves becomes nearly flat; such a plateau has been recognised
earlier in the case of unbiased walls. In our investigation we mainly focus on
the symmetry breaking potential $V_{breaking} \propto \chi^3$, and argue that
no significant modifications of the domain walls evolution take place if one
includes higher powers of $\chi$.
| [
{
"version": "v1",
"created": "Thu, 10 Apr 2025 16:26:08 GMT"
},
{
"version": "v2",
"created": "Tue, 22 Apr 2025 09:48:42 GMT"
}
] | 2025-04-23T00:00:00 | [
[
"Babichev",
"E.",
""
],
[
"Dankovsky",
"I.",
""
],
[
"Gorbunov",
"D.",
""
],
[
"Ramazanov",
"S.",
""
],
[
"Vikman",
"A.",
""
]
] |
2504.07989 | Raj Dandekar | Nirvan Patil, Malhar Abhay Inamdar, Agnivo Gosai, Guruprasad Pathak,
Anish Joshi, Aryan Sagavekar, Anish Joshirao, Raj Dandekar, Rajat Dandekar,
Sreedath Panat | Regional Tiny Stories: Using Small Models to Compare Language Learning
and Tokenizer Performance | 34 pages, 24 figures, 16 tables | null | null | null | cs.CL cs.AI | http://creativecommons.org/licenses/by/4.0/ | Small Language Models (SLMs) offer efficient alternatives to LLMs for
specific domains. The 2023 TinyStories study developed an English dataset that
allows SLMs with 1 to 10 million parameters to produce coherent outputs. Our
research expands this framework by translating the original dataset into Indian
languages and creating synthetic data using LLMs. We focus on Hindi, Marathi,
and Bengali, evaluating SLMs for regional language processing and understanding
linguistic complexity. We show that SLMs efficiently process regional languages
with significantly fewer parameters than LLMs, providing a complementary
framework for ``inference based evaluation" of tokenization strategies and
linguistic complexity. Our analysis shows that language-specific tokenizers
outperform general-purpose ones for Indian languages. Empirical validations,
supported by information-theoretic and morphological analyses, provides
fundamental understanding behind the better performance of Hindi models over
Marathi and Bengali. Additionally, we show that synthetic datasets outperform
translated content for training SLMs. Correlation analyses reveal
cross-linguistic patterns and language-specific relationships between
creativity, grammatical precision, and narrative completeness. These findings
advance both the practical application of SLMs to underserved languages and our
theoretical understanding of neural language development.
| [
{
"version": "v1",
"created": "Mon, 7 Apr 2025 10:33:14 GMT"
},
{
"version": "v2",
"created": "Tue, 22 Apr 2025 05:18:24 GMT"
}
] | 2025-04-23T00:00:00 | [
[
"Patil",
"Nirvan",
""
],
[
"Inamdar",
"Malhar Abhay",
""
],
[
"Gosai",
"Agnivo",
""
],
[
"Pathak",
"Guruprasad",
""
],
[
"Joshi",
"Anish",
""
],
[
"Sagavekar",
"Aryan",
""
],
[
"Joshirao",
"Anish",
""
],
[
"Dandekar",
"Raj",
""
],
[
"Dandekar",
"Rajat",
""
],
[
"Panat",
"Sreedath",
""
]
] |
2504.08365 | Xueping Zhang | Xueping Zhang, Yaxiong Chen, Ruilin Yao, Yunfei Zi, Shengwu Xiong | Location-Oriented Sound Event Localization and Detection with Spatial
Mapping and Regression Localization | null | null | null | null | cs.SD eess.AS | http://arxiv.org/licenses/nonexclusive-distrib/1.0/ | Sound Event Localization and Detection (SELD) combines the Sound Event
Detection (SED) with the corresponding Direction Of Arrival (DOA). Recently,
adopted event oriented multi-track methods affect the generality in polyphonic
environments due to the limitation of the number of tracks. To enhance the
generality in polyphonic environments, we propose Spatial Mapping and
Regression Localization for SELD (SMRL-SELD). SMRL-SELD segments the 3D spatial
space, mapping it to a 2D plane, and a new regression localization loss is
proposed to help the results converge toward the location of the corresponding
event. SMRL-SELD is location-oriented, allowing the model to learn event
features based on orientation. Thus, the method enables the model to process
polyphonic sounds regardless of the number of overlapping events. We conducted
experiments on STARSS23 and STARSS22 datasets and our proposed SMRL-SELD
outperforms the existing SELD methods in overall evaluation and polyphony
environments.
| [
{
"version": "v1",
"created": "Fri, 11 Apr 2025 09:00:53 GMT"
},
{
"version": "v2",
"created": "Tue, 22 Apr 2025 04:59:09 GMT"
}
] | 2025-04-23T00:00:00 | [
[
"Zhang",
"Xueping",
""
],
[
"Chen",
"Yaxiong",
""
],
[
"Yao",
"Ruilin",
""
],
[
"Zi",
"Yunfei",
""
],
[
"Xiong",
"Shengwu",
""
]
] |
2504.08681 | Gilles Pag\`es | Harald Luschgy and Gilles Pag\`es | Locally optimal Functional Quantization | 6 pages | null | null | null | math.PR | http://creativecommons.org/licenses/by/4.0/ | In this note we demonstrate that locally optimal functional quantizers for
probability distributions on a Banach space lying in the support of $P$ behave
exactly like globally optimal functional quantizers in terms of
stationarity/self-consistency.
| [
{
"version": "v1",
"created": "Fri, 11 Apr 2025 16:42:27 GMT"
},
{
"version": "v2",
"created": "Tue, 22 Apr 2025 07:47:48 GMT"
}
] | 2025-04-23T00:00:00 | [
[
"Luschgy",
"Harald",
""
],
[
"Pagès",
"Gilles",
""
]
] |
2504.08733 | Jano\v{s} Vidali | Jano\v{s} Vidali | Eigenspace embeddings of imprimitive association schemes | null | null | null | null | math.CO | http://creativecommons.org/licenses/by-sa/4.0/ | For a given symmetric association scheme $\mathcal{A}$ and its eigenspace
$S_j$ there exists a mapping of vertices of $\mathcal{A}$ to unit vectors of
$S_j$, known as the spherical representation of $\mathcal{A}$ in $S_j$, such
that the inner products of these vectors only depend on the relation between
the corresponding vertices; furthermore, these inner products only depend on
the parameters of $\mathcal{A}$. We consider parameters of imprimitive
association schemes listed as open cases in the list of parameters for
quotient-polynomial graphs recently published by Herman and Maleki, and study
embeddings of their substructures into some eigenspaces consistent with
spherical representations of the putative association schemes. Using this, we
obtain nonexistence for two parameter sets for $4$-class association schemes
and one parameter sets for a $5$-class association scheme passing all
previously known feasibility conditions, as well as uniqueness for two
parameter sets for $5$-class association schemes.
| [
{
"version": "v1",
"created": "Fri, 11 Apr 2025 17:58:47 GMT"
},
{
"version": "v2",
"created": "Tue, 22 Apr 2025 15:36:58 GMT"
}
] | 2025-04-23T00:00:00 | [
[
"Vidali",
"Janoš",
""
]
] |
2504.08920 | Nicolas Garrel | Nicolas Garrel | Witt invariants of quaternionic forms | 18 pages | null | null | null | math.RA | http://creativecommons.org/licenses/by/4.0/ | We describe all Witt invariants of anti-hermitian forms over a quaternion
algebra with its canonical involution, and in particular all Witt invariants of
orthogonal groups $O(A,\sigma)$ where $(A,\sigma)$ is an central simple algebra
with orthogonal involution and $A$ has index $2$. They are combinations of
appropriately defined $\lambda$-powers, similarly to the case of quadratic
forms, but the module of invariants is no longer free over those operations.
The method involves extending the scalars to a generic splitting field of $A$,
and controlling the residues of the invariants with respect to valuations
coming from closed points in the Severi-Brauer variety.
| [
{
"version": "v1",
"created": "Fri, 11 Apr 2025 19:01:26 GMT"
},
{
"version": "v2",
"created": "Tue, 22 Apr 2025 10:05:23 GMT"
}
] | 2025-04-23T00:00:00 | [
[
"Garrel",
"Nicolas",
""
]
] |
2504.09025 | Nam Nguyen | Nam Nguyen, Thuan Nguyen, Thinh Nguyen, Bella Bose | Universal Rate-Distortion-Classification Representations for Lossy
Compression | null | null | null | null | cs.IT math.IT | http://arxiv.org/licenses/nonexclusive-distrib/1.0/ | In lossy compression, Wang et al. [1] recently introduced the
rate-distortion-perception-classification function, which supports multi-task
learning by jointly optimizing perceptual quality, classification accuracy, and
reconstruction fidelity. Building on the concept of a universal encoder
introduced in [2], we investigate the universal representations that enable a
broad range of distortion-classification tradeoffs through a single shared
encoder coupled with multiple task-specific decoders. We establish, through
both theoretical analysis and numerical experiments, that for Gaussian source
under mean squared error (MSE) distortion, the entire distortion-classification
tradeoff region can be achieved using a single universal encoder. For general
sources, we characterize the achievable region and identify conditions under
which encoder reuse results in negligible distortion penalty. The experimental
result on the MNIST dataset further supports our theoretical findings. We show
that universal encoders can obtain distortion performance comparable to
task-specific encoders. These results demonstrate the practicality and
effectiveness of the proposed universal framework in multi-task compression
scenarios.
| [
{
"version": "v1",
"created": "Sat, 12 Apr 2025 00:43:59 GMT"
},
{
"version": "v2",
"created": "Tue, 22 Apr 2025 01:54:20 GMT"
}
] | 2025-04-23T00:00:00 | [
[
"Nguyen",
"Nam",
""
],
[
"Nguyen",
"Thuan",
""
],
[
"Nguyen",
"Thinh",
""
],
[
"Bose",
"Bella",
""
]
] |
2504.09192 | Zhiyong Wang | Zhiyong Wang | Towards More Efficient, Robust, Instance-adaptive, and Generalizable
Online Learning | Ph.D. Thesis | null | null | null | cs.LG | http://creativecommons.org/licenses/by/4.0/ | The primary goal of my Ph.D. study is to develop provably efficient and
practical algorithms for data-driven online sequential decision-making under
uncertainty. My work focuses on reinforcement learning (RL), multi-armed
bandits, and their applications, including recommendation systems, computer
networks, video analytics, and large language models (LLMs). Online learning
methods, such as bandits and RL, have demonstrated remarkable success - ranging
from outperforming human players in complex games like Atari and Go to
advancing robotics, recommendation systems, and fine-tuning LLMs. Despite these
successes, many established algorithms rely on idealized models that can fail
under model misspecifications or adversarial perturbations, particularly in
settings where accurate prior knowledge of the underlying model class is
unavailable or where malicious users operate within dynamic systems. These
challenges are pervasive in real-world applications, where robust and adaptive
solutions are critical. Furthermore, while worst-case guarantees provide
theoretical reliability, they often fail to capture instance-dependent
performance, which can lead to more efficient and practical solutions. Another
key challenge lies in generalizing to new, unseen environments, a crucial
requirement for deploying these methods in dynamic and unpredictable settings.
To address these limitations, my research aims to develop more efficient,
robust, instance-adaptive, and generalizable online learning algorithms for
both reinforcement learning and bandits. Towards this end, I focus on
developing more efficient, robust, instance-adaptive, and generalizable for
both general reinforcement learning (RL) and bandits.
| [
{
"version": "v1",
"created": "Sat, 12 Apr 2025 12:17:20 GMT"
},
{
"version": "v2",
"created": "Tue, 22 Apr 2025 02:02:31 GMT"
}
] | 2025-04-23T00:00:00 | [
[
"Wang",
"Zhiyong",
""
]
] |
2504.09397 | Dinh-Quan Tran | Dinh-Quan Tran, Peter J. Olver | Continuous Revival of the Periodic Schr\"{o}dinger Equation with
Piecewise $C^2$ Potential | Made minor corrections and added figures. 21 pages, 24 figures.
Comments are welcome! | null | null | null | math.AP | http://arxiv.org/licenses/nonexclusive-distrib/1.0/ | In this paper, we investigate the revivals of the one-dimensional periodic
Schr\"odinger equation with a piecewise $C^2$ potential function. As has been
observed through numerical simulations of the equation with various initial
data and potential functions, the solution, while remaining fractalized at
irrational times, exhibits a form of revival at rational times. The goal is to
prove that the solution at these rational times is given by a finite linear
combination of translations and dilations of the initial datum, plus an
additional continuous term, which we call "continuous revival". In pursuit of
this result, we present a review of relevant properties of the periodic
Schr\"odinger equation as an eigenvalue problem, including asymptotic results
on both the eigenvalues and eigenfunctions.
| [
{
"version": "v1",
"created": "Sun, 13 Apr 2025 01:43:50 GMT"
},
{
"version": "v2",
"created": "Tue, 22 Apr 2025 17:22:04 GMT"
}
] | 2025-04-23T00:00:00 | [
[
"Tran",
"Dinh-Quan",
""
],
[
"Olver",
"Peter J.",
""
]
] |
2504.09448 | Lin Zhu | Lin Zhu, Xinbing Wang, Chenghu Zhou, Nanyang Ye | Bayesian Cross-Modal Alignment Learning for Few-Shot Out-of-Distribution
Generalization | Accepted by AAAI2023 | null | null | null | cs.CV cs.LG | http://creativecommons.org/licenses/by/4.0/ | Recent advances in large pre-trained models showed promising results in
few-shot learning. However, their generalization ability on two-dimensional
Out-of-Distribution (OoD) data, i.e., correlation shift and diversity shift,
has not been thoroughly investigated. Researches have shown that even with a
significant amount of training data, few methods can achieve better performance
than the standard empirical risk minimization method (ERM) in OoD
generalization. This few-shot OoD generalization dilemma emerges as a
challenging direction in deep neural network generalization research, where the
performance suffers from overfitting on few-shot examples and OoD
generalization errors. In this paper, leveraging a broader supervision source,
we explore a novel Bayesian cross-modal image-text alignment learning method
(Bayes-CAL) to address this issue. Specifically, the model is designed as only
text representations are fine-tuned via a Bayesian modelling approach with
gradient orthogonalization loss and invariant risk minimization (IRM) loss. The
Bayesian approach is essentially introduced to avoid overfitting the base
classes observed during training and improve generalization to broader unseen
classes. The dedicated loss is introduced to achieve better image-text
alignment by disentangling the causal and non-casual parts of image features.
Numerical experiments demonstrate that Bayes-CAL achieved state-of-the-art OoD
generalization performances on two-dimensional distribution shifts. Moreover,
compared with CLIP-like models, Bayes-CAL yields more stable generalization
performances on unseen classes. Our code is available at
https://github.com/LinLLLL/BayesCAL.
| [
{
"version": "v1",
"created": "Sun, 13 Apr 2025 06:13:37 GMT"
},
{
"version": "v2",
"created": "Tue, 22 Apr 2025 10:59:00 GMT"
}
] | 2025-04-23T00:00:00 | [
[
"Zhu",
"Lin",
""
],
[
"Wang",
"Xinbing",
""
],
[
"Zhou",
"Chenghu",
""
],
[
"Ye",
"Nanyang",
""
]
] |
2504.09597 | Zhixuan Pan | Zhixuan Pan, Shaowen Wang, Jian Li | Understanding LLM Behaviors via Compression: Data Generation, Knowledge
Acquisition and Scaling Laws | null | null | null | null | cs.AI cs.IT cs.LG math.IT | http://creativecommons.org/licenses/by/4.0/ | Large Language Models (LLMs) have demonstrated remarkable capabilities across
numerous tasks, yet principled explanations for their underlying mechanisms and
several phenomena, such as scaling laws, hallucinations, and related behaviors,
remain elusive. In this work, we revisit the classical relationship between
compression and prediction, grounded in Kolmogorov complexity and Shannon
information theory, to provide deeper insights into LLM behaviors. By
leveraging the Kolmogorov Structure Function and interpreting LLM compression
as a two-part coding process, we offer a detailed view of how LLMs acquire and
store information across increasing model and data scales -- from pervasive
syntactic patterns to progressively rarer knowledge elements. Motivated by this
theoretical perspective and natural assumptions inspired by Heap's and Zipf's
laws, we introduce a simplified yet representative hierarchical data-generation
framework called the Syntax-Knowledge model. Under the Bayesian setting, we
show that prediction and compression within this model naturally lead to
diverse learning and scaling behaviors of LLMs. In particular, our theoretical
analysis offers intuitive and principled explanations for both data and model
scaling laws, the dynamics of knowledge acquisition during training and
fine-tuning, factual knowledge hallucinations in LLMs. The experimental results
validate our theoretical predictions.
| [
{
"version": "v1",
"created": "Sun, 13 Apr 2025 14:31:52 GMT"
},
{
"version": "v2",
"created": "Thu, 17 Apr 2025 16:53:17 GMT"
},
{
"version": "v3",
"created": "Mon, 21 Apr 2025 15:18:42 GMT"
},
{
"version": "v4",
"created": "Tue, 22 Apr 2025 14:11:33 GMT"
}
] | 2025-04-23T00:00:00 | [
[
"Pan",
"Zhixuan",
""
],
[
"Wang",
"Shaowen",
""
],
[
"Li",
"Jian",
""
]
] |
2504.09697 | Kenan Tang | Kenan Tang, Yanhong Li, Yao Qin | SPICE: A Synergistic, Precise, Iterative, and Customizable Image Editing
Workflow | 24 pages, 21 figures. Figure 9(b) has been accepted by CVPR AI Art
Gallery 2025 | null | null | null | cs.GR cs.CV cs.LG | http://creativecommons.org/licenses/by/4.0/ | Recent prompt-based image editing models have demonstrated impressive
prompt-following capability at structural editing tasks. However, existing
models still fail to perform local edits, follow detailed editing prompts, or
maintain global image quality beyond a single editing step. To address these
challenges, we introduce SPICE, a training-free workflow that accepts arbitrary
resolutions and aspect ratios, accurately follows user requirements, and
improves image quality consistently during more than 100 editing steps. By
synergizing the strengths of a base diffusion model and a Canny edge ControlNet
model, SPICE robustly handles free-form editing instructions from the user.
SPICE outperforms state-of-the-art baselines on a challenging realistic
image-editing dataset consisting of semantic editing (object addition, removal,
replacement, and background change), stylistic editing (texture changes), and
structural editing (action change) tasks. Not only does SPICE achieve the
highest quantitative performance according to standard evaluation metrics, but
it is also consistently preferred by users over existing image-editing methods.
We release the workflow implementation for popular diffusion model Web UIs to
support further research and artistic exploration.
| [
{
"version": "v1",
"created": "Sun, 13 Apr 2025 19:13:04 GMT"
}
] | 2025-04-23T00:00:00 | [
[
"Tang",
"Kenan",
""
],
[
"Li",
"Yanhong",
""
],
[
"Qin",
"Yao",
""
]
] |
2504.09723 | Yuxuan Lu | Dakuo Wang, Ting-Yao Hsu, Yuxuan Lu, Hansu Gu, Limeng Cui, Yaochen
Xie, William Headean, Bingsheng Yao, Akash Veeragouni, Jiapeng Liu, Sreyashi
Nag, Jessie Wang | AgentA/B: Automated and Scalable Web A/BTesting with Interactive LLM
Agents | null | null | null | null | cs.HC cs.CL | http://creativecommons.org/licenses/by-nc-nd/4.0/ | A/B testing experiment is a widely adopted method for evaluating UI/UX design
decisions in modern web applications. Yet, traditional A/B testing remains
constrained by its dependence on the large-scale and live traffic of human
participants, and the long time of waiting for the testing result. Through
formative interviews with six experienced industry practitioners, we identified
critical bottlenecks in current A/B testing workflows. In response, we present
AgentA/B, a novel system that leverages Large Language Model-based autonomous
agents (LLM Agents) to automatically simulate user interaction behaviors with
real webpages. AgentA/B enables scalable deployment of LLM agents with diverse
personas, each capable of navigating the dynamic webpage and interactively
executing multi-step interactions like search, clicking, filtering, and
purchasing. In a demonstrative controlled experiment, we employ AgentA/B to
simulate a between-subject A/B testing with 1,000 LLM agents Amazon.com, and
compare agent behaviors with real human shopping behaviors at a scale. Our
findings suggest AgentA/B can emulate human-like behavior patterns.
| [
{
"version": "v1",
"created": "Sun, 13 Apr 2025 21:10:56 GMT"
},
{
"version": "v2",
"created": "Mon, 21 Apr 2025 23:57:49 GMT"
}
] | 2025-04-23T00:00:00 | [
[
"Wang",
"Dakuo",
""
],
[
"Hsu",
"Ting-Yao",
""
],
[
"Lu",
"Yuxuan",
""
],
[
"Gu",
"Hansu",
""
],
[
"Cui",
"Limeng",
""
],
[
"Xie",
"Yaochen",
""
],
[
"Headean",
"William",
""
],
[
"Yao",
"Bingsheng",
""
],
[
"Veeragouni",
"Akash",
""
],
[
"Liu",
"Jiapeng",
""
],
[
"Nag",
"Sreyashi",
""
],
[
"Wang",
"Jessie",
""
]
] |
2504.09809 | Zhimin Li | Zhimin Li, Haichao Miao, Xinyuan Yan, Valerio Pascucci, Matthew
Berger, Shusen Liu | See or Recall: A Sanity Check for the Role of Vision in Solving
Visualization Question Answer Tasks with Multimodal LLMs | null | null | null | null | cs.HC cs.AI | http://creativecommons.org/licenses/by/4.0/ | Recent developments in multimodal large language models (MLLM) have equipped
language models to reason about vision and language jointly. This permits MLLMs
to both perceive and answer questions about data visualization across a variety
of designs and tasks. Applying MLLMs to a broad range of visualization tasks
requires us to properly evaluate their capabilities, and the most common way to
conduct evaluation is through measuring a model's visualization reasoning
capability, analogous to how we would evaluate human understanding of
visualizations (e.g., visualization literacy). However, we found that in the
context of visualization question answering (VisQA), how an MLLM perceives and
reasons about visualizations can be fundamentally different from how humans
approach the same problem. During the evaluation, even without visualization,
the model could correctly answer a substantial portion of the visualization
test questions, regardless of whether any selection options were provided. We
hypothesize that the vast amount of knowledge encoded in the language model
permits factual recall that supersedes the need to seek information from the
visual signal. It raises concerns that the current VisQA evaluation may not
fully capture the models' visualization reasoning capabilities. To address
this, we propose a comprehensive sanity check framework that integrates a
rule-based decision tree and a sanity check table to disentangle the effects of
"seeing" (visual processing) and "recall" (reliance on prior knowledge). This
validates VisQA datasets for evaluation, highlighting where models are truly
"seeing", positively or negatively affected by the factual recall, or relying
on inductive biases for question answering. Our study underscores the need for
careful consideration in designing future visualization understanding studies
when utilizing MLLMs.
| [
{
"version": "v1",
"created": "Mon, 14 Apr 2025 02:19:28 GMT"
},
{
"version": "v2",
"created": "Mon, 21 Apr 2025 20:52:11 GMT"
}
] | 2025-04-23T00:00:00 | [
[
"Li",
"Zhimin",
""
],
[
"Miao",
"Haichao",
""
],
[
"Yan",
"Xinyuan",
""
],
[
"Pascucci",
"Valerio",
""
],
[
"Berger",
"Matthew",
""
],
[
"Liu",
"Shusen",
""
]
] |
2504.09865 | Isabel Gallegos | Isabel O. Gallegos, Chen Shani, Weiyan Shi, Federico Bianchi, Izzy
Gainsburg, Dan Jurafsky, Robb Willer | Labeling Messages as AI-Generated Does Not Reduce Their Persuasive
Effects | null | null | null | null | cs.CY cs.AI cs.HC | http://arxiv.org/licenses/nonexclusive-distrib/1.0/ | As generative artificial intelligence (AI) enables the creation and
dissemination of information at massive scale and speed, it is increasingly
important to understand how people perceive AI-generated content. One prominent
policy proposal requires explicitly labeling AI-generated content to increase
transparency and encourage critical thinking about the information, but prior
research has not yet tested the effects of such labels. To address this gap, we
conducted a survey experiment (N=1601) on a diverse sample of Americans,
presenting participants with an AI-generated message about several public
policies (e.g., allowing colleges to pay student-athletes), randomly assigning
whether participants were told the message was generated by (a) an expert AI
model, (b) a human policy expert, or (c) no label. We found that messages were
generally persuasive, influencing participants' views of the policies by 9.74
percentage points on average. However, while 94.6% of participants assigned to
the AI and human label conditions believed the authorship labels, labels had no
significant effects on participants' attitude change toward the policies,
judgments of message accuracy, nor intentions to share the message with others.
These patterns were robust across a variety of participant characteristics,
including prior knowledge of the policy, prior experience with AI, political
party, education level, or age. Taken together, these results imply that, while
authorship labels would likely enhance transparency, they are unlikely to
substantially affect the persuasiveness of the labeled content, highlighting
the need for alternative strategies to address challenges posed by AI-generated
information.
| [
{
"version": "v1",
"created": "Mon, 14 Apr 2025 04:22:39 GMT"
},
{
"version": "v2",
"created": "Tue, 22 Apr 2025 02:47:23 GMT"
}
] | 2025-04-23T00:00:00 | [
[
"Gallegos",
"Isabel O.",
""
],
[
"Shani",
"Chen",
""
],
[
"Shi",
"Weiyan",
""
],
[
"Bianchi",
"Federico",
""
],
[
"Gainsburg",
"Izzy",
""
],
[
"Jurafsky",
"Dan",
""
],
[
"Willer",
"Robb",
""
]
] |
2504.09940 | Guowen Li | Guowen Li, Xintong Liu, Shilei Cao, Haoyuan Liang, Mengxuan Chen,
Lixian Zhang, Jinxiao Zhang, Jiuke Wang, Meng Jin, Juepeng Zheng, Haohuan Fu | TianQuan-Climate: A Subseasonal-to-Seasonal Global Weather Model via
Incorporate Climatology State | null | null | null | null | cs.LG | http://arxiv.org/licenses/nonexclusive-distrib/1.0/ | Subseasonal forecasting serves as an important support for Sustainable
Development Goals (SDGs), such as climate challenges, agricultural yield and
sustainable energy production. However, subseasonal forecasting is a complex
task in meteorology due to dissipating initial conditions and delayed external
forces. Although AI models are increasingly pushing the boundaries of this
forecasting limit, they face two major challenges: error accumulation and
Smoothness. To address these two challenges, we propose Climate Furnace
Subseasonal-to-Seasonal (TianQuan-Climate), a novel machine learning model
designed to provide global daily mean forecasts up to 45 days, covering five
upper-air atmospheric variables at 13 pressure levels and two surface
variables. Our proposed TianQuan-Climate has two advantages: 1) it utilizes a
multi-model prediction strategy to reduce system error impacts in long-term
subseasonal forecasts; 2) it incorporates a Content Fusion Module for
climatological integration and extends ViT with uncertainty blocks (UD-ViT) to
improve generalization by learning from uncertainty. We demonstrate the
effectiveness of TianQuan-Climate on benchmarks for weather forecasting and
climate projections within the 15 to 45-day range, where TianQuan-Climate
outperforms existing numerical and AI methods.
| [
{
"version": "v1",
"created": "Mon, 14 Apr 2025 07:02:34 GMT"
},
{
"version": "v2",
"created": "Sat, 19 Apr 2025 14:16:48 GMT"
},
{
"version": "v3",
"created": "Tue, 22 Apr 2025 02:29:49 GMT"
}
] | 2025-04-23T00:00:00 | [
[
"Li",
"Guowen",
""
],
[
"Liu",
"Xintong",
""
],
[
"Cao",
"Shilei",
""
],
[
"Liang",
"Haoyuan",
""
],
[
"Chen",
"Mengxuan",
""
],
[
"Zhang",
"Lixian",
""
],
[
"Zhang",
"Jinxiao",
""
],
[
"Wang",
"Jiuke",
""
],
[
"Jin",
"Meng",
""
],
[
"Zheng",
"Juepeng",
""
],
[
"Fu",
"Haohuan",
""
]
] |
2504.09944 | Alex Cowan | Alex Cowan | On the mean value of $\mathrm{GL}_1$ and $\mathrm{GL}_2$ $L$-functions,
with applications to murmurations | null | null | null | null | math.NT | http://arxiv.org/licenses/nonexclusive-distrib/1.0/ | We determine the mean value of $L$-functions attached to quadratic twists of
automorphic representations on $\mathrm{GL}_1$ or $\mathrm{GL}_2$ in large
regions of the critical strip. In the case of $\mathrm{GL}_1$, we go on to
exhibit a recently discovered type of fine structure called "murmurations"
unconditionally for all of our families. Our main tool is a new variant of the
approximate functional equation imbued with a mechanism for dynamically
rebalancing error terms while preserving holomorphicity.
| [
{
"version": "v1",
"created": "Mon, 14 Apr 2025 07:09:40 GMT"
},
{
"version": "v2",
"created": "Mon, 21 Apr 2025 20:56:55 GMT"
}
] | 2025-04-23T00:00:00 | [
[
"Cowan",
"Alex",
""
]
] |
2504.10027 | Yi Cui | Yi Cui, Rong Yu, and Weiqiang Yu | Deconfined Quantum Critical Point: A Review of Progress | Review article, to appear in Chinese Physics Letters | Chin. Phys. Lett. 42 047503 (2025) | 10.1088/0256-307X/42/4/047503 | null | cond-mat.str-el | http://creativecommons.org/licenses/by-nc-nd/4.0/ | Deconfined quantum critical points (DQCPs) have been proposed as a class of
continuous quantum phase transitions occurring between two ordered phases with
distinct symmetry-breaking patterns, beyond the conventional framework of
Landau-Ginzburg-Wilson (LGW) theory. At the DQCP, the system exhibits emergent
gauge fields, fractionalized excitations, and enhanced symmetries. Here we
review recent theoretical and experimental progress on exploring DQCPs in
condensed matter systems. We first introduce theoretical advancements in the
study of DQCPs over the past twenty years, particularly in magnetic models on
square lattices, honeycomb lattices, kagome lattices, and one-dimensional spin
chains. We then discuss recent progress on experimental realization of DQCP in
quantum magnetic systems. Experimentally, the Shastry-Sutherland model,
realized in SrCu$_2$(BO$_3$)$_2$, offers a particularly promising platform for
realizing DQCPs. The magnetic frustration inherent to this model drives phase
transitions between two distinct symmetry-breaking states: a valence bond solid
(VBS) phase and a N\'{e}el antiferromagnetic phase. Remarkably,
SrCu$_2$(BO$_3$)$_2$ has provided the first experimental evidence of a
proximate DQCP through a field-induced Bose-Einstein condensation,
transitioning from the VBS state to the N\'{e}el state. Nevertheless, the
direct experimental realization of a DQCP remains a significant challenge.
Despite this, it offers a promising platform for exploring emergent phenomena
through quantum phase transition in low-dimensional quantum systems.
| [
{
"version": "v1",
"created": "Mon, 14 Apr 2025 09:31:58 GMT"
}
] | 2025-04-23T00:00:00 | [
[
"Cui",
"Yi",
""
],
[
"Yu",
"Rong",
""
],
[
"Yu",
"Weiqiang",
""
]
] |
2504.10088 | Gyanendra Kumar Verma | Gyanendra K. Verma, Nupur Patanker and Abhay Kumar Singh | Code size constraints in b-symbol read channels: A bound analysis | null | null | null | null | cs.IT math.IT | http://creativecommons.org/licenses/by/4.0/ | In classical coding theory, error-correcting codes are designed to protect
against errors occurring at individual symbol positions in a codeword. However,
in practical storage and communication systems, errors often affect multiple
adjacent symbols rather than single symbols independently. To address this,
symbol-pair read channels were introduced \cite{Yuval2011}, and later
generalized to $b$-symbol read channels \cite{yaakobi2016} to better model such
error patterns. $b$-Symbol read channels generalize symbol-pair read channels
to account for clustered errors in modern storage and communication systems. By
developing bounds and efficient codes, researchers improve data reliability in
applications such as storage devices, wireless networks, and DNA-based storage.
Given integers $q$, $n$, $d$, and $b \geq 2$, let $A_b(n,d,q)$ denote the
largest possible code size for which there exists a $q$-ary code of length $n$
with minimum $b$-symbol distance at least $d$. In \cite{chen2022}, various
upper and lower bounds on $A_b(n,d,q)$ are given for $b=2$. In this paper, we
generalize some of these bounds to the $b$-symbol read channels for $b>2$ and
present several new bounds on $A_b(n,d,q)$. In particular, we establish the
linear programming bound, a recurrence relation on $A_b(n,d,q)$, the Johnson
bound (even), the restricted Johnson bound, the Gilbert-Varshamov-type bound,
and the Elias bound for the metric of symbols $b$, $b\geq 2$. Furthermore, we
provide examples demonstrating that the Gilbert-Varshamov bound we establish
offers a stronger lower bound than the one presented in \cite{Song2018}.
Additionally, we introduce an alternative approach to deriving the
Sphere-packing and Plotkin bounds.
| [
{
"version": "v1",
"created": "Mon, 14 Apr 2025 10:47:31 GMT"
},
{
"version": "v2",
"created": "Tue, 22 Apr 2025 08:44:22 GMT"
}
] | 2025-04-23T00:00:00 | [
[
"Verma",
"Gyanendra K.",
""
],
[
"Patanker",
"Nupur",
""
],
[
"Singh",
"Abhay Kumar",
""
]
] |
2504.10114 | Tatsuya Kaneko | Tatsuya Kaneko, Masataka Kakoi, Kazuhiko Kuroki | $t$-$J$ model for strongly correlated two-orbital systems: Application
to bilayer nickelate superconductors | 20 pages, 8 figures | null | null | null | cond-mat.str-el cond-mat.supr-con | http://arxiv.org/licenses/nonexclusive-distrib/1.0/ | We derive a $t$-$J$ model applicable to strongly correlated two-orbital
systems including bilayer nickelate superconductors. Using the Schrieffer-Wolff
transformation, we exclude the doubly occupied states raising the on-site
Coulomb energy and derive resulting spin interactions from the two-orbital
Hubbard model. We also introduce effective interactions attributed to the
interorbital Coulomb interaction. To adapt the effective model to bilayer
nickelates that exhibit high-temperature superconductivity, we quantitatively
evaluate the strengths of the spin interactions based on the hopping parameters
in La$_3$Ni$_2$O$_7$. Considering the evaluated effective interactions, we
propose a simplified $t$-$J$ model for bilayer nickelate superconductors.
| [
{
"version": "v1",
"created": "Mon, 14 Apr 2025 11:25:29 GMT"
},
{
"version": "v2",
"created": "Tue, 22 Apr 2025 06:36:22 GMT"
}
] | 2025-04-23T00:00:00 | [
[
"Kaneko",
"Tatsuya",
""
],
[
"Kakoi",
"Masataka",
""
],
[
"Kuroki",
"Kazuhiko",
""
]
] |
2504.10150 | Chen Zhang | Chen Zhang, Bo Hu, Weidong Chen, Zhendong Mao | HistLLM: A Unified Framework for LLM-Based Multimodal Recommendation
with User History Encoding and Compression | We want to withdraw this paper and revise its experimental details.
The revised version will be uploaded after further verification | null | null | null | cs.IR cs.MM | http://arxiv.org/licenses/nonexclusive-distrib/1.0/ | While large language models (LLMs) have proven effective in leveraging
textual data for recommendations, their application to multimodal
recommendation tasks remains relatively underexplored. Although LLMs can
process multimodal information through projection functions that map visual
features into their semantic space, recommendation tasks often require
representing users' history interactions through lengthy prompts combining text
and visual elements, which not only hampers training and inference efficiency
but also makes it difficult for the model to accurately capture user
preferences from complex and extended prompts, leading to reduced
recommendation performance. To address this challenge, we introduce HistLLM, an
innovative multimodal recommendation framework that integrates textual and
visual features through a User History Encoding Module (UHEM), compressing
multimodal user history interactions into a single token representation,
effectively facilitating LLMs in processing user preferences. Extensive
experiments demonstrate the effectiveness and efficiency of our proposed
mechanism.
| [
{
"version": "v1",
"created": "Mon, 14 Apr 2025 12:01:11 GMT"
},
{
"version": "v2",
"created": "Tue, 22 Apr 2025 02:41:47 GMT"
}
] | 2025-04-23T00:00:00 | [
[
"Zhang",
"Chen",
""
],
[
"Hu",
"Bo",
""
],
[
"Chen",
"Weidong",
""
],
[
"Mao",
"Zhendong",
""
]
] |
2504.10284 | Weiqi Wang Mr. | Weiqi Wang, Jiefu Ou, Yangqiu Song, Benjamin Van Durme, Daniel
Khashabi | Can LLMs Generate Tabular Summaries of Science Papers? Rethinking the
Evaluation Protocol | null | null | null | null | cs.CL | http://arxiv.org/licenses/nonexclusive-distrib/1.0/ | Literature review tables are essential for summarizing and comparing
collections of scientific papers. We explore the task of generating tables that
best fulfill a user's informational needs given a collection of scientific
papers. Building on recent work (Newman et al., 2024), we extend prior
approaches to address real-world complexities through a combination of
LLM-based methods and human annotations. Our contributions focus on three key
challenges encountered in real-world use: (i) User prompts are often
under-specified; (ii) Retrieved candidate papers frequently contain irrelevant
content; and (iii) Task evaluation should move beyond shallow text similarity
techniques and instead assess the utility of inferred tables for
information-seeking tasks (e.g., comparing papers). To support reproducible
evaluation, we introduce ARXIV2TABLE, a more realistic and challenging
benchmark for this task, along with a novel approach to improve literature
review table generation in real-world scenarios. Our extensive experiments on
this benchmark show that both open-weight and proprietary LLMs struggle with
the task, highlighting its difficulty and the need for further advancements.
Our dataset and code are available at https://github.com/JHU-CLSP/arXiv2Table.
| [
{
"version": "v1",
"created": "Mon, 14 Apr 2025 14:52:28 GMT"
},
{
"version": "v2",
"created": "Tue, 22 Apr 2025 03:09:56 GMT"
}
] | 2025-04-23T00:00:00 | [
[
"Wang",
"Weiqi",
""
],
[
"Ou",
"Jiefu",
""
],
[
"Song",
"Yangqiu",
""
],
[
"Van Durme",
"Benjamin",
""
],
[
"Khashabi",
"Daniel",
""
]
] |
2504.10305 | Fedor Vylegzhanin | Fedor Vylegzhanin and Yakov Veryovkin | Commutator subalgebra of the Lie algebra associated with a right-angled
Coxeter group | 16 pages. v2: Theorem 3.5 generalised; Section 6 rewritten, fixing
the proof of Theorem 7.2 | null | null | null | math.GR math.AT | http://creativecommons.org/licenses/by/4.0/ | We study the graded Lie algebra $L(RC_K)$ associated with the lower central
series of a right-angled Coxeter group $RC_K$. We prove that its commutator
subalgebra is a quotient of the polynomial ring over an auxiliary Lie
subalgebra $N_K$ of the graph Lie algebra $L_K$, and conjecture that the
quotient map is an isomorphism. The epimorphism is defined in terms of a new
operation in the associated Lie algebra, which corresponds to the squaring and
has an analogue in homotopy theory.
| [
{
"version": "v1",
"created": "Mon, 14 Apr 2025 15:13:49 GMT"
},
{
"version": "v2",
"created": "Tue, 22 Apr 2025 13:34:45 GMT"
}
] | 2025-04-23T00:00:00 | [
[
"Vylegzhanin",
"Fedor",
""
],
[
"Veryovkin",
"Yakov",
""
]
] |
2504.10601 | William Gyory | Ra\'ul A. Brice\~no, William Gyory, Thomas Iadecola, Srimoyee Sen | Simulating lattice fermion doubling with a Floquet drive | 6 pages, 1 figure, supplemental material (6 pages) | null | null | null | hep-lat cond-mat.mes-hall hep-th nucl-th quant-ph | http://creativecommons.org/licenses/by/4.0/ | We consider a recently discovered mathematical correspondence between the
spectra of a naively discretized lattice fermion and that of a periodically
driven (i.e., Floquet) quantum system and turn it into an infrared equivalence
between the two systems. The equivalence can be framed as a duality relation,
allowing us to simulate a two-flavor discrete-time fermion theory on the
lattice side, where the two flavors arise from time discretization, using a
single-flavor fermion theory on the Floquet side. Our demonstration establishes
an equivalence between (i) the fermion content, (ii) the correlation functions,
and consequently (iii) observables of the two theories in the infrared. We also
show how interactions may be incorporated into this equivalence.
| [
{
"version": "v1",
"created": "Mon, 14 Apr 2025 18:01:40 GMT"
},
{
"version": "v2",
"created": "Tue, 22 Apr 2025 01:39:56 GMT"
}
] | 2025-04-23T00:00:00 | [
[
"Briceño",
"Raúl A.",
""
],
[
"Gyory",
"William",
""
],
[
"Iadecola",
"Thomas",
""
],
[
"Sen",
"Srimoyee",
""
]
] |
2504.10621 | Chase Hatcher | Chase Hatcher, Lily Donis, Adrienne Traxler, Madison Swirtz, Camila
Manni, Justin Gutzwa, Charles Henderson, Ram\'on Barthelemy | Egocentric Mixed-Methods SNA: Analyzing Interviews with Women and/or
Queer and LGBT+ Ph.D. Physicists | null | null | null | null | physics.ed-ph | http://creativecommons.org/licenses/by/4.0/ | Social network analysis (SNA) has been widely used in physics education
research (PER) in recent years, but mostly in a limited range of the available
modalities. This paper describes a unique approach to egocentric, mixed-methods
SNA applied to qualitative network data obtained from 100 interviews with women
and/or queer professional physicists. We focus on our methods for obtaining
quantitative network data from these qualitative sources and present novel
techniques for analysis of the networks. We also examine the ways in which
egocentric and mixed-methods SNA techniques are aligned with critical methods
and well-suited to the study of difference, non-normativity, and experiences of
marginalization in physics spaces and communities. We explore the limitations
and potential applications of these methods and situate this work in the larger
context of our study of these interviews. This work bridges a methodological
gap between SNA and qualitative work on identity in PER and begins to develop
our understanding of the way gender and sexual minority physicists experience
support.
| [
{
"version": "v1",
"created": "Mon, 14 Apr 2025 18:21:30 GMT"
},
{
"version": "v2",
"created": "Tue, 22 Apr 2025 17:02:01 GMT"
}
] | 2025-04-23T00:00:00 | [
[
"Hatcher",
"Chase",
""
],
[
"Donis",
"Lily",
""
],
[
"Traxler",
"Adrienne",
""
],
[
"Swirtz",
"Madison",
""
],
[
"Manni",
"Camila",
""
],
[
"Gutzwa",
"Justin",
""
],
[
"Henderson",
"Charles",
""
],
[
"Barthelemy",
"Ramón",
""
]
] |
2504.10852 | Xuesong Li | Pengxiao Han, Changkun Ye, Jinguang Tong, Cuicui Jiang, Jie Hong, Li
Fang, Xuesong Li | Enhancing Features in Long-tailed Data Using Large Vision Model | null | null | null | null | cs.CV | http://arxiv.org/licenses/nonexclusive-distrib/1.0/ | Language-based foundation models, such as large language models (LLMs) or
large vision-language models (LVLMs), have been widely studied in long-tailed
recognition. However, the need for linguistic data is not applicable to all
practical tasks. In this study, we aim to explore using large vision models
(LVMs) or visual foundation models (VFMs) to enhance long-tailed data features
without any language information. Specifically, we extract features from the
LVM and fuse them with features in the baseline network's map and latent space
to obtain the augmented features. Moreover, we design several prototype-based
losses in the latent space to further exploit the potential of the augmented
features. In the experimental section, we validate our approach on two
benchmark datasets: ImageNet-LT and iNaturalist2018.
| [
{
"version": "v1",
"created": "Tue, 15 Apr 2025 04:21:50 GMT"
},
{
"version": "v2",
"created": "Tue, 22 Apr 2025 12:31:32 GMT"
}
] | 2025-04-23T00:00:00 | [
[
"Han",
"Pengxiao",
""
],
[
"Ye",
"Changkun",
""
],
[
"Tong",
"Jinguang",
""
],
[
"Jiang",
"Cuicui",
""
],
[
"Hong",
"Jie",
""
],
[
"Fang",
"Li",
""
],
[
"Li",
"Xuesong",
""
]
] |
2504.10870 | Apurva Tiwari | Apurva Tiwari, Jason Iaconis, Jezer Jojo, Sayonee Ray, Martin
Roetteler, Chris Hill, Jay Pathak | Algorithmic Advances Towards a Realizable Quantum Lattice Boltzmann
Method | 11 pages, 11 figures | null | null | null | quant-ph cs.ET physics.comp-ph | http://creativecommons.org/licenses/by-nc-nd/4.0/ | The Quantum Lattice Boltzmann Method (QLBM) is one of the most promising
approaches for realizing the potential of quantum computing in simulating
computational fluid dynamics. Many recent works mostly focus on classical
simulation, and rely on full state tomography. Several key algorithmic issues
like observable readout, data encoding, and impractical circuit depth remain
unsolved. As a result, these are not directly realizable on any quantum
hardware. We present a series of novel algorithmic advances which allow us to
implement the QLBM algorithm, for the first time, on a quantum computer.
Hardware results for the time evolution of a 2D Gaussian initial density
distribution subject to a uniform advection-diffusion field are presented.
Furthermore, 3D simulation results are presented for particular non-uniform
advection fields, devised so as to avoid the problem of diminishing probability
of success due to repeated post-selection operations required for multiple
timesteps. We demonstrate the evolution of an initial quantum state governed by
the advection-diffusion equation, accounting for the iterative nature of the
explicit QLBM algorithm. A tensor network encoding scheme is used to represent
the initial condition supplied to the advection-diffusion equation,
significantly reducing the two-qubit gate count affording a shorter circuit
depth. Further reductions are made in the collision and streaming operators.
Collectively, these advances give a path to realizing more practical, 2D and 3D
QLBM applications with non-trivial velocity fields on quantum hardware.
| [
{
"version": "v1",
"created": "Tue, 15 Apr 2025 05:02:41 GMT"
}
] | 2025-04-23T00:00:00 | [
[
"Tiwari",
"Apurva",
""
],
[
"Iaconis",
"Jason",
""
],
[
"Jojo",
"Jezer",
""
],
[
"Ray",
"Sayonee",
""
],
[
"Roetteler",
"Martin",
""
],
[
"Hill",
"Chris",
""
],
[
"Pathak",
"Jay",
""
]
] |
2504.10897 | Prashanti Priya Angara | Prashanti Priya Angara, Emily Martins, Ulrike Stege, Hausi M\"uller | SCOOP: A Quantum-Computing Framework for Constrained Combinatorial
Optimization | 11 pages,4 figures, 1 table | null | null | null | quant-ph | http://creativecommons.org/licenses/by/4.0/ | While the ultimate goal of solving computationally intractable problems is to
find a provably optimal solutions, practical constraints of real-world
scenarios often necessitate focusing on efficiently obtaining high-quality,
near-optimal solutions. The Quantum Approximate Optimization Algorithm (QAOA)
is a state-of-the-art hybrid quantum-classical approach for tackling these
challenging problems that are encoded using quadratic and higher-order
unconstrained binary optimization problems (QUBO and HUBO). We present SCOOP, a
novel QAOA-based framework for solving constrained optimization problems. SCOOP
transforms a constrained problem into an unconstrained counterpart, forming
SCOOP problem twins. The QAOA quantum algorithm operates on the unconstrained
twin to identify potential optimal and near-optimal solutions. Effective
classical post-processing reduces the solution set to the constrained problem
space. Our SCOOP approach is solution-enhanced, objective-function-compatible,
and scalable. We demonstrate the framework on three NP-hard problems, Minimum
Dominating Set, Minimum Maximal Matching, and Minimum Set Cover appearing in
practical application domains such as resource allocation, communication
networks, and machine learning. We validate SCOOP's feasibility and
effectiveness on Xanadu PennyLane simulators.
| [
{
"version": "v1",
"created": "Tue, 15 Apr 2025 06:17:23 GMT"
},
{
"version": "v2",
"created": "Mon, 21 Apr 2025 23:41:31 GMT"
}
] | 2025-04-23T00:00:00 | [
[
"Angara",
"Prashanti Priya",
""
],
[
"Martins",
"Emily",
""
],
[
"Stege",
"Ulrike",
""
],
[
"Müller",
"Hausi",
""
]
] |
2504.10915 | Rajesh Ranjan | Rajesh Ranjan, Shailja Gupta, Surya Narayan Singh | LOKA Protocol: A Decentralized Framework for Trustworthy and Ethical AI
Agent Ecosystems | 4 Figures, 1 Table | null | null | null | cs.MA cs.AI cs.CY | http://creativecommons.org/licenses/by-nc-sa/4.0/ | The rise of autonomous AI agents, capable of perceiving, reasoning, and
acting independently, signals a profound shift in how digital ecosystems
operate, govern, and evolve. As these agents proliferate beyond centralized
infrastructures, they expose foundational gaps in identity, accountability, and
ethical alignment. Three critical questions emerge: Identity: Who or what is
the agent? Accountability: Can its actions be verified, audited, and trusted?
Ethical Consensus: Can autonomous systems reliably align with human values and
prevent harmful emergent behaviors? We present the novel LOKA Protocol (Layered
Orchestration for Knowledgeful Agents), a unified, systems-level architecture
for building ethically governed, interoperable AI agent ecosystems. LOKA
introduces a proposed Universal Agent Identity Layer (UAIL) for decentralized,
verifiable identity; intent-centric communication protocols for semantic
coordination across diverse agents; and a Decentralized Ethical Consensus
Protocol (DECP) that could enable agents to make context-aware decisions
grounded in shared ethical baselines. Anchored in emerging standards such as
Decentralized Identifiers (DIDs), Verifiable Credentials (VCs), and
post-quantum cryptography, LOKA proposes a scalable, future-resilient blueprint
for multi-agent AI governance. By embedding identity, trust, and ethics into
the protocol layer itself, LOKA proposes the foundation for a new era of
responsible, transparent, and autonomous AI ecosystems operating across digital
and physical domains.
| [
{
"version": "v1",
"created": "Tue, 15 Apr 2025 06:51:35 GMT"
},
{
"version": "v2",
"created": "Tue, 22 Apr 2025 02:59:48 GMT"
}
] | 2025-04-23T00:00:00 | [
[
"Ranjan",
"Rajesh",
""
],
[
"Gupta",
"Shailja",
""
],
[
"Singh",
"Surya Narayan",
""
]
] |
2504.10972 | Yihang Liu | Yihang Liu, Lianghua He, Ying Wen, Longzhen Yang, Hongzhou Chen | AFiRe: Anatomy-Driven Self-Supervised Learning for Fine-Grained
Representation in Radiographic Images | null | null | 10.1609/aaai.v39i18.34091 | null | cs.CV | http://creativecommons.org/licenses/by/4.0/ | Current self-supervised methods, such as contrastive learning, predominantly
focus on global discrimination, neglecting the critical fine-grained anatomical
details required for accurate radiographic analysis. To address this challenge,
we propose an Anatomy-driven self-supervised framework for enhancing
Fine-grained Representation in radiographic image analysis (AFiRe). The core
idea of AFiRe is to align the anatomical consistency with the unique
token-processing characteristics of Vision Transformer. Specifically, AFiRe
synergistically performs two self-supervised schemes: (i) Token-wise
anatomy-guided contrastive learning, which aligns image tokens based on
structural and categorical consistency, thereby enhancing fine-grained
spatial-anatomical discrimination; (ii) Pixel-level anomaly-removal
restoration, which particularly focuses on local anomalies, thereby refining
the learned discrimination with detailed geometrical information. Additionally,
we propose Synthetic Lesion Mask to enhance anatomical diversity while
preserving intra-consistency, which is typically corrupted by traditional data
augmentations, such as Cropping and Affine transformations. Experimental
results show that AFiRe: (i) provides robust anatomical discrimination,
achieving more cohesive feature clusters compared to state-of-the-art
contrastive learning methods; (ii) demonstrates superior generalization,
surpassing 7 radiography-specific self-supervised methods in multi-label
classification tasks with limited labeling; and (iii) integrates fine-grained
information, enabling precise anomaly detection using only image-level
annotations.
| [
{
"version": "v1",
"created": "Tue, 15 Apr 2025 08:29:54 GMT"
},
{
"version": "v2",
"created": "Tue, 22 Apr 2025 08:38:33 GMT"
}
] | 2025-04-23T00:00:00 | [
[
"Liu",
"Yihang",
""
],
[
"He",
"Lianghua",
""
],
[
"Wen",
"Ying",
""
],
[
"Yang",
"Longzhen",
""
],
[
"Chen",
"Hongzhou",
""
]
] |
2504.11035 | Hannah Correia | Hannah E. Correia | A conceptual synthesis of causal assumptions for causal discovery and
inference | Withdrawn for incorporation as supplementary material in a broader
collaborative manuscript; integration will be reflected in a new preprint | null | null | null | stat.ME q-bio.QM stat.AP stat.OT | http://creativecommons.org/licenses/by-nc-nd/4.0/ | This work presents a conceptual synthesis of causal discovery and inference
frameworks, with a focus on how foundational assumptions -- causal sufficiency,
causal faithfulness, and the causal Markov condition -- are formalized and
operationalized across methodological traditions. Through structured tables and
comparative summaries, I map core assumptions, tasks, and analytical choices
from multiple causal frameworks, highlighting their connections and
differences. The synthesis provides practical guidance for researchers
designing causal studies, especially in settings where observational or
experimental constraints challenge standard approaches. This guide spans all
phases of causal analysis, including question formulation, formalization of
background knowledge, selection of appropriate frameworks, choice of study
design or algorithm, and interpretation. It is intended as a tool to support
rigorous causal reasoning across diverse empirical domains.
| [
{
"version": "v1",
"created": "Tue, 15 Apr 2025 09:58:06 GMT"
},
{
"version": "v2",
"created": "Mon, 21 Apr 2025 18:55:20 GMT"
}
] | 2025-04-23T00:00:00 | [
[
"Correia",
"Hannah E.",
""
]
] |
2504.11086 | Fernando M\'ario de Oliveira Filho | Christine Bachoc, Bram Bekker, Philippe Moustrou, Fernando M\'ario de
Oliveira Filho | Obtuse almost-equiangular sets | 29 pages; fixed problem with references from previous version | null | null | null | math.CO math.MG math.OC | http://arxiv.org/licenses/nonexclusive-distrib/1.0/ | For $t \in [-1, 1)$, a set of points on the $(n-1)$-dimensional unit sphere
is called $t$-almost equiangular if among any three distinct points there is a
pair with inner product $t$. We propose a semidefinite programming upper bound
for the maximum cardinality $\alpha(n, t)$ of such a set based on an extension
of the Lov\'asz theta number to hypergraphs. This bound is at least as good as
previously known bounds and for many values of $n$ and $t$ it is better.
We also refine existing spectral methods to show that $\alpha(n, t) \leq
2(n+1)$ for all $n$ and $t \leq 0$, with equality only at $t = -1/n$. This
allows us to show the uniqueness of the optimal construction at $t = -1/n$ for
$n \leq 5$ and to enumerate all possible constructions for $n \leq 3$ and $t
\leq 0$.
| [
{
"version": "v1",
"created": "Tue, 15 Apr 2025 11:32:46 GMT"
},
{
"version": "v2",
"created": "Tue, 22 Apr 2025 11:21:59 GMT"
}
] | 2025-04-23T00:00:00 | [
[
"Bachoc",
"Christine",
""
],
[
"Bekker",
"Bram",
""
],
[
"Moustrou",
"Philippe",
""
],
[
"Filho",
"Fernando Mário de Oliveira",
""
]
] |
2504.11116 | Jeonggyu Huh | Jeonggyu Huh, Jaegi Jeon, Hyeng Keun Koo | Breaking the Dimensional Barrier: A Pontryagin-Guided Direct Policy
Optimization for Continuous-Time Multi-Asset Portfolio | null | null | null | null | q-fin.PM q-fin.CP | http://arxiv.org/licenses/nonexclusive-distrib/1.0/ | Solving large-scale, continuous-time portfolio optimization problems
involving numerous assets and state-dependent dynamics has long been challenged
by the curse of dimensionality. Traditional dynamic programming and PDE-based
methods, while rigorous, typically become computationally intractable beyond
few state variables (3~6 limit in prior studies). To overcome this critical
barrier, we introduce the Pontryagin-Guided Direct Policy Optimization (PG-DPO)
framework. PG-DPO leverages Pontryagin's Maximum Principle (PMP) and
backpropagation-through-time (BPTT) to guide neural network policies, handling
exogenous states without dense grids. This PMP-guided approach holds potential
for a broad class of sufficiently regular continuous-time control problems.
Crucially, our computationally efficient "Two-Stage" variant exploits rapidly
stabilizing BPTT costate estimates, converting them into near-optimal
Pontryagin controls after only a short warm-up, significantly reducing training
overhead. This enables a breakthrough in scalability: numerical experiments
show PG-DPO successfully tackles problems with dimensions previously considered
far out of reach (up to 50 assets and 10 state variables). The framework
delivers near-optimal policies, offering a practical and powerful alternative
for high-dimensional continuous-time portfolio choice.
| [
{
"version": "v1",
"created": "Tue, 15 Apr 2025 12:03:14 GMT"
},
{
"version": "v2",
"created": "Tue, 22 Apr 2025 10:59:52 GMT"
}
] | 2025-04-23T00:00:00 | [
[
"Huh",
"Jeonggyu",
""
],
[
"Jeon",
"Jaegi",
""
],
[
"Koo",
"Hyeng Keun",
""
]
] |
2504.11392 | Mohamed ElKabbash | Mohamed ElKabbash | Probing General Relativity-Induced Decoherence Using an on-chip Sagnac
Interferometer | null | null | null | null | physics.optics quant-ph | http://creativecommons.org/licenses/by/4.0/ | The intersection of quantum mechanics and general relativity remains an open
frontier in fundamental physics, with few experimentally accessible phenomena
connecting the two. Recent theoretical proposals suggest that relativistic
proper time can act as a source of decoherence in quantum systems, providing a
testable overlap between the two theories. Here, we propose a chip-integrated
Sagnac interferometer where rotation induces a proper time difference between
clockwise and counterclockwise single-photon paths. When this time delay
exceeds the photon's coherence time, interference visibility is predicted to
decrease, offering a direct signature of relativistic time dilation-induced
decoherence. We theoretically derive the proper time difference arising from
the Sagnac effect and estimate that for a loop radius of 18.9 cm and a rotation
speed of 1000 rad/s, decoherence should occur for single-photon wavepackets
with a coherence time of 10 femtoseconds. We also present a practical chip
design that accommodates the required high-speed mechanical rotation and
includes an all-optical readout scheme to eliminate wiring constraints. This
approach enables a stable, on-chip implementation using realistic parameters,
with rotation speed serving as a continuously tunable knob to control
decoherence. Our platform opens a new route for experimentally probing the
interplay between quantum coherence and relativistic proper time in a scalable
and compact form.
| [
{
"version": "v1",
"created": "Tue, 15 Apr 2025 17:01:16 GMT"
},
{
"version": "v2",
"created": "Tue, 22 Apr 2025 05:17:04 GMT"
}
] | 2025-04-23T00:00:00 | [
[
"ElKabbash",
"Mohamed",
""
]
] |
2504.11403 | Qiutong Wang | Qiutong Wang | Counting irreducible representations of general linear groups and
unitary groups | null | null | null | null | math.RT | http://arxiv.org/licenses/nonexclusive-distrib/1.0/ | Let $G$ be a general linear group over $\BR$, $\BC$, or $\BH$, or a real
unitary group. In this paper, we precisely describe the number of isomorphism
classes of irreducible Casselman-Wallach representations of $G$ with a given
infinitesimal character and a given associated variety, expressed in terms of
certain combinatorial data called painted Young diagrams and assigned Young
diagrams.
| [
{
"version": "v1",
"created": "Tue, 15 Apr 2025 17:18:20 GMT"
},
{
"version": "v2",
"created": "Tue, 22 Apr 2025 10:32:56 GMT"
}
] | 2025-04-23T00:00:00 | [
[
"Wang",
"Qiutong",
""
]
] |
2504.11445 | Hengkai Ding | Hengkai Ding, Yiping Shu, Yun Chen, Nan Li, Xiaoyue Cao, James
Nightingale, Qiuhan He, Lei Wang and Hui Li | Pixel-level modelling of group-scale strong lens CASSOWARY 19 | 25 pages, 9 figures, 3 tables, accepted for publication in RAA | null | null | null | astro-ph.GA | http://arxiv.org/licenses/nonexclusive-distrib/1.0/ | We present the first high-precision model for the group-scale strong lensing
system CASSOWARY 19 (CSWA19), utilising images from the Hubble Space Telescope
(HST). Sixteen member galaxies identified via the red-sequence method, and the
main halo, all modelled as the dual Pseudo Isothermal Elliptical profile
(dPIE), are incorporated into a parametric lens model alongside an external
shear field. To model the system, we adopt the PyAutoLens software package,
employing a progressive search chain strategy for realizing the transition of
source model from multiple S\'ersic profiles to a brightness-adaptive
pixelization, which uses 1000 pixels in the source plane to reconstruct the
background source corresponding to 177,144 image pixels in the image plane. Our
results indicate that the total mass within the Einstein radius is
$M_{\theta_\mathrm{E}}$ $\approx 1.41\times10^{13}$M$_{\odot}$ and the average
slope of the total mass density $\rho (r)\propto r^{-\gamma}$ is
$\tilde{\gamma}=1.33$ within the effective radius. This slope is shallower than
those measured in galaxies and groups but is closer to those of galaxy
clusters. In addition, our approach successfully resolves the two merging
galaxies in the background source and yields a total magnification of
$\mu=103.18^{+0.23}_{-0.19}$, which is significantly higher than the outcomes
from previous studies of CSWA19. In summary, our research demonstrates the
effectiveness of the brightness-adaptive pixelization source reconstruction
technique for modelling group-scale strong lensing systems. It can serve as a
technical reference for future investigations into pixel-level modelling of the
group- and cluster-scale strong lensing systems.
| [
{
"version": "v1",
"created": "Tue, 15 Apr 2025 17:56:11 GMT"
},
{
"version": "v2",
"created": "Wed, 16 Apr 2025 12:26:12 GMT"
},
{
"version": "v3",
"created": "Tue, 22 Apr 2025 10:28:07 GMT"
}
] | 2025-04-23T00:00:00 | [
[
"Ding",
"Hengkai",
""
],
[
"Shu",
"Yiping",
""
],
[
"Chen",
"Yun",
""
],
[
"Li",
"Nan",
""
],
[
"Cao",
"Xiaoyue",
""
],
[
"Nightingale",
"James",
""
],
[
"He",
"Qiuhan",
""
],
[
"Wang",
"Lei",
""
],
[
"Li",
"Hui",
""
]
] |
2504.11676 | Wenshuai Hu | Wenshuai Hu and Guanghua Ji | Maximum bound principle for Q-tensor gradient flow with low regularity
integrators | 33 pages,60 figures | null | null | null | math.NA cs.NA | http://arxiv.org/licenses/nonexclusive-distrib/1.0/ | We investigate low-regularity integrator (LRI) methods for the Q-tensor model
governing nematic liquid-crystalline semilinear parabolic equation. First- and
second-order temporal discretizations are developed using Duhamel's formula,
and we rigorously prove that both schemes preserve the maximum bound principle
(MBP) and energy dissipation under minimal regularity requirements. Optimal
convergence rates are established for the proposed methods. Numerical
experiments validate the theoretical findings, demonstrating that the
eigenvalues of Q remain strictly confined within the physical range
(-1/3},2/3).
| [
{
"version": "v1",
"created": "Wed, 16 Apr 2025 00:22:05 GMT"
},
{
"version": "v2",
"created": "Tue, 22 Apr 2025 04:52:04 GMT"
}
] | 2025-04-23T00:00:00 | [
[
"Hu",
"Wenshuai",
""
],
[
"Ji",
"Guanghua",
""
]
] |
2504.11759 | Ziyu Xu | Ziyu Xu and Lasse Fischer and Aaditya Ramdas | Bringing closure to FDR control: beating the e-Benjamini-Hochberg
procedure | 18 pages, 1 figure | null | null | null | stat.ME math.ST stat.TH | http://arxiv.org/licenses/nonexclusive-distrib/1.0/ | False discovery rate (FDR) has been a key metric for error control in
multiple hypothesis testing, and many methods have developed for FDR control
across a diverse cross-section of settings and applications. We develop a
closure principle for all FDR controlling procedures, i.e., we provide a
characterization based on e-values for all admissible FDR controlling
procedures. A general version of this closure principle can recover any
multiple testing error metric and allows one to choose the error metric
post-hoc. We leverage this idea to formulate the closed eBH procedure, a
(usually strict) improvement over the eBH procedure for FDR control when
provided with e-values. This also yields a closed BY procedure that dominates
the Benjamini-Yekutieli (BY) procedure for FDR control with arbitrarily
dependent p-values, thus proving that the latter is inadmissibile. We
demonstrate the practical performance of our new procedures in simulations.
| [
{
"version": "v1",
"created": "Wed, 16 Apr 2025 04:36:12 GMT"
},
{
"version": "v2",
"created": "Tue, 22 Apr 2025 04:56:11 GMT"
}
] | 2025-04-23T00:00:00 | [
[
"Xu",
"Ziyu",
""
],
[
"Fischer",
"Lasse",
""
],
[
"Ramdas",
"Aaditya",
""
]
] |
2504.11867 | Xingyu Li | Qi Liu, Yanchen Liu, Ruifeng Li, Chenhong Cao, Yufeng Li, Xingyu Li,
Peng Wang, Runhan Feng, Shiyang Bu | MDHP-Net: Detecting an Emerging Time-exciting Threat in IVN | This work was intended as a replacement of arXiv:2411.10258 and any
subsequent updates will appear there | null | null | null | cs.CR | http://arxiv.org/licenses/nonexclusive-distrib/1.0/ | The integration of intelligent and connected technologies in modern vehicles,
while offering enhanced functionalities through Electronic Control Unit (ECU)
and interfaces like OBD-II and telematics, also exposes the vehicle's
in-vehicle network (IVN) to potential cyberattacks. Unlike prior work, we
identify a new time-exciting threat model against IVN. These attacks inject
malicious messages that exhibit a time-exciting effect, gradually manipulating
network traffic to disrupt vehicle operations and compromise safety-critical
functions. We systematically analyze the characteristics of the threat:
dynamism, time-exciting impact, and low prior knowledge dependency. To validate
its practicality, we replicate the attack on a real Advanced Driver Assistance
System via Controller Area Network (CAN), exploiting Unified Diagnostic Service
vulnerabilities and proposing four attack strategies. While CAN's integrity
checks mitigate attacks, Ethernet migration (e.g., DoIP/SOME/IP) introduces new
surfaces. We further investigate the feasibility of time-exciting threat under
SOME/IP. To detect time-exciting threat, we introduce MDHP-Net, leveraging
Multi-Dimentional Hawkes Process (MDHP) and temporal and message-wise feature
extracting structures. Meanwhile, to estimate MDHP parameters, we developed the
first GPU-optimized gradient descent solver for MDHP (MDHP-GDS). These modules
significantly improves the detection rate under time-exciting attacks in
multi-ECU IVN system. To address data scarcity, we release STEIA9, the first
open-source dataset for time-exciting attacks, covering 9 Ethernet-based attack
scenarios. Extensive experiments on STEIA9 (9 attack scenarios) show MDHP-Net
outperforms 3 baselines, confirming attack feasibility and detection efficacy.
| [
{
"version": "v1",
"created": "Wed, 16 Apr 2025 08:41:24 GMT"
},
{
"version": "v2",
"created": "Tue, 22 Apr 2025 01:20:29 GMT"
}
] | 2025-04-23T00:00:00 | [
[
"Liu",
"Qi",
""
],
[
"Liu",
"Yanchen",
""
],
[
"Li",
"Ruifeng",
""
],
[
"Cao",
"Chenhong",
""
],
[
"Li",
"Yufeng",
""
],
[
"Li",
"Xingyu",
""
],
[
"Wang",
"Peng",
""
],
[
"Feng",
"Runhan",
""
],
[
"Bu",
"Shiyang",
""
]
] |
2504.11907 | Gabriele Calzolari | Gabriele Calzolari (1), Vidya Sumathy (1), Christoforos Kanellakis
(1), George Nikolakopoulos (1) ((1) Lule{\aa} University of Technology) | A Graph-Based Reinforcement Learning Approach with Frontier Potential
Based Reward for Safe Cluttered Environment Exploration | 6 pages, 4 figures, submitted to the 2025 IEEE/RSJ International
Conference on Intelligent Robots and Systems (IROS 2025) | null | null | null | cs.RO | http://arxiv.org/licenses/nonexclusive-distrib/1.0/ | Autonomous exploration of cluttered environments requires efficient
exploration strategies that guarantee safety against potential collisions with
unknown random obstacles. This paper presents a novel approach combining a
graph neural network-based exploration greedy policy with a safety shield to
ensure safe navigation goal selection. The network is trained using
reinforcement learning and the proximal policy optimization algorithm to
maximize exploration efficiency while reducing the safety shield interventions.
However, if the policy selects an infeasible action, the safety shield
intervenes to choose the best feasible alternative, ensuring system
consistency. Moreover, this paper proposes a reward function that includes a
potential field based on the agent's proximity to unexplored regions and the
expected information gain from reaching them. Overall, the approach
investigated in this paper merges the benefits of the adaptability of
reinforcement learning-driven exploration policies and the guarantee ensured by
explicit safety mechanisms. Extensive evaluations in simulated environments
demonstrate that the approach enables efficient and safe exploration in
cluttered environments.
| [
{
"version": "v1",
"created": "Wed, 16 Apr 2025 09:31:14 GMT"
},
{
"version": "v2",
"created": "Tue, 22 Apr 2025 12:15:25 GMT"
}
] | 2025-04-23T00:00:00 | [
[
"Calzolari",
"Gabriele",
"",
"Luleå University of Technology"
],
[
"Sumathy",
"Vidya",
"",
"Luleå University of Technology"
],
[
"Kanellakis",
"Christoforos",
"",
"Luleå University of Technology"
],
[
"Nikolakopoulos",
"George",
"",
"Luleå University of Technology"
]
] |
2504.11932 | Rintaro Karashima | Rintaro Karashima and Hiroyasu Inoue | Technological Complexity Based on Japanese Patent Data | null | null | null | null | cs.SI | http://creativecommons.org/licenses/by/4.0/ | As international competition intensifies in technologies, nations need to
identify key technologies to foster innovation. However, the identification is
difficult because a technology is independent, therefore has complex nature.
Here, this study aims to assess patent technological fields by applying
Technological Complexity Index from a corporate perspective, addressing its
underutilization in Japan despite its potential. By utilizing carefully
processed patent data from fiscal years 1981 to 2010, we analyze the bipartite
network which consists of 1,938 corporations and 35 or 124 technological
fields. Our findings provide quantitative characteristics of ubiquity and
sophistication for patent fields, the detailed technological trends that
reflect the social context, and methodological stability for policymakers and
researchers, contributing to targeted innovation strategies in Japan.
| [
{
"version": "v1",
"created": "Wed, 16 Apr 2025 10:12:00 GMT"
},
{
"version": "v2",
"created": "Tue, 22 Apr 2025 05:11:47 GMT"
}
] | 2025-04-23T00:00:00 | [
[
"Karashima",
"Rintaro",
""
],
[
"Inoue",
"Hiroyasu",
""
]
] |
2504.11972 | Xanh Ho Thi | Xanh Ho, Jiahao Huang, Florian Boudin, and Akiko Aizawa | LLM-as-a-Judge: Reassessing the Performance of LLMs in Extractive QA | 17 pages; code and data are available at
https://github.com/Alab-NII/llm-judge-extract-qa | null | null | null | cs.CL | http://creativecommons.org/licenses/by/4.0/ | Extractive reading comprehension question answering (QA) datasets are
typically evaluated using Exact Match (EM) and F1-score, but these metrics
often fail to fully capture model performance. With the success of large
language models (LLMs), they have been employed in various tasks, including
serving as judges (LLM-as-a-judge). In this paper, we reassess the performance
of QA models using LLM-as-a-judge across four reading comprehension QA
datasets. We examine different families of LLMs and various answer types to
evaluate the effectiveness of LLM-as-a-judge in these tasks. Our results show
that LLM-as-a-judge is highly correlated with human judgments and can replace
traditional EM/F1 metrics. By using LLM-as-a-judge, the correlation with human
judgments improves significantly, from 0.22 (EM) and 0.40 (F1-score) to 0.85.
These findings confirm that EM and F1 metrics underestimate the true
performance of the QA models. While LLM-as-a-judge is not perfect for more
difficult answer types (e.g., job), it still outperforms EM/F1, and we observe
no bias issues, such as self-preference, when the same model is used for both
the QA and judgment tasks.
| [
{
"version": "v1",
"created": "Wed, 16 Apr 2025 11:08:46 GMT"
},
{
"version": "v2",
"created": "Tue, 22 Apr 2025 05:04:40 GMT"
}
] | 2025-04-23T00:00:00 | [
[
"Ho",
"Xanh",
""
],
[
"Huang",
"Jiahao",
""
],
[
"Boudin",
"Florian",
""
],
[
"Aizawa",
"Akiko",
""
]
] |
2504.12060 | David Rasmussen Lolck | Nairen Cao, Vincent Cohen-Addad, Euiwoong Lee, Shi Li, David Rasmussen
Lolck, Alantha Newman, Mikkel Thorup, Lukas Vogl, Shuyi Yan, Hanwen Zhang | Static to Dynamic Correlation Clustering | null | null | null | null | cs.DS | http://creativecommons.org/licenses/by/4.0/ | Correlation clustering is a well-studied problem, first proposed by Bansal,
Blum, and Chawla [BBC04]. The input is an unweighted, undirected graph. The
problem is to cluster the vertices so as to minimizing the number of edges
between vertices in different clusters and missing edges between vertices
inside the same cluster. This problem has a wide application in data mining and
machine learning. We introduce a general framework that transforms existing
static correlation clustering algorithms into fully-dynamic ones that work
against an adaptive adversary.
We show how to apply our framework to known efficient correlation clustering
algorithms, starting from the classic $3$-approximate Pivot algorithm from
[ACN08]. Applied to the most recent near-linear $1.437$-approximation algorithm
from [CCL+25], we get a $1.437$-approximation fully-dynamic algorithm that
works with worst-case constant update time. The original static algorithm gets
its approximation factor with constant probability, and we get the same against
an adaptive adversary in the sense that for any given update step not known to
our algorithm, our solution is a $1.437$-approximation with constant
probability when we reach this update.
Previous dynamic algorithms had approximation factors around $3$ in
expectation, and they could only handle an oblivious adversary.
| [
{
"version": "v1",
"created": "Wed, 16 Apr 2025 13:13:16 GMT"
},
{
"version": "v2",
"created": "Tue, 22 Apr 2025 12:54:09 GMT"
}
] | 2025-04-23T00:00:00 | [
[
"Cao",
"Nairen",
""
],
[
"Cohen-Addad",
"Vincent",
""
],
[
"Lee",
"Euiwoong",
""
],
[
"Li",
"Shi",
""
],
[
"Lolck",
"David Rasmussen",
""
],
[
"Newman",
"Alantha",
""
],
[
"Thorup",
"Mikkel",
""
],
[
"Vogl",
"Lukas",
""
],
[
"Yan",
"Shuyi",
""
],
[
"Zhang",
"Hanwen",
""
]
] |
2504.12157 | Xiaojun Ye | Xiaojun Ye, Chun Wang, Yiren Song, Sheng Zhou, Liangcheng Li, Jiajun
Bu | FocusedAD: Character-centric Movie Audio Description | Code and Demo link: https://github.com/Thorin215/FocusedAD | null | null | null | cs.CV | http://arxiv.org/licenses/nonexclusive-distrib/1.0/ | Movie Audio Description (AD) aims to narrate visual content during
dialogue-free segments, particularly benefiting blind and visually impaired
(BVI) audiences. Compared with general video captioning, AD demands
plot-relevant narration with explicit character name references, posing unique
challenges in movie understanding.To identify active main characters and focus
on storyline-relevant regions, we propose FocusedAD, a novel framework that
delivers character-centric movie audio descriptions. It includes: (i) a
Character Perception Module(CPM) for tracking character regions and linking
them to names; (ii) a Dynamic Prior Module(DPM) that injects contextual cues
from prior ADs and subtitles via learnable soft prompts; and (iii) a Focused
Caption Module(FCM) that generates narrations enriched with plot-relevant
details and named characters. To overcome limitations in character
identification, we also introduce an automated pipeline for building character
query banks. FocusedAD achieves state-of-the-art performance on multiple
benchmarks, including strong zero-shot results on MAD-eval-Named and our newly
proposed Cinepile-AD dataset. Code and data will be released at
https://github.com/Thorin215/FocusedAD .
| [
{
"version": "v1",
"created": "Wed, 16 Apr 2025 15:04:14 GMT"
},
{
"version": "v2",
"created": "Tue, 22 Apr 2025 12:25:51 GMT"
}
] | 2025-04-23T00:00:00 | [
[
"Ye",
"Xiaojun",
""
],
[
"Wang",
"Chun",
""
],
[
"Song",
"Yiren",
""
],
[
"Zhou",
"Sheng",
""
],
[
"Li",
"Liangcheng",
""
],
[
"Bu",
"Jiajun",
""
]
] |
2504.12198 | Kevin Carter-Fenk | Kevin Carter-Fenk | Diagrammatic Simplification of Linearized Coupled Cluster Theory | 11 pages, 6 figures | null | null | null | physics.chem-ph | http://creativecommons.org/licenses/by/4.0/ | Linearized Coupled Cluster Doubles (LinCCD) often provides near-singular
energies in small-gap systems that exhibit static correlation. This has been
attributed to the lack of quadratic $T_2^2$ terms that typically balance out
small energy denominators in the CCD amplitude equations. Herein, I show that
exchange contributions to ring and crossed-ring contractions (not small
denominators per se) cause the divergent behavior of LinCC(S)D approaches.
Rather than omitting exchange terms, I recommend a regular and size-consistent
method that retains only linear ladder diagrams. As LinCCD and configuration
interaction doubles (CID) equations are isomorphic, this also implies that
simplification (rather than quadratic extensions) of CID amplitude equations
can lead to a size-consistent theory. Linearized ladder CCD (LinLCCD) is robust
in statically-correlated systems and can be made
$O(n_{\text{occ}}^4n_{\text{vir}}^2)$ with a hole-hole approximation. The
relationship between LinLCCD and random-phase approximation sets the stage for
the development of next-generation double-hybrid density functionals that can
describe static correlation.
| [
{
"version": "v1",
"created": "Wed, 16 Apr 2025 15:48:33 GMT"
},
{
"version": "v2",
"created": "Tue, 22 Apr 2025 14:57:56 GMT"
}
] | 2025-04-23T00:00:00 | [
[
"Carter-Fenk",
"Kevin",
""
]
] |
2504.12362 | Azam Gholami | S. Gore, I. Gholami, S.O. Ahmed, T. Doskhozhina, S.V.R. Ambadipudi,
A.J. Bae and A. Gholami | Boundary Effects and Oxygen Deficiency-Driven Pattern Transitions in
Algal Bioconvection | null | null | null | null | physics.flu-dyn physics.bio-ph | http://creativecommons.org/licenses/by-nc-nd/4.0/ | Suspensions of motile microorganisms can spontaneously give rise to large
scale fluid motion, known as bioconvection, which is characterized by dense,
cell-rich downwelling plumes interspersed with broad upwelling regions. In this
study, we investigate bioconvection in shallow suspensions of Chlamydomonas
reinhardtii cells confined within spiral-shaped boundaries, combining detailed
experimental observations with three dimensional simulations. Under open
liquid-air interface conditions, cells accumulate near the surface due to
negative gravitaxis, forming spiral shaped density patterns that subsequently
fragment into lattice-like structures and give rise to downwelling plumes.
Space-time analyses reveal coherent rotational dynamics, with inward moving
patterns near the spiral core and bidirectional motion farther from the center.
Introducing confinement by sealing the top boundary with an air impermeable
transparent wall triggers striking transitions in the bioconvection patterns,
driven by oxygen depletion: initially stable structures reorganize into new
patterns with reduced characteristic wavelengths. Complementary 3D simulations,
based on the incompressible Navier-Stokes equations and incorporating negative
buoyancy and active stress from swimming cells, capture the initial pattern
formation and its subsequent instability, reproducing the fragmentation of
spiral-shaped accumulations into downwelling plumes and the emergence of strong
vortical flows, nearly an order of magnitude faster than individual cell
swimming speeds. However, these models do not capture the oxygen-driven pattern
transitions observed experimentally. Our findings reveal that confinement
geometry, oxygen dynamics, and metabolic transitions critically govern
bioconvection pattern evolution, offering new strategies to control microbial
self-organization and flow through environmental and geometric design.
| [
{
"version": "v1",
"created": "Wed, 16 Apr 2025 11:25:04 GMT"
},
{
"version": "v2",
"created": "Sat, 19 Apr 2025 10:23:41 GMT"
}
] | 2025-04-23T00:00:00 | [
[
"Gore",
"S.",
""
],
[
"Gholami",
"I.",
""
],
[
"Ahmed",
"S. O.",
""
],
[
"Doskhozhina",
"T.",
""
],
[
"Ambadipudi",
"S. V. R.",
""
],
[
"Bae",
"A. J.",
""
],
[
"Gholami",
"A.",
""
]
] |
2504.12391 | Yuri Nikolayevsky | Yuri Nikolayevsky and Wolfgang Ziller | Non-singular geodesic orbit nilmanifolds | null | null | null | null | math.DG math.GT | http://arxiv.org/licenses/nonexclusive-distrib/1.0/ | A Riemannian manifold is called a geodesic orbit manifolds, GO for short, if
any geodesic is an orbit of a one-parameter group of isometries. By a result of
C.Gordon, a non-flat GO nilmanifold is necessarily a two-step nilpotent Lie
group with a left-invariant metric. We give a complete classification of
non-singular GO nilmanifolds. Besides previously known examples, there are new
families with 3-dimensional center, and two one-parameter families of
dimensions 14 and 15.
| [
{
"version": "v1",
"created": "Wed, 16 Apr 2025 18:01:05 GMT"
},
{
"version": "v2",
"created": "Tue, 22 Apr 2025 06:16:27 GMT"
}
] | 2025-04-23T00:00:00 | [
[
"Nikolayevsky",
"Yuri",
""
],
[
"Ziller",
"Wolfgang",
""
]
] |
2504.12675 | Pengtao Dang | Pengtao Dang, Tingbo Guo, Melissa Fishel, Guang Lin, Wenzhuo Wu, Sha
Cao, Chi Zhang | Physics Informed Constrained Learning of Dynamics from Static Data | 39 pages, 10 figures | null | null | null | cs.LG physics.bio-ph q-bio.MN | http://arxiv.org/licenses/nonexclusive-distrib/1.0/ | A physics-informed neural network (PINN) models the dynamics of a system by
integrating the governing physical laws into the architecture of a neural
network. By enforcing physical laws as constraints, PINN overcomes challenges
with data scarsity and potentially high dimensionality. Existing PINN
frameworks rely on fully observed time-course data, the acquisition of which
could be prohibitive for many systems. In this study, we developed a new PINN
learning paradigm, namely Constrained Learning, that enables the approximation
of first-order derivatives or motions using non-time course or partially
observed data. Computational principles and a general mathematical formulation
of Constrained Learning were developed. We further introduced MPOCtrL (Message
Passing Optimization-based Constrained Learning) an optimization approach
tailored for the Constrained Learning framework that strives to balance the
fitting of physical models and observed data. Its code is available at github
link: https://github.com/ptdang1001/MPOCtrL Experiments on synthetic and
real-world data demonstrated that MPOCtrL can effectively detect the nonlinear
dependency between observed data and the underlying physical properties of the
system. In particular, on the task of metabolic flux analysis, MPOCtrL
outperforms all existing data-driven flux estimators.
| [
{
"version": "v1",
"created": "Thu, 17 Apr 2025 06:06:53 GMT"
},
{
"version": "v2",
"created": "Tue, 22 Apr 2025 05:57:01 GMT"
}
] | 2025-04-23T00:00:00 | [
[
"Dang",
"Pengtao",
""
],
[
"Guo",
"Tingbo",
""
],
[
"Fishel",
"Melissa",
""
],
[
"Lin",
"Guang",
""
],
[
"Wu",
"Wenzhuo",
""
],
[
"Cao",
"Sha",
""
],
[
"Zhang",
"Chi",
""
]
] |
2504.12822 | Folkert M\"uller-Hoissen | Folkert M\"uller-Hoissen and Rusuo Ye | Miura transformation in bidifferential calculus and a vectorial Darboux
transformation for the Fokas-Lenells equation | 25 pages, 5 figures, minor amendments in the second version | null | null | null | nlin.SI math-ph math.MP | http://arxiv.org/licenses/nonexclusive-distrib/1.0/ | Using a general result of bidifferential calculus and recent results of other
authors, a vectorial binary Darboux transformation is derived for the first
member of the "negative" part of the potential Kaup-Newell hierarchy, which is
a system of two coupled Fokas-Lenells equations. Miura transformations are
found from the latter to the first member of the negative part of the AKNS
hierarchy and also to its "pseudodual". The reduction to the Fokas-Lenells
equation is implemented and exact solutions with a plane wave seed generated.
| [
{
"version": "v1",
"created": "Thu, 17 Apr 2025 10:32:39 GMT"
},
{
"version": "v2",
"created": "Tue, 22 Apr 2025 11:45:40 GMT"
}
] | 2025-04-23T00:00:00 | [
[
"Müller-Hoissen",
"Folkert",
""
],
[
"Ye",
"Rusuo",
""
]
] |
2504.12867 | Guanrou Yang | Guanrou Yang, Chen Yang, Qian Chen, Ziyang Ma, Wenxi Chen, Wen Wang,
Tianrui Wang, Yifan Yang, Zhikang Niu, Wenrui Liu, Fan Yu, Zhihao Du, Zhifu
Gao, ShiLiang Zhang, Xie Chen | EmoVoice: LLM-based Emotional Text-To-Speech Model with Freestyle Text
Prompting | null | null | null | null | eess.AS cs.AI cs.CL | http://arxiv.org/licenses/nonexclusive-distrib/1.0/ | Human speech goes beyond the mere transfer of information; it is a profound
exchange of emotions and a connection between individuals. While Text-to-Speech
(TTS) models have made huge progress, they still face challenges in controlling
the emotional expression in the generated speech. In this work, we propose
EmoVoice, a novel emotion-controllable TTS model that exploits large language
models (LLMs) to enable fine-grained freestyle natural language emotion
control, and a phoneme boost variant design that makes the model output phoneme
tokens and audio tokens in parallel to enhance content consistency, inspired by
chain-of-thought (CoT) and chain-of-modality (CoM) techniques. Besides, we
introduce EmoVoice-DB, a high-quality 40-hour English emotion dataset featuring
expressive speech and fine-grained emotion labels with natural language
descriptions. EmoVoice achieves state-of-the-art performance on the English
EmoVoice-DB test set using only synthetic training data, and on the Chinese
Secap test set using our in-house data. We further investigate the reliability
of existing emotion evaluation metrics and their alignment with human
perceptual preferences, and explore using SOTA multimodal LLMs GPT-4o-audio and
Gemini to assess emotional speech. Demo samples are available at
https://yanghaha0908.github.io/EmoVoice/. Dataset, code, and checkpoints will
be released.
| [
{
"version": "v1",
"created": "Thu, 17 Apr 2025 11:50:04 GMT"
},
{
"version": "v2",
"created": "Fri, 18 Apr 2025 08:18:11 GMT"
},
{
"version": "v3",
"created": "Tue, 22 Apr 2025 02:09:12 GMT"
}
] | 2025-04-23T00:00:00 | [
[
"Yang",
"Guanrou",
""
],
[
"Yang",
"Chen",
""
],
[
"Chen",
"Qian",
""
],
[
"Ma",
"Ziyang",
""
],
[
"Chen",
"Wenxi",
""
],
[
"Wang",
"Wen",
""
],
[
"Wang",
"Tianrui",
""
],
[
"Yang",
"Yifan",
""
],
[
"Niu",
"Zhikang",
""
],
[
"Liu",
"Wenrui",
""
],
[
"Yu",
"Fan",
""
],
[
"Du",
"Zhihao",
""
],
[
"Gao",
"Zhifu",
""
],
[
"Zhang",
"ShiLiang",
""
],
[
"Chen",
"Xie",
""
]
] |
2504.12875 | Phung Lai | Phung Lai, Guanxiong Liu, NhatHai Phan, Issa Khalil, Abdallah
Khreishah, Xintao Wu | A Client-level Assessment of Collaborative Backdoor Poisoning in Non-IID
Federated Learning | null | 2025 International Conference on Distributed Computing Systems
(ICDCS) | null | null | cs.LG | http://creativecommons.org/licenses/by/4.0/ | Federated learning (FL) enables collaborative model training using
decentralized private data from multiple clients. While FL has shown robustness
against poisoning attacks with basic defenses, our research reveals new
vulnerabilities stemming from non-independent and identically distributed
(non-IID) data among clients. These vulnerabilities pose a substantial risk of
model poisoning in real-world FL scenarios.
To demonstrate such vulnerabilities, we develop a novel collaborative
backdoor poisoning attack called CollaPois. In this attack, we distribute a
single pre-trained model infected with a Trojan to a group of compromised
clients. These clients then work together to produce malicious gradients,
causing the FL model to consistently converge towards a low-loss region
centered around the Trojan-infected model. Consequently, the impact of the
Trojan is amplified, especially when the benign clients have diverse local data
distributions and scattered local gradients. CollaPois stands out by achieving
its goals while involving only a limited number of compromised clients, setting
it apart from existing attacks. Also, CollaPois effectively avoids noticeable
shifts or degradation in the FL model's performance on legitimate data samples,
allowing it to operate stealthily and evade detection by advanced robust FL
algorithms.
Thorough theoretical analysis and experiments conducted on various benchmark
datasets demonstrate the superiority of CollaPois compared to state-of-the-art
backdoor attacks. Notably, CollaPois bypasses existing backdoor defenses,
especially in scenarios where clients possess diverse data distributions.
Moreover, the results show that CollaPois remains effective even when involving
a small number of compromised clients. Notably, clients whose local data is
closely aligned with compromised clients experience higher risks of backdoor
infections.
| [
{
"version": "v1",
"created": "Thu, 17 Apr 2025 12:03:02 GMT"
},
{
"version": "v2",
"created": "Mon, 21 Apr 2025 20:19:21 GMT"
}
] | 2025-04-23T00:00:00 | [
[
"Lai",
"Phung",
""
],
[
"Liu",
"Guanxiong",
""
],
[
"Phan",
"NhatHai",
""
],
[
"Khalil",
"Issa",
""
],
[
"Khreishah",
"Abdallah",
""
],
[
"Wu",
"Xintao",
""
]
] |
2504.12876 | XinXin Qi | XinXin Qi, Hao Sun | A two-component dark matter model with $Z_2 \times Z_4$ symmetry | null | null | null | null | hep-ph | http://creativecommons.org/licenses/by/4.0/ | We consider a two-component dark matter model with $Z_2 \times Z_4$ symmetry,
where a singlet scalar $S$ and a Majorana fermion $\chi$ are introduced as dark
matter candidates. We also introduce another singlet scalar $S_0$ with a
non-zero vacuum expectation value to the SM so that the fermion dark matter can
obtain mass after spontaneous symmetry breaking. We have a new Higgs boson in
the model and in the case of the decoupling limit, the fermion dark matter
production is only determined by $S$ and the new Higgs boson. The mass
hierarchy of these new particles can make a difference in the reaction rate of
dark matter annihilation processes, contributing to different viable parameter
spaces for different mass orderings. We randomly scanned the parameter space
with six various cases under relic density constraint and found that when
$\chi$ is the lightest among the dark sector, $\chi$ production is generated
via the so-called forbidden channels. Moreover, we consider the combined limits
arising from Higgs invisible decay, dark matter relic density and direct
detection constraints. Within the chosen parameter space, direct detection
results put the most stringent constraint, and we have a more flexible value
for the scalar dark matter mass when the mass of $\chi$ is not smaller than the
new Higgs boson mass.
| [
{
"version": "v1",
"created": "Thu, 17 Apr 2025 12:04:13 GMT"
},
{
"version": "v2",
"created": "Tue, 22 Apr 2025 01:51:48 GMT"
}
] | 2025-04-23T00:00:00 | [
[
"Qi",
"XinXin",
""
],
[
"Sun",
"Hao",
""
]
] |
2504.12915 | Ebrahim Norouzi | Ebrahim Norouzi, Sven Hertling, Harald Sack | ConExion: Concept Extraction with Large Language Models | null | null | null | null | cs.CL cs.IR | http://creativecommons.org/licenses/by/4.0/ | In this paper, an approach for concept extraction from documents using
pre-trained large language models (LLMs) is presented. Compared with
conventional methods that extract keyphrases summarizing the important
information discussed in a document, our approach tackles a more challenging
task of extracting all present concepts related to the specific domain, not
just the important ones. Through comprehensive evaluations of two widely used
benchmark datasets, we demonstrate that our method improves the F1 score
compared to state-of-the-art techniques. Additionally, we explore the potential
of using prompts within these models for unsupervised concept extraction. The
extracted concepts are intended to support domain coverage evaluation of
ontologies and facilitate ontology learning, highlighting the effectiveness of
LLMs in concept extraction tasks. Our source code and datasets are publicly
available at https://github.com/ISE-FIZKarlsruhe/concept_extraction.
| [
{
"version": "v1",
"created": "Thu, 17 Apr 2025 13:05:14 GMT"
},
{
"version": "v2",
"created": "Tue, 22 Apr 2025 11:11:50 GMT"
}
] | 2025-04-23T00:00:00 | [
[
"Norouzi",
"Ebrahim",
""
],
[
"Hertling",
"Sven",
""
],
[
"Sack",
"Harald",
""
]
] |
2504.12977 | Maksim Vishnevskiy | Maksim Vishnevskiy | A Phenomenological Approach to Analyzing User Queries in IT Systems
Using Heidegger's Fundamental Ontology | 12 pages, no figures | null | 10.5281/Zenodo.15241370 | null | cs.SE cs.AI cs.CL cs.HC | http://creativecommons.org/licenses/by/4.0/ | This paper presents a novel research analytical IT system grounded in Martin
Heidegger's Fundamental Ontology, distinguishing between beings (das Seiende)
and Being (das Sein). The system employs two modally distinct, descriptively
complete languages: a categorical language of beings for processing user inputs
and an existential language of Being for internal analysis. These languages are
bridged via a phenomenological reduction module, enabling the system to analyze
user queries (including questions, answers, and dialogues among IT
specialists), identify recursive and self-referential structures, and provide
actionable insights in categorical terms. Unlike contemporary systems limited
to categorical analysis, this approach leverages Heidegger's phenomenological
existential analysis to uncover deeper ontological patterns in query
processing, aiding in resolving logical traps in complex interactions, such as
metaphor usage in IT contexts. The path to full realization involves
formalizing the language of Being by a research team based on Heidegger's
Fundamental Ontology; given the existing completeness of the language of
beings, this reduces the system's computability to completeness, paving the way
for a universal query analysis tool. The paper presents the system's
architecture, operational principles, technical implementation, use
cases--including a case based on real IT specialist dialogues--comparative
evaluation with existing tools, and its advantages and limitations.
| [
{
"version": "v1",
"created": "Thu, 17 Apr 2025 14:29:25 GMT"
}
] | 2025-04-23T00:00:00 | [
[
"Vishnevskiy",
"Maksim",
""
]
] |
2504.12988 | Yannis Montreuil | Yannis Montreuil, Axel Carlier, Lai Xing Ng, Wei Tsang Ooi | Why Ask One When You Can Ask $k$? Two-Stage Learning-to-Defer to the
Top-$k$ Experts | null | null | null | null | cs.LG stat.ML | http://creativecommons.org/licenses/by-nc-nd/4.0/ | Learning-to-Defer (L2D) enables decision-making systems to improve
reliability by selectively deferring uncertain predictions to more competent
agents. However, most existing approaches focus exclusively on single-agent
deferral, which is often inadequate in high-stakes scenarios that require
collective expertise. We propose Top-$k$ Learning-to-Defer, a generalization of
the classical two-stage L2D framework that allocates each query to the $k$ most
confident agents instead of a single one. To further enhance flexibility and
cost-efficiency, we introduce Top-$k(x)$ Learning-to-Defer, an adaptive
extension that learns the optimal number of agents to consult for each query,
based on input complexity, agent competency distributions, and consultation
costs. For both settings, we derive a novel surrogate loss and prove that it is
Bayes-consistent and $(\mathcal{R}, \mathcal{G})$-consistent, ensuring
convergence to the Bayes-optimal allocation. Notably, we show that the
well-established model cascades paradigm arises as a restricted instance of our
Top-$k$ and Top-$k(x)$ formulations. Extensive experiments across diverse
benchmarks demonstrate the effectiveness of our framework on both
classification and regression tasks.
| [
{
"version": "v1",
"created": "Thu, 17 Apr 2025 14:50:40 GMT"
},
{
"version": "v2",
"created": "Tue, 22 Apr 2025 07:02:20 GMT"
}
] | 2025-04-23T00:00:00 | [
[
"Montreuil",
"Yannis",
""
],
[
"Carlier",
"Axel",
""
],
[
"Ng",
"Lai Xing",
""
],
[
"Ooi",
"Wei Tsang",
""
]
] |
2504.13083 | Liran Shirizly | Liran Shirizly, Dekel Meirom, Malcolm Carroll and Haggai Landa | Feedforward suppression of readout-induced faults in quantum error
correction | v2 contain small fixes | null | null | null | quant-ph | http://creativecommons.org/licenses/by/4.0/ | We propose a method to reduce readout-induced faults, applicable in settings
like quantum error correction with repeated measurement cycles. The method
consists of an adaptive readout sequence conditioned on each check qubit's
readout result from the previous cycle. For readout errors, correlated
preparation errors and measurement-induced leakage that are stronger in a
particular qubit state, this feedforward protocol can suppress the physical
qubit errors. Focusing on a simple realization of conditionally flipping (by an
X gate) the state of check qubits before their measurement, we investigate the
effect of such state-dependent errors using simulations in the setup of a
low-density parity check code. We show that the suggested protocol can reduce
both logical errors and decoding time, two important aspects of fault-tolerant
quantum computations.
| [
{
"version": "v1",
"created": "Thu, 17 Apr 2025 16:50:26 GMT"
},
{
"version": "v2",
"created": "Tue, 22 Apr 2025 12:13:31 GMT"
}
] | 2025-04-23T00:00:00 | [
[
"Shirizly",
"Liran",
""
],
[
"Meirom",
"Dekel",
""
],
[
"Carroll",
"Malcolm",
""
],
[
"Landa",
"Haggai",
""
]
] |
2504.13273 | Jacob Dorn | Jacob Dorn | How Much Weak Overlap Can Doubly Robust T-Statistics Handle? | null | null | null | null | econ.EM math.ST stat.ME stat.TH | http://arxiv.org/licenses/nonexclusive-distrib/1.0/ | In the presence of sufficiently weak overlap, it is known that no regular
root-n-consistent estimators exist and standard estimators may fail to be
asymptotically normal. This paper shows that a thresholded version of the
standard doubly robust estimator is asymptotically normal with well-calibrated
Wald confidence intervals even when constructed using nonparametric estimates
of the propensity score and conditional mean outcome. The analysis implies a
cost of weak overlap in terms of black-box nuisance rates, borne when the
semiparametric bound is infinite, and the contribution of outcome smoothness to
the outcome regression rate, which is incurred even when the semiparametric
bound is finite. As a byproduct of this analysis, I show that under weak
overlap, the optimal global regression rate is the same as the optimal
pointwise regression rate, without the usual polylogarithmic penalty. The
high-level conditions yield new rules of thumb for thresholding in practice. In
simulations, thresholded AIPW can exhibit moderate overrejection in small
samples, but I am unable to reject a null hypothesis of exact coverage in large
samples. In an empirical application, the clipped AIPW estimator that targets
the standard average treatment effect yields similar precision to a heuristic
10% fixed-trimming approach that changes the target sample.
| [
{
"version": "v1",
"created": "Thu, 17 Apr 2025 18:32:49 GMT"
},
{
"version": "v2",
"created": "Tue, 22 Apr 2025 01:53:10 GMT"
}
] | 2025-04-23T00:00:00 | [
[
"Dorn",
"Jacob",
""
]
] |
2504.13336 | Lea Kunkel | Lea Kunkel, Mathias Trabs | On the minimax optimality of Flow Matching through the connection to
kernel density estimation | null | null | null | null | stat.ML cs.LG math.ST stat.TH | http://arxiv.org/licenses/nonexclusive-distrib/1.0/ | Flow Matching has recently gained attention in generative modeling as a
simple and flexible alternative to diffusion models, the current state of the
art. While existing statistical guarantees adapt tools from the analysis of
diffusion models, we take a different perspective by connecting Flow Matching
to kernel density estimation. We first verify that the kernel density estimator
matches the optimal rate of convergence in Wasserstein distance up to
logarithmic factors, improving existing bounds for the Gaussian kernel. Based
on this result, we prove that for sufficiently large networks, Flow Matching
also achieves the optimal rate up to logarithmic factors, providing a
theoretical foundation for the empirical success of this method. Finally, we
provide a first justification of Flow Matching's effectiveness in
high-dimensional settings by showing that rates improve when the target
distribution lies on a lower-dimensional linear subspace.
| [
{
"version": "v1",
"created": "Thu, 17 Apr 2025 21:06:41 GMT"
}
] | 2025-04-23T00:00:00 | [
[
"Kunkel",
"Lea",
""
],
[
"Trabs",
"Mathias",
""
]
] |
Subsets and Splits
No community queries yet
The top public SQL queries from the community will appear here once available.