id
stringlengths 9
16
| submitter
stringlengths 1
64
⌀ | authors
stringlengths 3
62.8k
| title
stringlengths 1
382
| comments
stringlengths 1
1.15k
⌀ | journal-ref
stringlengths 1
557
⌀ | doi
stringlengths 7
153
⌀ | report-no
stringlengths 1
509
⌀ | categories
stringlengths 5
125
| license
stringclasses 9
values | abstract
stringlengths 6
6.09k
| versions
listlengths 1
187
| update_date
timestamp[s]date 2007-05-23 00:00:00
2025-04-25 00:00:00
| authors_parsed
sequencelengths 1
2.83k
|
|---|---|---|---|---|---|---|---|---|---|---|---|---|---|
2504.14622 | Rebecca Silva | Rebecca B. Silva, Bin Cheng, Shing M. Lee | Dose optimization design accounting for unknown patient heterogeneity in
cancer clinical trials | null | null | null | null | stat.AP stat.ME | http://creativecommons.org/licenses/by-nc-nd/4.0/ | Project Optimus, an initiative by the FDA's Oncology Center of Excellence,
seeks to reform the dose-optimization and dose-selection paradigm in oncology.
We propose a dose-optimization design that considers plateau efficacy profiles,
integrates pharmacokinetic data to inform the exposure-toxicity curve, and
accounts for patient characteristics that may contribute to heterogeneity in
response. The dose-optimization design is carried out in two stages. First, a
toxicity-driven stage estimates a safe set of doses. Then, a dose-ranging
efficacy-driven stage explores the set using response and patient
characteristic data, employing Bayesian Sparse Group Selection to understand
patient heterogeneity. Between stages, the design integrates pharmacokinetic
data and uses futility assessments to identify the target population among the
general phase I patient population. An optimal dose is recommended for each
identified subpopulation within the target population. The simulation study
demonstrates that a model-based approach to identifying the target population
can be effective; patient characteristics relating to heterogeneity were
identified and different optimal doses were recommended for each identified
target subpopulation. Most designs that account for patient heterogeneity are
intended for trials where heterogeneity is known and pre-defined subpopulations
are specified. However, given the limited information at such an early stage,
subpopulations should be learned through the design.
| [
{
"version": "v1",
"created": "Sun, 20 Apr 2025 14:01:24 GMT"
}
] | 2025-04-22T00:00:00 | [
[
"Silva",
"Rebecca B.",
""
],
[
"Cheng",
"Bin",
""
],
[
"Lee",
"Shing M.",
""
]
] |
2504.14623 | Benjamin Monmege | B\'eatrice B\'erard, Benjamin Monmege, B Srivathsan, Arnab Sur | Synthesising Asynchronous Automata from Fair Specifications | null | null | null | null | cs.FL | http://creativecommons.org/licenses/by-sa/4.0/ | Asynchronous automata are a model of distributed finite state processes
synchronising on shared actions. A celebrated result by Zielonka shows how a
deterministic asynchronous automaton (AA) can be synthesised, starting from two
inputs: a global specification as a deterministic finite-state automaton (DFA)
and a distribution of the alphabet into local alphabets for each process. The
translation is particularly complex and has been revisited several times. In
this work, we revisit this construction on a restricted class of fair
specifications: a DFA described a fair specification if in every loop, all
processes participate in at least one action - so, no process is starved. For
fair specifications, we present a new construction to synthesise an AA. Our
construction is conceptually simpler and results in an AA where every process
has a number of local states that is linear in the number of states of the DFA,
and where the only exponential explosion is related to a parameter of fairness
(the length of the longest word that can be read in the DFA in which not every
process participates). Finally, we show how this construction can be combined
with an existing construction for hierarchical process architectures.
| [
{
"version": "v1",
"created": "Sun, 20 Apr 2025 14:01:46 GMT"
}
] | 2025-04-22T00:00:00 | [
[
"Bérard",
"Béatrice",
""
],
[
"Monmege",
"Benjamin",
""
],
[
"Srivathsan",
"B",
""
],
[
"Sur",
"Arnab",
""
]
] |
2504.14624 | Polina Gordienko | Polina Gordienko, Christoph Jansen, Thomas Augustin, Martin Rechenauer | Consensus in Motion: A Case of Dynamic Rationality of Sequential
Learning in Probability Aggregation | Submitted to the International Conference on Modeling Decisions for
Artificial Intelligence (MDAI 2025) | null | null | null | cs.AI | http://creativecommons.org/licenses/by/4.0/ | We propose a framework for probability aggregation based on propositional
probability logic. Unlike conventional judgment aggregation, which focuses on
static rationality, our model addresses dynamic rationality by ensuring that
collective beliefs update consistently with new information. We show that any
consensus-compatible and independent aggregation rule on a non-nested agenda is
necessarily linear. Furthermore, we provide sufficient conditions for a fair
learning process, where individuals initially agree on a specified subset of
propositions known as the common ground, and new information is restricted to
this shared foundation. This guarantees that updating individual judgments via
Bayesian conditioning-whether performed before or after aggregation-yields the
same collective belief. A distinctive feature of our framework is its treatment
of sequential decision-making, which allows new information to be incorporated
progressively through multiple stages while maintaining the established common
ground. We illustrate our findings with a running example in a political
scenario concerning healthcare and immigration policies.
| [
{
"version": "v1",
"created": "Sun, 20 Apr 2025 14:04:39 GMT"
}
] | 2025-04-22T00:00:00 | [
[
"Gordienko",
"Polina",
""
],
[
"Jansen",
"Christoph",
""
],
[
"Augustin",
"Thomas",
""
],
[
"Rechenauer",
"Martin",
""
]
] |
2504.14625 | Haiyan Qin | Haiyan Qin, Jiahao Feng, Xiaotong Feng, Wei W. Xing, Wang Kang | Towards Optimal Circuit Generation: Multi-Agent Collaboration Meets
Collective Intelligence | 9 pages, 6 figures | null | null | null | cs.AR cs.AI | http://creativecommons.org/licenses/by-nc-sa/4.0/ | Large language models (LLMs) have transformed code generation, yet their
application in hardware design produces gate counts 38\%--1075\% higher than
human designs. We present CircuitMind, a multi-agent framework that achieves
human-competitive efficiency through three key innovations: syntax locking
(constraining generation to basic logic gates), retrieval-augmented generation
(enabling knowledge-driven design), and dual-reward optimization (balancing
correctness with efficiency). To evaluate our approach, we introduce TC-Bench,
the first gate-level benchmark harnessing collective intelligence from the
TuringComplete ecosystem -- a competitive circuit design platform with hundreds
of thousands of players. Experiments show CircuitMind enables 55.6\% of model
implementations to match or exceed top-tier human experts in composite
efficiency metrics. Most remarkably, our framework elevates the 14B Phi-4 model
to outperform both GPT-4o mini and Gemini 2.0 Flash, achieving efficiency
comparable to the top 25\% of human experts without requiring specialized
training. These innovations establish a new paradigm for hardware optimization
where collaborative AI systems leverage collective human expertise to achieve
optimal circuit designs. Our model, data, and code are open-source at
https://github.com/BUAA-CLab/CircuitMind.
| [
{
"version": "v1",
"created": "Sun, 20 Apr 2025 14:05:17 GMT"
}
] | 2025-04-22T00:00:00 | [
[
"Qin",
"Haiyan",
""
],
[
"Feng",
"Jiahao",
""
],
[
"Feng",
"Xiaotong",
""
],
[
"Xing",
"Wei W.",
""
],
[
"Kang",
"Wang",
""
]
] |
2504.14626 | Santanu Roy Dr | Santanu Roy, Shweta Singh, Palak Sahu, Ashvath Suresh, Debashish Das | MSAD-Net: Multiscale and Spatial Attention-based Dense Network for Lung
Cancer Classification | null | null | null | null | cs.CV | http://creativecommons.org/licenses/by/4.0/ | Lung cancer, a severe form of malignant tumor that originates in the tissues
of the lungs, can be fatal if not detected in its early stages. It ranks among
the top causes of cancer-related mortality worldwide. Detecting lung cancer
manually using chest X-Ray image or Computational Tomography (CT) scans image
poses significant challenges for radiologists. Hence, there is a need for
automatic diagnosis system of lung cancers from radiology images. With the
recent emergence of deep learning, particularly through Convolutional Neural
Networks (CNNs), the automated detection of lung cancer has become a much
simpler task. Nevertheless, numerous researchers have addressed that the
performance of conventional CNNs may be hindered due to class imbalance issue,
which is prevalent in medical images. In this research work, we have proposed a
novel CNN architecture ``Multi-Scale Dense Network (MSD-Net)''
(trained-from-scratch). The novelties we bring in the proposed model are (I) We
introduce novel dense modules in the 4th block and 5th block of the CNN model.
We have leveraged 3 depthwise separable convolutional (DWSC) layers, and one
1x1 convolutional layer in each dense module, in order to reduce complexity of
the model considerably. (II) Additionally, we have incorporated one skip
connection from 3rd block to 5th block and one parallel branch connection from
4th block to Global Average Pooling (GAP) layer. We have utilized dilated
convolutional layer (with dilation rate=2) in the last parallel branch in order
to extract multi-scale features. Extensive experiments reveal that our proposed
model has outperformed latest CNN model ConvNext-Tiny, recent trend Vision
Transformer (ViT), Pooling-based ViT (PiT), and other existing models by
significant margins.
| [
{
"version": "v1",
"created": "Sun, 20 Apr 2025 14:07:21 GMT"
}
] | 2025-04-22T00:00:00 | [
[
"Roy",
"Santanu",
""
],
[
"Singh",
"Shweta",
""
],
[
"Sahu",
"Palak",
""
],
[
"Suresh",
"Ashvath",
""
],
[
"Das",
"Debashish",
""
]
] |
2504.14627 | Arash Yadegari Ghahderijani | Arash Yadegari Ghahderijani and Hande Naz Turgay and Dimka
Karastoyanova | Change Logging and Mining of Change Logs of Business Processes -- A
Literature Review | null | null | null | null | cs.SE | http://creativecommons.org/licenses/by-nc-sa/4.0/ | Context: Change mining enables organizations to understand the changes that
occurred in their business processes. This allows them to enhance their
business processes and adapt to dynamic environments. Therefore, change mining
is becoming a topic of interest for researchers, scholars, and practitioners.
Objective: Motivated by the goal of establishing the state of the art in this
area, this paper aims to investigate the literature in change logging and
mining in process-aware information systems, provide an overview of the methods
that are used in the existing publications, and identify gaps in the research
on the topic of logging and mining process changes.
Method: A literature review is conducted with the objective to identify and
define methods to mine, store, and record changes in business processes. From
1136 publications, we selected 6 papers related to changes in business process
and extended the list to 9 papers by including the relevant articles referenced
by the papers that we selected originally.
Results: In answer of our research questions, we have identified two classes
of change mining methods, two ways of recording the changes into change logs,
five formats for change log representation, and four objectives to be learned
from changes.
Conclusion: The literature review provides a summary of existing change
mining and logging methods in process-aware information systems and identifies
a number of research gaps in the area.
| [
{
"version": "v1",
"created": "Sun, 20 Apr 2025 14:08:38 GMT"
}
] | 2025-04-22T00:00:00 | [
[
"Ghahderijani",
"Arash Yadegari",
""
],
[
"Turgay",
"Hande Naz",
""
],
[
"Karastoyanova",
"Dimka",
""
]
] |
2504.14628 | Shunxin Guo | Shunxin Guo, Jiaqi Lv, Qiufeng Wang, and Xin Geng | GENE-FL: Gene-Driven Parameter-Efficient Dynamic Federated Learning | null | null | null | null | cs.DC cs.LG | http://creativecommons.org/licenses/by/4.0/ | Real-world \underline{F}ederated \underline{L}earning systems often encounter
\underline{D}ynamic clients with \underline{A}gnostic and highly heterogeneous
data distributions (DAFL), which pose challenges for efficient communication
and model initialization. To address these challenges, we draw inspiration from
the recently proposed Learngene paradigm, which compresses the large-scale
model into lightweight, cross-task meta-information fragments. Learngene
effectively encapsulates and communicates core knowledge, making it
particularly well-suited for DAFL, where dynamic client participation requires
communication efficiency and rapid adaptation to new data distributions. Based
on this insight, we propose a Gene-driven parameter-efficient dynamic Federated
Learning (GENE-FL) framework. First, local models perform quadratic constraints
based on parameters with high Fisher values in the global model, as these
parameters are considered to encapsulate generalizable knowledge. Second, we
apply the strategy of parameter sensitivity analysis in local model parameters
to condense the \textit{learnGene} for interaction. Finally, the server
aggregates these small-scale trained \textit{learnGene}s into a robust
\textit{learnGene} with cross-task generalization capability, facilitating the
rapid initialization of dynamic agnostic client models. Extensive experimental
results demonstrate that GENE-FL reduces \textbf{4 $\times$} communication
costs compared to FEDAVG and effectively initializes agnostic client models
with only about \textbf{9.04} MB.
| [
{
"version": "v1",
"created": "Sun, 20 Apr 2025 14:10:02 GMT"
}
] | 2025-04-22T00:00:00 | [
[
"Guo",
"Shunxin",
""
],
[
"Lv",
"Jiaqi",
""
],
[
"Wang",
"Qiufeng",
""
],
[
"Geng",
"Xin",
""
]
] |
2504.14629 | Ivan Mikhailov | Ivan N. Mikhailov | New geodesic lines in the Gromov-Hausdorff class lying in the cloud of
the real line | 9 pages, 2 pictures, accepted for publication in Chebyshevskii
Sbornik | null | null | null | math.MG | http://creativecommons.org/licenses/by/4.0/ | In the paper we prove that, for arbitrary unbounded subset $A\subset R$ and
an arbitrary bounded metric space~$X$, a curve $A\times_{\ell^1} (tX)$,
$t\in[0,\,\infty)$ is a geodesic line in the Gromov--Hausdorff class. We also
show that, for abitrary $\lambda > 1$, $n\in\mathbb{N}$, the following
inequality holds:
$d_{GH}\bigl(\mathbb{Z}^n,\,\lambda\mathbb{Z}^n\bigr)\ge\frac{1}{2}$. We
conclude that a curve $t\mathbb{Z}^n$, $t\in(0,\,\infty)$ is not continuous
with respect to the Gromov--Hausdorff distance, and, therefore, is not a
gedesic line. Moreover, it follows that multiplication of all metric spaces
lying on the finite Gromov--Hausdorff distance from $\mathbb{R}^n$ on
some~$\lambda > 0$ is also discontinous with respect to the Gromov--Hausdorff
distance.
| [
{
"version": "v1",
"created": "Sun, 20 Apr 2025 14:11:57 GMT"
}
] | 2025-04-22T00:00:00 | [
[
"Mikhailov",
"Ivan N.",
""
]
] |
2504.14630 | Hossein Hassani | Rondik Hadi Abdulrahman and Hossein Hassani | Automatic Text Summarization (ATS) for Research Documents in Sorani
Kurdish | 18 pages, 11 figures, 8 tables | null | null | null | cs.CL | http://creativecommons.org/licenses/by-nc-nd/4.0/ | Extracting concise information from scientific documents aids learners,
researchers, and practitioners. Automatic Text Summarization (ATS), a key
Natural Language Processing (NLP) application, automates this process. While
ATS methods exist for many languages, Kurdish remains underdeveloped due to
limited resources. This study develops a dataset and language model based on
231 scientific papers in Sorani Kurdish, collected from four academic
departments in two universities in the Kurdistan Region of Iraq (KRI),
averaging 26 pages per document. Using Sentence Weighting and Term
Frequency-Inverse Document Frequency (TF-IDF) algorithms, two experiments were
conducted, differing in whether the conclusions were included. The average word
count was 5,492.3 in the first experiment and 5,266.96 in the second. Results
were evaluated manually and automatically using ROUGE-1, ROUGE-2, and ROUGE-L
metrics, with the best accuracy reaching 19.58%. Six experts conducted manual
evaluations using three criteria, with results varying by document. This
research provides valuable resources for Kurdish NLP researchers to advance ATS
and related fields.
| [
{
"version": "v1",
"created": "Sun, 20 Apr 2025 14:17:17 GMT"
}
] | 2025-04-22T00:00:00 | [
[
"Abdulrahman",
"Rondik Hadi",
""
],
[
"Hassani",
"Hossein",
""
]
] |
2504.14631 | Thomas Weber | Thomas Weber | Explainability for Embedding AI: Aspirations and Actuality | Second Workshop on Engineering Interactive Systems Embedding AI
Technologies at EICS 2024, Tuesday June 25th, 2024 - Cagliary, Sardinia,
Italy | null | null | null | cs.HC cs.SE | http://creativecommons.org/licenses/by/4.0/ | With artificial intelligence (AI) embedded in many everyday software systems,
effectively and reliably developing and maintaining AI systems becomes an
essential skill for software developers. However, the complexity inherent to AI
poses new challenges. Explainable AI (XAI) may allow developers to understand
better the systems they build, which, in turn, can help with tasks like
debugging. In this paper, we report insights from a series of surveys with
software developers that highlight that there is indeed an increased need for
explanatory tools to support developers in creating AI systems. However, the
feedback also indicates that existing XAI systems still fall short of this
aspiration. Thus, we see an unmet need to provide developers with adequate
support mechanisms to cope with this complexity so they can embed AI into
high-quality software in the future.
| [
{
"version": "v1",
"created": "Sun, 20 Apr 2025 14:20:01 GMT"
}
] | 2025-04-22T00:00:00 | [
[
"Weber",
"Thomas",
""
]
] |
2504.14632 | Shu Li | Shu Li, Binxiang Dai | Hopf bifurcation in a memory-based diffusion competition model with
spatial heterogeneity | null | null | null | null | math.DS | http://arxiv.org/licenses/nonexclusive-distrib/1.0/ | In this paper, we investigate a Lotka-Volterra competition-diffusion system
with self-memory effects and spatial heterogeneity under Dirichlet boundary
conditions. We focus on how memory strength influences the coexistence and
stability of competing species. By analyzing the characteristic equation, we
establish the existence and stability of a spatially nonhomogeneous positive
steady state and demonstrate the occurrence of Hopf bifurcation as memory delay
increases. Our results reveal that both weak and some opposing memory effects
of two competing species promote stable coexistence, while strong memory may
destabilize the system and lead to periodic oscillations. Spatial heterogeneity
further enriches the dynamical behaviors. Numerical simulations are presented
to confirm the theoretical results.
| [
{
"version": "v1",
"created": "Sun, 20 Apr 2025 14:21:30 GMT"
}
] | 2025-04-22T00:00:00 | [
[
"Li",
"Shu",
""
],
[
"Dai",
"Binxiang",
""
]
] |
2504.14633 | Ji-Jun Park | Soo-joon Choi, Ji-jun Park | Harnessing Generative LLMs for Enhanced Financial Event Entity
Extraction Performance | null | null | null | null | cs.CL | http://arxiv.org/licenses/nonexclusive-distrib/1.0/ | Financial event entity extraction is a crucial task for analyzing market
dynamics and building financial knowledge graphs, yet it presents significant
challenges due to the specialized language and complex structures in financial
texts. Traditional approaches often rely on sequence labeling models, which can
struggle with long-range dependencies and the inherent complexity of extracting
multiple, potentially overlapping entities. Motivated by the advanced language
understanding and generative capabilities of Large Language Models (LLMs), we
propose a novel method that reframes financial event entity extraction as a
text-to-structured-output generation task. Our approach involves fine-tuning a
pre-trained LLM using Parameter-Efficient Fine-Tuning (PEFT) to directly
generate a structured representation, such as a JSON object, containing the
extracted entities and their precise character spans from the input text. We
evaluate our method on the challenging CCKS 2019 Financial Event Entity
Extraction dataset, comparing its performance against strong sequence labeling
baselines, including SEBERTNets and sebertNets. Experimental results
demonstrate that our generative LLM method achieves a new state-of-the-art F1
score on this benchmark, significantly outperforming previous methods. Through
detailed quantitative analysis across event types, entity types, and instance
complexity, as well as human evaluation, we show that our approach is more
effective at handling the nuances of financial text and extracting high-quality
entities. This work validates the potential of applying generative LLMs
directly to complex, domain-specific information extraction tasks requiring
structured output.
| [
{
"version": "v1",
"created": "Sun, 20 Apr 2025 14:23:31 GMT"
}
] | 2025-04-22T00:00:00 | [
[
"Choi",
"Soo-joon",
""
],
[
"Park",
"Ji-jun",
""
]
] |
2504.14635 | Thierry Dufour | Thierry Dufour, Elliot Ogden | Positive and negative DC glow discharges: a comparative study to
characterize self-organized patterns on water surface | null | Physics of Plasma, Vol. 32, Issue 4 (2025) | 10.1063/5.0251603 | null | physics.plasm-ph nlin.AO | http://creativecommons.org/licenses/by/4.0/ | Self-organized patterns (SOP) in plasma discharges arise from the complex
interplay of electric field, reactive species and charged particles, driven by
non-linear plasma dynamics. While studies have explored SOP formation in
various configurations, no systematic comparison of positive and negative DC
glow discharges has been conducted to explain why SOP form exclusively when
polarization is negative. This study aims to analyze SOP formation mechanisms
by comparing electrical, optical and spectral properties of positive and
negative DC glow discharges interacting with a grounded water surface. Key
differences in gas temperature, electric field and reactive species
distribution are hence identified. For positive DC glow discharges (PGD), the
gas temperature remains in the 350-370 K range, while the reduced electric
field remains below 100 Td across the gap. The plasma is dominated by OH and
N2* species, whose excitation results from direct electron impact and energy
transfer in a low-field environment. The absence of strong ionization and
electric field gradients leads to a spatially homogeneous emission layer on the
liquid surface, resulting in a circular uniform plasma (CUP) pattern without
self-organization. In contrast, SOP emerge exclusively under negative DC glow
discharges (NGD) at currents above 15 mA. These discharges present a reduced
electric field, peaking at 485 Td at 1mm from the cathode pin, dropping below
100 Td in the central gap and rising to 460 Td near the water surface. There,
the plasma layer still contains OH and N2* species but also N2+ ions, the
latter being critical for SOP formation. SOP morphology evolves with gap size:
at 7 mm, patterns transition from specks to filaments, with pattern diameter
and thickness as high as 5.5 mm and 210 {\mu}m respectively. Lowering water
surface tension with surfactants reduces SOP size and modifies pattern
morphology.
| [
{
"version": "v1",
"created": "Sun, 20 Apr 2025 14:25:18 GMT"
}
] | 2025-04-22T00:00:00 | [
[
"Dufour",
"Thierry",
""
],
[
"Ogden",
"Elliot",
""
]
] |
2504.14636 | Binjie Guo | Binjie Guo, Hanyu Zheng, Guowei Su, Ru Zhang, Haohan Jiang, Xurong
Lin, Hongyan Wei, Aisheng Mo, Jie Li, Zhiyuan Qian, Zhuhao Zhang, Xiaoyuan
Cheng | AlphaZero-Edu: Making AlphaZero Accessible to Everyone | null | null | null | null | cs.LG cs.AI | http://creativecommons.org/licenses/by/4.0/ | Recent years have witnessed significant progress in reinforcement learning,
especially with Zero-like paradigms, which have greatly boosted the
generalization and reasoning abilities of large-scale language models.
Nevertheless, existing frameworks are often plagued by high implementation
complexity and poor reproducibility. To tackle these challenges, we present
AlphaZero-Edu, a lightweight, education-focused implementation built upon the
mathematical framework of AlphaZero. It boasts a modular architecture that
disentangles key components, enabling transparent visualization of the
algorithmic processes. Additionally, it is optimized for resource-efficient
training on a single NVIDIA RTX 3090 GPU and features highly parallelized
self-play data generation, achieving a 3.2-fold speedup with 8 processes. In
Gomoku matches, the framework has demonstrated exceptional performance,
achieving a consistently high win rate against human opponents. AlphaZero-Edu
has been open-sourced at https://github.com/StarLight1212/AlphaZero_Edu,
providing an accessible and practical benchmark for both academic research and
industrial applications.
| [
{
"version": "v1",
"created": "Sun, 20 Apr 2025 14:29:39 GMT"
}
] | 2025-04-22T00:00:00 | [
[
"Guo",
"Binjie",
""
],
[
"Zheng",
"Hanyu",
""
],
[
"Su",
"Guowei",
""
],
[
"Zhang",
"Ru",
""
],
[
"Jiang",
"Haohan",
""
],
[
"Lin",
"Xurong",
""
],
[
"Wei",
"Hongyan",
""
],
[
"Mo",
"Aisheng",
""
],
[
"Li",
"Jie",
""
],
[
"Qian",
"Zhiyuan",
""
],
[
"Zhang",
"Zhuhao",
""
],
[
"Cheng",
"Xiaoyuan",
""
]
] |
2504.14637 | Thierry Dufour | M. Soulier, T. Vacek, K. Geraud, T. Dufour | Transferred plasma catheter for endotherapeutic applications: a
parametric study of guided streamers dynamics | null | Physics of Plasma, Vol. 32, Issue 4 (2025) | 10.1063/5.0254402 | null | physics.plasm-ph physics.med-ph | http://creativecommons.org/licenses/by/4.0/ | Non-thermal atmospheric pressure plasma jets (APPJs) are increasingly used in
biomedical applications due to their low temperatures and ability to generate
reactive oxygen and nitrogen species (RONS), making them suitable for sensitive
environments like medical therapies. The transferred plasma catheter (TPC), a
variant of APPJ, shows promise for endoscopic applications but requires precise
control of plasma dynamics in confined spaces to ensure safety and efficacy.
Despite extensive studies on guided streamers in traditional APPJs, there is
limited understanding of streamer behavior in TPC configurations, particularly
in challenging scenarios involving grounded metallic surfaces. This study
examines the spatiotemporal dynamics of guided streamers generated by TPCs
under varying gap distances to establish a robust framework for safe and
effective plasma delivery in endoscopic settings. Combining electrical and
optical diagnostics, the study characterizes streamer propagation, electric
field profiles, and plasma-induced currents in a helium-driven TPC delivering
cold plasma to a grounded metal target across gaps of 2 to 18 mm. Results show
that streamers maintain charge stability and effectively interact with the
target for gap distances below 12 mm, producing significant therapeutic
currents. Beyond this threshold, propagation deteriorates due to recombination
and reduced electric field intensity. For shorter gaps, counter-propagating
waves and secondary streamer interactions are observed, while larger gaps lead
to charge dissipation and reduced efficacy. These findings highlight the
importance of optimizing gap distances for plasma-assisted endoscopic
procedures and demonstrate the TPC's robustness in adverse conditions.
| [
{
"version": "v1",
"created": "Sun, 20 Apr 2025 14:29:57 GMT"
}
] | 2025-04-22T00:00:00 | [
[
"Soulier",
"M.",
""
],
[
"Vacek",
"T.",
""
],
[
"Geraud",
"K.",
""
],
[
"Dufour",
"T.",
""
]
] |
2504.14638 | Junyuan Fang | Junyuan Fang (1 and 2), Zihan Wang (1), Yejun Zhang (1), Shuzhe Wang
(1), Iaroslav Melekhov (1), Juho Kannala (1 and 3) ((1) Aalto University,
Espoo, Finland, (2) University of Helsinki, Helsinki, Finland, (3) University
of Oulu, Oulu, Finland) | NVSMask3D: Hard Visual Prompting with Camera Pose Interpolation for 3D
Open Vocabulary Instance Segmentation | 15 pages, 4 figures, Scandinavian Conference on Image Analysis 2025 | null | null | null | cs.CV | http://creativecommons.org/licenses/by-nc-sa/4.0/ | Vision-language models (VLMs) have demonstrated impressive zero-shot transfer
capabilities in image-level visual perception tasks. However, they fall short
in 3D instance-level segmentation tasks that require accurate localization and
recognition of individual objects. To bridge this gap, we introduce a novel 3D
Gaussian Splatting based hard visual prompting approach that leverages camera
interpolation to generate diverse viewpoints around target objects without any
2D-3D optimization or fine-tuning. Our method simulates realistic 3D
perspectives, effectively augmenting existing hard visual prompts by enforcing
geometric consistency across viewpoints. This training-free strategy seamlessly
integrates with prior hard visual prompts, enriching object-descriptive
features and enabling VLMs to achieve more robust and accurate 3D instance
segmentation in diverse 3D scenes.
| [
{
"version": "v1",
"created": "Sun, 20 Apr 2025 14:39:27 GMT"
}
] | 2025-04-22T00:00:00 | [
[
"Fang",
"Junyuan",
"",
"1 and 2"
],
[
"Wang",
"Zihan",
"",
"1 and 3"
],
[
"Zhang",
"Yejun",
"",
"1 and 3"
],
[
"Wang",
"Shuzhe",
"",
"1 and 3"
],
[
"Melekhov",
"Iaroslav",
"",
"1 and 3"
],
[
"Kannala",
"Juho",
"",
"1 and 3"
]
] |
2504.14639 | Changming Huang | Xiaoyang Wang, Qidong Fu, Changming Huang | Topological linear and nonlinear gap modes in defected dimer lattice
with fourth-order diffraction | 8 pages, 7 figures, to appear in Chaos, Solitons and Fractals | null | null | null | physics.optics nlin.PS | http://arxiv.org/licenses/nonexclusive-distrib/1.0/ | In this study, we investigate the optical properties of linear and nonlinear
topological in-phase, out-of-phase, and edge modes in a defected dimer lattice
with fourth-order diffraction, encompassing their bifurcation characteristics,
localized field distributions, and stability properties. Both focusing and
defocusing nonlinearities are considered. Within the nontrivial regime,
in-phase, out-of-phase, and edge modes can emerge within the gap. Numerical
results reveal that three types of topological gap solitons bifurcate from
their corresponding linear localized modes. The dimerization parameter can be
tuned to drive the system from a trivial to a nontrivial configuration. We
observe that as the strength of the fourth-order diffraction increases, the
size of the spectral gap also enlarges, with this parameter effectively
broadening the stability window for solitons. Our findings offer novel insights
into the properties of both linear and nonlinear modes in topologically
defected structures.
| [
{
"version": "v1",
"created": "Sun, 20 Apr 2025 14:41:11 GMT"
}
] | 2025-04-22T00:00:00 | [
[
"Wang",
"Xiaoyang",
""
],
[
"Fu",
"Qidong",
""
],
[
"Huang",
"Changming",
""
]
] |
2504.14640 | Yuheng Huang | Yuheng Huang, Lei Ma, Keizaburo Nishikino, Takumi Akazaki | Risk Assessment Framework for Code LLMs via Leveraging Internal States | To appear in the 33rd ACM International Conference on the Foundations
of Software Engineering (FSE Companion'25 Industry Track), June 23-28, 2025,
Trondheim, Norway. This work was supported by Fujitsu Limited | Companion Proceedings of the 33rd ACM International Conference on
the Foundations of Software Engineering (FSE'25 Industry Track), June 2025 | 10.1145/3696630.3728566 | null | cs.SE cs.AI cs.CL | http://arxiv.org/licenses/nonexclusive-distrib/1.0/ | The pre-training paradigm plays a key role in the success of Large Language
Models (LLMs), which have been recognized as one of the most significant
advancements of AI recently. Building on these breakthroughs, code LLMs with
advanced coding capabilities bring huge impacts on software engineering,
showing the tendency to become an essential part of developers' daily routines.
However, the current code LLMs still face serious challenges related to
trustworthiness, as they can generate incorrect, insecure, or unreliable code.
Recent exploratory studies find that it can be promising to detect such risky
outputs by analyzing LLMs' internal states, akin to how the human brain
unconsciously recognizes its own mistakes. Yet, most of these approaches are
limited to narrow sub-domains of LLM operations and fall short of achieving
industry-level scalability and practicability. To address these challenges, in
this paper, we propose PtTrust, a two-stage risk assessment framework for code
LLM based on internal state pre-training, designed to integrate seamlessly with
the existing infrastructure of software companies. The core idea is that the
risk assessment framework could also undergo a pre-training process similar to
LLMs. Specifically, PtTrust first performs unsupervised pre-training on
large-scale unlabeled source code to learn general representations of LLM
states. Then, it uses a small, labeled dataset to train a risk predictor. We
demonstrate the effectiveness of PtTrust through fine-grained, code line-level
risk assessment and demonstrate that it generalizes across tasks and different
programming languages. Further experiments also reveal that PtTrust provides
highly intuitive and interpretable features, fostering greater user trust. We
believe PtTrust makes a promising step toward scalable and trustworthy
assurance for code LLMs.
| [
{
"version": "v1",
"created": "Sun, 20 Apr 2025 14:44:18 GMT"
}
] | 2025-04-22T00:00:00 | [
[
"Huang",
"Yuheng",
""
],
[
"Ma",
"Lei",
""
],
[
"Nishikino",
"Keizaburo",
""
],
[
"Akazaki",
"Takumi",
""
]
] |
2504.14641 | Grace Li Zhang | Kangwei Xu, Bing Li, Grace Li Zhang, Ulf Schlichtmann | HLSTester: Efficient Testing of Behavioral Discrepancies with LLMs for
High-Level Synthesis | null | null | null | null | cs.SE cs.SY eess.SY | http://creativecommons.org/licenses/by-nc-sa/4.0/ | In high-level synthesis (HLS), C/C++ programs with synthesis directives are
used to generate circuits for FPGA implementations. However, hardware-specific
and platform-dependent characteristics in these implementations can introduce
behavioral discrepancies between the original C/C++ programs and the circuits
after high-level synthesis. Existing methods for testing behavioral
discrepancies in HLS are still immature, and the testing workflow requires
significant human efforts. To address this challenge, we propose HLSTester, a
large language model (LLM) aided testing framework that efficiently detects
behavioral discrepancies in HLS. To mitigate hallucinations in LLMs and enhance
prompt quality, the testbenches for original C/C++ programs are leveraged to
guide LLMs in generating HLS-compatible testbenches, effectively eliminating
certain traditional C/C++ constructs that are incompatible with HLS tools. Key
variables are pinpointed through a backward slicing technique in both C/C++ and
HLS programs to monitor their runtime spectra, enabling an in-depth analysis of
the discrepancy symptoms. To reduce test time, a testing input generation
mechanism is introduced to integrate dynamic mutation with insights from an
LLM-based progressive reasoning chain. In addition, repetitive hardware testing
is skipped by a redundancy-aware filtering technique for the generated test
inputs. Experimental results demonstrate that the proposed LLM-aided testing
framework significantly accelerates the testing workflow while achieving higher
testbench simulation pass rates compared with the traditional method and the
direct use of LLMs on the same HLS programs.
| [
{
"version": "v1",
"created": "Sun, 20 Apr 2025 14:45:01 GMT"
}
] | 2025-04-22T00:00:00 | [
[
"Xu",
"Kangwei",
""
],
[
"Li",
"Bing",
""
],
[
"Zhang",
"Grace Li",
""
],
[
"Schlichtmann",
"Ulf",
""
]
] |
2504.14642 | Lin Li | Lin Li, Wei Chen, Jiahui Li, Long Chen | Relation-R1: Cognitive Chain-of-Thought Guided Reinforcement Learning
for Unified Relational Comprehension | Ongoing project | null | null | null | cs.CV | http://arxiv.org/licenses/nonexclusive-distrib/1.0/ | Recent advances in multi-modal large language models (MLLMs) have
significantly improved object-level grounding and region captioning, but remain
limited in visual relation understanding (\eg, scene graph generation),
particularly in modeling \textit{N}-ary relationships that identify multiple
semantic roles among an action event. Such a lack of \textit{semantic
dependencies} modeling among multi-entities leads to unreliable outputs,
intensifying MLLMs' hallucinations and over-reliance on language priors. To
this end, we propose Relation-R1, the first unified relational comprehension
framework that explicitly integrates cognitive chain-of-thought (CoT)-guided
Supervised Fine-Tuning (SFT) and Group Relative Policy Optimization (GRPO)
within a reinforcement learning (RL) paradigm. Specifically, we first establish
foundational reasoning capabilities via SFT, enforcing structured outputs with
thinking processes. Then, GRPO is utilized to refine these outputs via
multi-reward optimization, prioritizing visual-semantic grounding over
language-induced biases, thereby improving generalization capability. Extensive
experiments on widely-used PSG and SWiG datasets demonstrate that Relation-R1
achieves state-of-the-art performance in both binary and \textit{N}-ary
relation understanding.
| [
{
"version": "v1",
"created": "Sun, 20 Apr 2025 14:50:49 GMT"
}
] | 2025-04-22T00:00:00 | [
[
"Li",
"Lin",
""
],
[
"Chen",
"Wei",
""
],
[
"Li",
"Jiahui",
""
],
[
"Chen",
"Long",
""
]
] |
2504.14643 | Robin Blume-Kohout | Robin Blume-Kohout and Kevin Young | Estimating detector error models from syndrome data | null | null | null | null | quant-ph | http://creativecommons.org/licenses/by/4.0/ | Protecting quantum information using quantum error correction (QEC) requires
repeatedly measuring stabilizers to extract error syndromes that are used to
identify and correct errors. Syndrome extraction data provides information
about the processes that cause errors. The collective effects of these
processes can be described by a detector error model (DEM). We show how to
estimate probabilities of individual DEM events, and of aggregated classes of
DEM events, using data from multiple cycles of syndrome extraction.
| [
{
"version": "v1",
"created": "Sun, 20 Apr 2025 14:56:56 GMT"
}
] | 2025-04-22T00:00:00 | [
[
"Blume-Kohout",
"Robin",
""
],
[
"Young",
"Kevin",
""
]
] |
2504.14645 | Philipp Altmann | Philipp Altmann, C\'eline Davignon, Maximilian Zorn, Fabian Ritz,
Claudia Linnhoff-Popien, Thomas Gabor | Surrogate Fitness Metrics for Interpretable Reinforcement Learning | 30 pages, 7 figures, under review | null | null | null | cs.LG cs.AI | http://arxiv.org/licenses/nonexclusive-distrib/1.0/ | We employ an evolutionary optimization framework that perturbs initial states
to generate informative and diverse policy demonstrations. A joint surrogate
fitness function guides the optimization by combining local diversity,
behavioral certainty, and global population diversity. To assess demonstration
quality, we apply a set of evaluation metrics, including the reward-based
optimality gap, fidelity interquartile means (IQMs), fitness composition
analysis, and trajectory visualizations. Hyperparameter sensitivity is also
examined to better understand the dynamics of trajectory optimization. Our
findings demonstrate that optimizing trajectory selection via surrogate fitness
metrics significantly improves interpretability of RL policies in both discrete
and continuous environments. In gridworld domains, evaluations reveal
significantly enhanced demonstration fidelities compared to random and ablated
baselines. In continuous control, the proposed framework offers valuable
insights, particularly for early-stage policies, while fidelity-based
optimization proves more effective for mature policies. By refining and
systematically analyzing surrogate fitness functions, this study advances the
interpretability of RL models. The proposed improvements provide deeper
insights into RL decision-making, benefiting applications in safety-critical
and explainability-focused domains.
| [
{
"version": "v1",
"created": "Sun, 20 Apr 2025 15:01:19 GMT"
}
] | 2025-04-22T00:00:00 | [
[
"Altmann",
"Philipp",
""
],
[
"Davignon",
"Céline",
""
],
[
"Zorn",
"Maximilian",
""
],
[
"Ritz",
"Fabian",
""
],
[
"Linnhoff-Popien",
"Claudia",
""
],
[
"Gabor",
"Thomas",
""
]
] |
2504.14646 | Michael Kinyon | Alexander Grishkov, Michael Kinyon and Petr Vojt\v{e}chovsk\'y | Bol loops of order 27 | 12 pages, amsart | null | null | null | math.GR | http://creativecommons.org/licenses/by/4.0/ | We classify Bol loops of order $27$, using a combination of theoretical
results and computer search. There are $15$ Bol loops of order $27$, including
five groups. New constructions for the ten nonassociative Bol loops of order
$27$ are given.
| [
{
"version": "v1",
"created": "Sun, 20 Apr 2025 15:01:52 GMT"
}
] | 2025-04-22T00:00:00 | [
[
"Grishkov",
"Alexander",
""
],
[
"Kinyon",
"Michael",
""
],
[
"Vojtěchovský",
"Petr",
""
]
] |
2504.14647 | Ashis Mandal | Alice Fialowski and Ashis Mandal | Deformations of the five dimensional Heisenberg Lie algebra | 9 pages | null | null | null | math.QA math.KT | http://arxiv.org/licenses/nonexclusive-distrib/1.0/ | In this note we explicitly give all the equivalent classes of deformations of
the 5-dimensional Heisenberg Lie algebra $\mathfrak{h}_2$ over complex or real
number fields. We show that there are altogether 20 infinitesimal deformations
(families), 18 of them being extendable to real deformations and 2 of them are
only infinitesimal.
| [
{
"version": "v1",
"created": "Sun, 20 Apr 2025 15:11:49 GMT"
}
] | 2025-04-22T00:00:00 | [
[
"Fialowski",
"Alice",
""
],
[
"Mandal",
"Ashis",
""
]
] |
2504.14648 | Si-Qiang Luo | Si-Qiang Luo, Xiang Liu | Identifying triple-strangeness $\Omega$ hyperons in light of recent
experimental results | 10 pages, 1 figure, 7 tables | null | null | null | hep-ph hep-ex | http://arxiv.org/licenses/nonexclusive-distrib/1.0/ | In this work, we systematically calculate the mass spectra and decay widths
of $\Omega$ baryons with $N\leq3$. In our scheme, the newly observed
$\Omega(2109)$ is identified as a $N=2$ candidate, with its quantum numbers
tentatively assigned as $\Omega(2S,3/2^+)$. However, the experimentally
measured mass exhibits a deviation of approximately 50 MeV below the
theoretical predictions. The possible reason for this deviation, along with the
suggestion for experiments, are discussed in this paper. Additionally, we give
possible assignments of $\Omega(2012)^-$, $\Omega(2250)^-$, $\Omega(2380)^-$,
and $\Omega(2470)^-$. Furthermore, we calculate the masses and widths of other
missing $\Omega$ states, which could provide valuable clues for experimental
searches aimed at discovering these states.
| [
{
"version": "v1",
"created": "Sun, 20 Apr 2025 15:12:01 GMT"
}
] | 2025-04-22T00:00:00 | [
[
"Luo",
"Si-Qiang",
""
],
[
"Liu",
"Xiang",
""
]
] |
2504.14649 | Tianqi Song | Eugene Tang KangJie, Tianqi Song, Zicheng Zhu, Jingshu Li, Yi-Chieh
Lee | AI Literacy Education for Older Adults: Motivations, Challenges and
Preferences | Extended Abstracts of the CHI Conference on Human Factors in
Computing Systems | null | null | null | cs.HC | http://creativecommons.org/licenses/by/4.0/ | As Artificial Intelligence (AI) becomes increasingly integrated into older
adults' daily lives, equipping them with the knowledge and skills to understand
and use AI is crucial. However, most research on AI literacy education has
focused on students and children, leaving a gap in understanding the unique
needs of older adults when learning about AI. To address this, we surveyed 103
older adults aged 50 and above (Mean = 64, SD = 7). Results revealed that they
found it important and were motivated to learn about AI because they wish to
harness the benefits and avoid the dangers of AI, seeing it as necessary to
cope in the future. However, they expressed learning challenges such as
difficulties in understanding and not knowing how to start learning AI.
Particularly, a strong preference for hands-on learning was indicated. We
discussed design opportunities to support AI literacy education for older
adults.
| [
{
"version": "v1",
"created": "Sun, 20 Apr 2025 15:12:04 GMT"
}
] | 2025-04-22T00:00:00 | [
[
"KangJie",
"Eugene Tang",
""
],
[
"Song",
"Tianqi",
""
],
[
"Zhu",
"Zicheng",
""
],
[
"Li",
"Jingshu",
""
],
[
"Lee",
"Yi-Chieh",
""
]
] |
2504.14650 | Yuting Huang | Yuting Huang, Leilei Ding, Zhipeng Tang, Tianfu Wang, Xinrui Lin,
Wuyang Zhang, Mingxiao Ma, Yanyong Zhang | A Framework for Benchmarking and Aligning Task-Planning Safety in
LLM-Based Embodied Agents | 16 pages, 10 figures | null | null | null | cs.AI | http://arxiv.org/licenses/nonexclusive-distrib/1.0/ | Large Language Models (LLMs) exhibit substantial promise in enhancing
task-planning capabilities within embodied agents due to their advanced
reasoning and comprehension. However, the systemic safety of these agents
remains an underexplored frontier. In this study, we present Safe-BeAl, an
integrated framework for the measurement (SafePlan-Bench) and alignment
(Safe-Align) of LLM-based embodied agents' behaviors. SafePlan-Bench
establishes a comprehensive benchmark for evaluating task-planning safety,
encompassing 2,027 daily tasks and corresponding environments distributed
across 8 distinct hazard categories (e.g., Fire Hazard). Our empirical analysis
reveals that even in the absence of adversarial inputs or malicious intent,
LLM-based agents can exhibit unsafe behaviors. To mitigate these hazards, we
propose Safe-Align, a method designed to integrate physical-world safety
knowledge into LLM-based embodied agents while maintaining task-specific
performance. Experiments across a variety of settings demonstrate that
Safe-BeAl provides comprehensive safety validation, improving safety by 8.55 -
15.22%, compared to embodied agents based on GPT-4, while ensuring successful
task completion.
| [
{
"version": "v1",
"created": "Sun, 20 Apr 2025 15:12:14 GMT"
}
] | 2025-04-22T00:00:00 | [
[
"Huang",
"Yuting",
""
],
[
"Ding",
"Leilei",
""
],
[
"Tang",
"Zhipeng",
""
],
[
"Wang",
"Tianfu",
""
],
[
"Lin",
"Xinrui",
""
],
[
"Zhang",
"Wuyang",
""
],
[
"Ma",
"Mingxiao",
""
],
[
"Zhang",
"Yanyong",
""
]
] |
2504.14651 | Luca Giacomelli | Luca Giacomelli, Michel H. Devoret, Cristiano Ciuti | Exact Duality at Low Energy in a Josephson Tunnel Junction Coupled to a
Transmission Line | 16 pages, 10 figures | null | null | null | quant-ph cond-mat.mes-hall cond-mat.supr-con | http://arxiv.org/licenses/nonexclusive-distrib/1.0/ | We theoretically explore the low-energy behavior of a Josephson tunnel
junction coupled to a finite-length, charge-biased transmission line and
compare it to its flux-biased counterpart. For transmission lines of increasing
length, we show that the low-energy charge-dependent energy bands of the
charge-biased configuration can be exactly mapped onto those of the flux-biased
system via a well-defined duality transformation of circuit parameters. In the
limit of an infinitely long transmission line, the influence of boundary
conditions vanishes, and both circuits reduce to a resistively shunted
Josephson junction. This convergence reveals the system's intrinsic
self-duality and critical behavior. Our exact formulation of charge-flux
duality provides a foundation for generalizations to more complex
superconductor-insulator phase transitions.
| [
{
"version": "v1",
"created": "Sun, 20 Apr 2025 15:13:10 GMT"
}
] | 2025-04-22T00:00:00 | [
[
"Giacomelli",
"Luca",
""
],
[
"Devoret",
"Michel H.",
""
],
[
"Ciuti",
"Cristiano",
""
]
] |
2504.14652 | Nishu Goyal | N. Goyal, S. Aplin, A. Balerna, P. Bell, J. Casas, M. Cascella, S.
Chatterji, C. Cohen, E. Collet, G. Dennis, P. Fajardo, E. N. Gimenez, H.
Graafsma, H. Hiresmann, F. J. Iguaz, K. Klementiev, T. Kolodziej, L.
Manzanillas, T. Martin, R. H. Menk, M. Porro, M. Quispe, B. Schmitt, S.
Scully, M. Turcato, C. Ward, E. Welter | Progress in the Development of Multi-Element Monolithic Germanium
Detectors in LEAPS-INNOV Project: Insights from Detector Performance
Simulation | null | null | null | null | physics.ins-det | http://creativecommons.org/licenses/by/4.0/ | This study presents a detailed simulation-based analysis of the detection
limits of multi-element monolithic Germanium (Ge) detectors to cadmium traces
in environmental soil samples. Using the capabilities of the Geant4 Monte Carlo
toolkit in combination with the Solid State Detector Package, we evaluated the
detection limit variation with the sample-to-detector distances and photon
flux. These simulations were conducted to mimic realistic conditions, with a
photon flux measured by the SAMBA beamline at the SOLEIL synchrotron facility.
Our findings for the detection limit for trace amounts of pollutants in low
concentrations like cadmium in the soil provide valuable insights for
optimizing experimental setups in environmental monitoring and
synchrotron-based applications, where precise detection of trace elements is
critical.
| [
{
"version": "v1",
"created": "Sun, 20 Apr 2025 15:16:25 GMT"
}
] | 2025-04-22T00:00:00 | [
[
"Goyal",
"N.",
""
],
[
"Aplin",
"S.",
""
],
[
"Balerna",
"A.",
""
],
[
"Bell",
"P.",
""
],
[
"Casas",
"J.",
""
],
[
"Cascella",
"M.",
""
],
[
"Chatterji",
"S.",
""
],
[
"Cohen",
"C.",
""
],
[
"Collet",
"E.",
""
],
[
"Dennis",
"G.",
""
],
[
"Fajardo",
"P.",
""
],
[
"Gimenez",
"E. N.",
""
],
[
"Graafsma",
"H.",
""
],
[
"Hiresmann",
"H.",
""
],
[
"Iguaz",
"F. J.",
""
],
[
"Klementiev",
"K.",
""
],
[
"Kolodziej",
"T.",
""
],
[
"Manzanillas",
"L.",
""
],
[
"Martin",
"T.",
""
],
[
"Menk",
"R. H.",
""
],
[
"Porro",
"M.",
""
],
[
"Quispe",
"M.",
""
],
[
"Schmitt",
"B.",
""
],
[
"Scully",
"S.",
""
],
[
"Turcato",
"M.",
""
],
[
"Ward",
"C.",
""
],
[
"Welter",
"E.",
""
]
] |
2504.14653 | Fenghao Zhu | Fenghao Zhu, Xinquan Wang, Xinyi Li, Maojun Zhang, Yixuan Chen,
Chongwen Huang, Zhaohui Yang, Xiaoming Chen, Zhaoyang Zhang, Richeng Jin,
Yongming Huang, Wei Feng, Tingting Yang, Baoming Bai, Feifei Gao, Kun Yang,
Yuanwen Liu, Sami Muhaidat, Chau Yuen, Kaibin Huang, Kai-Kit Wong, Dusit
Niyato and M\'erouane Debbah | Wireless Large AI Model: Shaping the AI-Native Future of 6G and Beyond | null | null | null | null | cs.IT eess.SP math.IT | http://arxiv.org/licenses/nonexclusive-distrib/1.0/ | The emergence of sixth-generation and beyond communication systems is
expected to fundamentally transform digital experiences through introducing
unparalleled levels of intelligence, efficiency, and connectivity. A promising
technology poised to enable this revolutionary vision is the wireless large AI
model (WLAM), characterized by its exceptional capabilities in data processing,
inference, and decision-making. In light of these remarkable capabilities, this
paper provides a comprehensive survey of WLAM, elucidating its fundamental
principles, diverse applications, critical challenges, and future research
opportunities. We begin by introducing the background of WLAM and analyzing the
key synergies with wireless networks, emphasizing the mutual benefits.
Subsequently, we explore the foundational characteristics of WLAM, delving into
their unique relevance in wireless environments. Then, the role of WLAM in
optimizing wireless communication systems across various use cases and the
reciprocal benefits are systematically investigated. Furthermore, we discuss
the integration of WLAM with emerging technologies, highlighting their
potential to enable transformative capabilities and breakthroughs in wireless
communication. Finally, we thoroughly examine the high-level challenges
hindering the practical implementation of WLAM and discuss pivotal future
research directions.
| [
{
"version": "v1",
"created": "Sun, 20 Apr 2025 15:25:58 GMT"
}
] | 2025-04-22T00:00:00 | [
[
"Zhu",
"Fenghao",
""
],
[
"Wang",
"Xinquan",
""
],
[
"Li",
"Xinyi",
""
],
[
"Zhang",
"Maojun",
""
],
[
"Chen",
"Yixuan",
""
],
[
"Huang",
"Chongwen",
""
],
[
"Yang",
"Zhaohui",
""
],
[
"Chen",
"Xiaoming",
""
],
[
"Zhang",
"Zhaoyang",
""
],
[
"Jin",
"Richeng",
""
],
[
"Huang",
"Yongming",
""
],
[
"Feng",
"Wei",
""
],
[
"Yang",
"Tingting",
""
],
[
"Bai",
"Baoming",
""
],
[
"Gao",
"Feifei",
""
],
[
"Yang",
"Kun",
""
],
[
"Liu",
"Yuanwen",
""
],
[
"Muhaidat",
"Sami",
""
],
[
"Yuen",
"Chau",
""
],
[
"Huang",
"Kaibin",
""
],
[
"Wong",
"Kai-Kit",
""
],
[
"Niyato",
"Dusit",
""
],
[
"Debbah",
"Mérouane",
""
]
] |
2504.14654 | Merve G\"ulmez | Hossam ElAtali, Merve G\"ulmez, Thomas Nyman, N. Asokan | BLACKOUT: Data-Oblivious Computation with Blinded Capabilities | null | null | null | null | cs.CR | http://arxiv.org/licenses/nonexclusive-distrib/1.0/ | Lack of memory-safety and exposure to side channels are two prominent,
persistent challenges for the secure implementation of software. Memory-safe
programming languages promise to significantly reduce the prevalence of
memory-safety bugs, but make it more difficult to implement
side-channel-resistant code. We aim to address both memory-safety and
side-channel resistance by augmenting memory-safe hardware with the ability for
data-oblivious programming. We describe an extension to the CHERI capability
architecture to provide blinded capabilities that allow data-oblivious
computation to be carried out by userspace tasks. We also present BLACKOUT, our
realization of blinded capabilities on a FPGA softcore based on the speculative
out-of-order CHERI-Toooba processor and extend the CHERI-enabled Clang/LLVM
compiler and the CheriBSD operating system with support for blinded
capabilities. BLACKOUT makes writing side-channel-resistant code easier by
making non-data-oblivious operations via blinded capabilities explicitly fault.
Through rigorous evaluation we show that BLACKOUT ensures memory operated on
through blinded capabilities is securely allocated, used, and reclaimed and
demonstrate that, in benchmarks comparable to those used by previous work,
BLACKOUT imposes only a small performance degradation (1.5% geometric mean)
compared to the baseline CHERI-Toooba processor.
| [
{
"version": "v1",
"created": "Sun, 20 Apr 2025 15:25:59 GMT"
}
] | 2025-04-22T00:00:00 | [
[
"ElAtali",
"Hossam",
""
],
[
"Gülmez",
"Merve",
""
],
[
"Nyman",
"Thomas",
""
],
[
"Asokan",
"N.",
""
]
] |
2504.14655 | Wei Shen | Yunhui Xia, Wei Shen, Yan Wang, Jason Klein Liu, Huifeng Sun, Siyue
Wu, Jian Hu, Xiaolong Xu | LeetCodeDataset: A Temporal Dataset for Robust Evaluation and Efficient
Training of Code LLMs | null | null | null | null | cs.LG cs.CL cs.SE | http://arxiv.org/licenses/nonexclusive-distrib/1.0/ | We introduce LeetCodeDataset, a high-quality benchmark for evaluating and
training code-generation models, addressing two key challenges in LLM research:
the lack of reasoning-focused coding benchmarks and self-contained training
testbeds. By curating LeetCode Python problems with rich metadata, broad
coverage, 100+ test cases per problem, and temporal splits (pre/post July
2024), our dataset enables contamination-free evaluation and efficient
supervised fine-tuning (SFT). Experiments show reasoning models significantly
outperform non-reasoning counterparts, while SFT with only 2.6K model-generated
solutions achieves performance comparable to 110K-sample counterparts. The
dataset and evaluation framework are available on Hugging Face and Github.
| [
{
"version": "v1",
"created": "Sun, 20 Apr 2025 15:28:16 GMT"
}
] | 2025-04-22T00:00:00 | [
[
"Xia",
"Yunhui",
""
],
[
"Shen",
"Wei",
""
],
[
"Wang",
"Yan",
""
],
[
"Liu",
"Jason Klein",
""
],
[
"Sun",
"Huifeng",
""
],
[
"Wu",
"Siyue",
""
],
[
"Hu",
"Jian",
""
],
[
"Xu",
"Xiaolong",
""
]
] |
2504.14656 | Hayato Shimabukuro | Hayato Shimabukuro, Yidong Xu and Yue Shao | Analyzing the 21cm forest with Wavelet Scattering Transform: Insight
into non-Gaussian features of the 21cm forest | 16 pages, 9 figures | null | null | null | astro-ph.CO | http://creativecommons.org/publicdomain/zero/1.0/ | The 21cm forest, narrow absorption features in the spectra of high redshift
radio sources caused by intervening neutral hydrogen, offers a unique probe of
the intergalactic medium and small-scale structures during reionization. While
traditional power spectrum methods have been widely used for analyzing the 21cm
forest, these techniques are limited in capturing the non-Gaussian nature of
the signal. In this work, we introduce the Wavelet Scattering Transform (WST)
as a novel diagnostic tool for the 21cm forest, which allows for the extraction
of higher-order statistical features that power spectrum methods cannot easily
capture. By decomposing simulated brightness temperature spectra into a
hierarchy of scattering coefficients, the WST isolates both local intensity
fluctuations (first-order coefficients) and scale-scale correlations
(second-order coefficients), revealing the complex, multi-scale non-Gaussian
interactions inherent in the 21cm forest. This approach enhances the power of
21cm forest in distinguishing between different cosmological models, such as
Cold Dark Matter (CDM) and Warm Dark Matter (WDM), as well as scenarios with
enhanced X-ray heating. Unlike traditional methods, which focus primarily on
Gaussian statistics, the WST captures richer astrophysical and cosmological
information. Our analysis shows that WST can significantly improve constraints
on key parameters, such as the X-ray heating efficiency and the WDM particle
mass, providing deeper insights into the early stages of cosmic structure
formation.
| [
{
"version": "v1",
"created": "Sun, 20 Apr 2025 15:35:04 GMT"
}
] | 2025-04-22T00:00:00 | [
[
"Shimabukuro",
"Hayato",
""
],
[
"Xu",
"Yidong",
""
],
[
"Shao",
"Yue",
""
]
] |
2504.14657 | Simon A. Lee | Yihan Lin and Zhirong Bella Yu and Simon Lee | A Case Study Exploring the Current Landscape of Synthetic Medical Record
Generation with Commercial LLMs | Accepted at the Conference of Health, Inference, Learning (CHIL 2025)
in Berkeley, CA. To appear in PMLR later in 2025 | null | null | null | cs.CL cs.AI cs.LG | http://creativecommons.org/licenses/by/4.0/ | Synthetic Electronic Health Records (EHRs) offer a valuable opportunity to
create privacy preserving and harmonized structured data, supporting numerous
applications in healthcare. Key benefits of synthetic data include precise
control over the data schema, improved fairness and representation of patient
populations, and the ability to share datasets without concerns about
compromising real individuals privacy. Consequently, the AI community has
increasingly turned to Large Language Models (LLMs) to generate synthetic data
across various domains. However, a significant challenge in healthcare is
ensuring that synthetic health records reliably generalize across different
hospitals, a long standing issue in the field. In this work, we evaluate the
current state of commercial LLMs for generating synthetic data and investigate
multiple aspects of the generation process to identify areas where these models
excel and where they fall short. Our main finding from this work is that while
LLMs can reliably generate synthetic health records for smaller subsets of
features, they struggle to preserve realistic distributions and correlations as
the dimensionality of the data increases, ultimately limiting their ability to
generalize across diverse hospital settings.
| [
{
"version": "v1",
"created": "Sun, 20 Apr 2025 15:37:05 GMT"
}
] | 2025-04-22T00:00:00 | [
[
"Lin",
"Yihan",
""
],
[
"Yu",
"Zhirong Bella",
""
],
[
"Lee",
"Simon",
""
]
] |
2504.14659 | Elad Domanovitz | Elad Domanovitz, Anatoly Khina | Markovian Continuity of the MMSE | null | null | null | null | eess.SP cs.IT math.IT math.ST stat.TH | http://creativecommons.org/licenses/by-sa/4.0/ | Minimum mean square error (MMSE) estimation is widely used in signal
processing and related fields. While it is known to be non-continuous with
respect to all standard notions of stochastic convergence, it remains robust in
practical applications. In this work, we review the known counterexamples to
the continuity of the MMSE. We observe that, in these counterexamples, the
discontinuity arises from an element in the converging measurement sequence
providing more information about the estimand than the limit of the measurement
sequence. We argue that this behavior is uncharacteristic of real-world
applications and introduce a new stochastic convergence notion, termed
Markovian convergence, to address this issue. We prove that the MMSE is, in
fact, continuous under this new notion. We supplement this result with
semi-continuity and continuity guarantees of the MMSE in other settings and
prove the continuity of the MMSE under linear estimation.
| [
{
"version": "v1",
"created": "Sun, 20 Apr 2025 15:42:41 GMT"
}
] | 2025-04-22T00:00:00 | [
[
"Domanovitz",
"Elad",
""
],
[
"Khina",
"Anatoly",
""
]
] |
2504.14660 | Deepu Kumar | Deepu Kumar, Joydev Khatua, Nguyen The Hoang, Yumin Sim, Rajesh Kumar
Ulaganathan, Raju Kalaivanan, Raman Sankar, Maeng-Je Seong, and Kwang-Yong
Choi | Manipulation of anisotropic Zhang-Rice exciton in van der Waals
antiferromagnets NiPS3-xSex by anion substitution | null | null | null | null | cond-mat.mtrl-sci | http://arxiv.org/licenses/nonexclusive-distrib/1.0/ | Spin-entangled excitons have emerged as intriguing quasi-particle excitations
in van der Waals magnets. Among them, the recently observed Zhang-Rice (ZR)
exciton in NiPS3 has garnered significant research interest due to its strong
correlation with magnetic ordering and its exceptionally long-lived coherence.
Herein, we present our in-depth temperature- and polarization-dependent
photoluminescence (PL) study of anion-substituted NiPS3-xSex (x=0.008,0.03,0.06
and 0.09 ) to explore the nature and dynamics of the ZR exciton. Our results
reveal that, similar to the cation substitution, a small percentage of anion
substitution effectively destroys and modulates the ZR exciton, as evidenced by
the emergence of a weaker, lower-energy PL peak in addition to the primary ZR
peak. The primary and secondary PL peaks exhibit the same anisotropic
polarization but differ in their peak energy shift and intensity evolution with
Se substitution, suggesting varying charge transfers of p-orbitals. Notably,
the ZR exciton undergoes rapid thermal destabilization at much lower
temperatures than two-magnon excitations, highlighting that p-orbital
inhomogeneity beyond the magnetic ordering structure is a decisive factor in
driving its thermal quenching.
| [
{
"version": "v1",
"created": "Sun, 20 Apr 2025 15:43:07 GMT"
}
] | 2025-04-22T00:00:00 | [
[
"Kumar",
"Deepu",
""
],
[
"Khatua",
"Joydev",
""
],
[
"Hoang",
"Nguyen The",
""
],
[
"Sim",
"Yumin",
""
],
[
"Ulaganathan",
"Rajesh Kumar",
""
],
[
"Kalaivanan",
"Raju",
""
],
[
"Sankar",
"Raman",
""
],
[
"Seong",
"Maeng-Je",
""
],
[
"Choi",
"Kwang-Yong",
""
]
] |
2504.14661 | Hui-Huang Chen | Hui-Huang Chen, Xin-Liang Zhou, Jiayu Yin, Ming Zhang | Symmetry resolved entanglement entropy after an inhomogeneous quench | 17 pages, 1 figure | null | null | null | hep-th cond-mat.stat-mech | http://creativecommons.org/licenses/by/4.0/ | We investigate the non-equilibrium time evolution of symmetry-resolved
entanglement entropy (SREE) following an inhomogeneous quench in a critical
one-dimensional free fermion system. Using conformal field theory, we derive an
exact expression for the SREE and analyze its behavior. We find that, at
leading order in the long-time limit, the SREE grows logarithmically as $\log
t$. While the equipartition of entanglement holds at leading order, we identify
subleading corrections that break it. Our numerical simulations corroborate the
analytical predictions with excellent agreement.
| [
{
"version": "v1",
"created": "Sun, 20 Apr 2025 15:47:02 GMT"
}
] | 2025-04-22T00:00:00 | [
[
"Chen",
"Hui-Huang",
""
],
[
"Zhou",
"Xin-Liang",
""
],
[
"Yin",
"Jiayu",
""
],
[
"Zhang",
"Ming",
""
]
] |
2504.14662 | Yeoreum Lee | Yeoreum Lee, Jinwook Jung, Sungyong Baik | Mitigating Parameter Interference in Model Merging via Sharpness-Aware
Fine-Tuning | ICLR 2025 | null | null | null | cs.LG cs.CV | http://creativecommons.org/licenses/by/4.0/ | Large-scale deep learning models with a pretraining-finetuning paradigm have
led to a surge of numerous task-specific models fine-tuned from a common
pre-trained model. Recently, several research efforts have been made on merging
these large models into a single multi-task model, particularly with simple
arithmetic on parameters. Such merging methodology faces a central challenge:
interference between model parameters fine-tuned on different tasks. Few recent
works have focused on designing a new fine-tuning scheme that can lead to small
parameter interference, however at the cost of the performance of each
task-specific fine-tuned model and thereby limiting that of a merged model. To
improve the performance of a merged model, we note that a fine-tuning scheme
should aim for (1) smaller parameter interference and (2) better performance of
each fine-tuned model on the corresponding task. In this work, we aim to design
a new fine-tuning objective function to work towards these two goals. In the
course of this process, we find such objective function to be strikingly
similar to sharpness-aware minimization (SAM) objective function, which aims to
achieve generalization by finding flat minima. Drawing upon our observation, we
propose to fine-tune pre-trained models via sharpness-aware minimization. The
experimental and theoretical results showcase the effectiveness and
orthogonality of our proposed approach, improving performance upon various
merging and fine-tuning methods. Our code is available at
https://github.com/baiklab/SAFT-Merge.
| [
{
"version": "v1",
"created": "Sun, 20 Apr 2025 15:57:12 GMT"
}
] | 2025-04-22T00:00:00 | [
[
"Lee",
"Yeoreum",
""
],
[
"Jung",
"Jinwook",
""
],
[
"Baik",
"Sungyong",
""
]
] |
2504.14663 | Edvaldo R. Wassouf Jr. | Edvaldo Wassouf Jr, D\'ebora Paiva, Kiev Gama and Awdren Font\~ao | The Developer Experience of LGBTQIA+ People in Agile Teams: a Multivocal
Literature Review | Preprint with the original submission accepted for publication at
CHASE (2025) | null | null | null | cs.SE | http://creativecommons.org/licenses/by/4.0/ | Research on underrepresented populations is essential for fostering greater
diversity within the software industry. Team diversity is important for reasons
that go beyond ethics. Diversity contributes to greater innovation and
productivity, helping decrease turnover rates and reduce team conflicts. Within
this context, LGBTQIA+ software engineering professionals face unique
challenges, e.g., self-isolation and invisibility feeling. Developer Experience
(DX) encompasses cognitive, emotional, and motivational considerations,
supporting the idea that improving how DX can enhance team performance,
strengthen collaboration, and lead to more successful software projects. This
study aimed to examine traditional and grey literature data through a
Multivocal Literature Review focused on the DX of LGBTQIA+ professionals in
agile teams. Our findings reveal that issues such as invisibility, prejudice,
and discrimination adversely affect their experiences, compounded by the
predominance of heterosexual males in the field. Conversely, professionals who
feel welcomed by their teams and organizations, especially in processes
tailored to their needs, report more positive team dynamics and engagement.
| [
{
"version": "v1",
"created": "Sun, 20 Apr 2025 16:00:37 GMT"
}
] | 2025-04-22T00:00:00 | [
[
"Wassouf",
"Edvaldo",
"Jr"
],
[
"Paiva",
"Débora",
""
],
[
"Gama",
"Kiev",
""
],
[
"Fontão",
"Awdren",
""
]
] |
2504.14664 | Sun Jixiang | Jixiang Sun, Fei Lei, Jiawei Zhang, Wenxiu Sun, Yujiu Yang | Frequency-domain Learning with Kernel Prior for Blind Image Deblurring | null | null | null | null | cs.CV | http://creativecommons.org/licenses/by/4.0/ | While achieving excellent results on various datasets, many deep learning
methods for image deblurring suffer from limited generalization capabilities
with out-of-domain data. This limitation is likely caused by their dependence
on certain domain-specific datasets. To address this challenge, we argue that
it is necessary to introduce the kernel prior into deep learning methods, as
the kernel prior remains independent of the image context. For effective fusion
of kernel prior information, we adopt a rational implementation method inspired
by traditional deblurring algorithms that perform deconvolution in the
frequency domain. We propose a module called Frequency Integration Module (FIM)
for fusing the kernel prior and combine it with a frequency-based deblurring
Transfomer network. Experimental results demonstrate that our method
outperforms state-of-the-art methods on multiple blind image deblurring tasks,
showcasing robust generalization abilities. Source code will be available soon.
| [
{
"version": "v1",
"created": "Sun, 20 Apr 2025 16:00:38 GMT"
}
] | 2025-04-22T00:00:00 | [
[
"Sun",
"Jixiang",
""
],
[
"Lei",
"Fei",
""
],
[
"Zhang",
"Jiawei",
""
],
[
"Sun",
"Wenxiu",
""
],
[
"Yang",
"Yujiu",
""
]
] |
2504.14665 | A S M Sharifuzzaman Sagar | A S M Sharifuzzaman Sagar, Yu Chen and Jun Hoong Chan | DMPCN: Dynamic Modulated Predictive Coding Network with Hybrid Feedback
Representations | null | null | null | null | cs.CV | http://creativecommons.org/licenses/by/4.0/ | Traditional predictive coding networks, inspired by theories of brain
function, consistently achieve promising results across various domains,
extending their influence into the field of computer vision. However, the
performance of the predictive coding networks is limited by their error
feedback mechanism, which traditionally employs either local or global
recurrent updates, leading to suboptimal performance in processing both local
and broader details simultaneously. In addition, traditional predictive coding
networks face difficulties in dynamically adjusting to the complexity and
context of varying input data, which is crucial for achieving high levels of
performance in diverse scenarios. Furthermore, there is a gap in the
development and application of specific loss functions that could more
effectively guide the model towards optimal performance. To deal with these
issues, this paper introduces a hybrid prediction error feedback mechanism with
dynamic modulation for deep predictive coding networks by effectively combining
global contexts and local details while adjusting feedback based on input
complexity. Additionally, we present a loss function tailored to this framework
to improve accuracy by focusing on precise prediction error minimization.
Experimental results demonstrate the superiority of our model over other
approaches, showcasing faster convergence and higher predictive accuracy in
CIFAR-10, CIFAR-100, MNIST, and FashionMNIST datasets.
| [
{
"version": "v1",
"created": "Sun, 20 Apr 2025 16:14:07 GMT"
}
] | 2025-04-22T00:00:00 | [
[
"Sagar",
"A S M Sharifuzzaman",
""
],
[
"Chen",
"Yu",
""
],
[
"Chan",
"Jun Hoong",
""
]
] |
2504.14666 | Kaihang Pan | Kaihang Pan, Wang Lin, Zhongqi Yue, Tenglong Ao, Liyu Jia, Wei Zhao,
Juncheng Li, Siliang Tang, Hanwang Zhang | Generative Multimodal Pretraining with Discrete Diffusion Timestep
Tokens | Accepted by CVPR 2025 (Oral) | null | null | null | cs.CV | http://arxiv.org/licenses/nonexclusive-distrib/1.0/ | Recent endeavors in Multimodal Large Language Models (MLLMs) aim to unify
visual comprehension and generation by combining LLM and diffusion models, the
state-of-the-art in each task, respectively. Existing approaches rely on
spatial visual tokens, where image patches are encoded and arranged according
to a spatial order (e.g., raster scan). However, we show that spatial tokens
lack the recursive structure inherent to languages, hence form an impossible
language for LLM to master. In this paper, we build a proper visual language by
leveraging diffusion timesteps to learn discrete, recursive visual tokens. Our
proposed tokens recursively compensate for the progressive attribute loss in
noisy images as timesteps increase, enabling the diffusion model to reconstruct
the original image at any timestep. This approach allows us to effectively
integrate the strengths of LLMs in autoregressive reasoning and diffusion
models in precise image generation, achieving seamless multimodal comprehension
and generation within a unified framework. Extensive experiments show that we
achieve superior performance for multimodal comprehension and generation
simultaneously compared with other MLLMs. Project Page:
https://DDT-LLaMA.github.io/.
| [
{
"version": "v1",
"created": "Sun, 20 Apr 2025 16:14:28 GMT"
}
] | 2025-04-22T00:00:00 | [
[
"Pan",
"Kaihang",
""
],
[
"Lin",
"Wang",
""
],
[
"Yue",
"Zhongqi",
""
],
[
"Ao",
"Tenglong",
""
],
[
"Jia",
"Liyu",
""
],
[
"Zhao",
"Wei",
""
],
[
"Li",
"Juncheng",
""
],
[
"Tang",
"Siliang",
""
],
[
"Zhang",
"Hanwang",
""
]
] |
2504.14667 | Kai Zhao | Kai Zhao, Zhaohui Yang | Efficient Federated Split Learning for Large Language Models over
Communication Networks | null | null | null | null | cs.LG cs.NI | http://arxiv.org/licenses/nonexclusive-distrib/1.0/ | Fine-tuning pre-trained large language models (LLM) in a distributed manner
poses significant challenges on resource-constrained edge devices. To address
this challenge, we propose FedsLLM, a novel framework that integrates split
federated learning with parameter-efficient fine-tuning techniques. By
leveraging model splitting and Low-Rank Adaptation (LoRA), FedsLLM reduces the
computational burden on edge devices. Furthermore, the introduction of a
federated server facilitates parallel training and enhances privacy. To
accommodate heterogeneous communication conditions and diverse computational
capabilities of edge devices, as well as the impact of LoRA rank selection on
model convergence and training cost, we formulate a joint optimization problem.
The formulated problem jointly optimizes subchannel allocation, power control,
model splitting point selection, and LoRA rank configuration, all aimed at
minimizing total training delay. An alternating optimization algorithm is
developed to efficiently solve this problem and accelerate the training
process. Simulation results demonstrate that the proposed FedsLLM framework
achieves comparable model accuracy while significantly reducing client-side
computational requirements. Furthermore, the proposed resource allocation
scheme and adaptive LoRA rank selection strategy notably reduce the training
latency compared to conventional approaches.
| [
{
"version": "v1",
"created": "Sun, 20 Apr 2025 16:16:54 GMT"
}
] | 2025-04-22T00:00:00 | [
[
"Zhao",
"Kai",
""
],
[
"Yang",
"Zhaohui",
""
]
] |
2504.14668 | John DeVadoss | John deVadoss and Matthias Artzt | A Byzantine Fault Tolerance Approach towards AI Safety | 14 pages | null | null | null | cs.DC | http://creativecommons.org/licenses/by/4.0/ | Ensuring that an AI system behaves reliably and as intended, especially in
the presence of unexpected faults or adversarial conditions, is a complex
challenge. Inspired by the field of Byzantine Fault Tolerance (BFT) from
distributed computing, we explore a fault tolerance architecture for AI safety.
By drawing an analogy between unreliable, corrupt, misbehaving or malicious AI
artifacts and Byzantine nodes in a distributed system, we propose an
architecture that leverages consensus mechanisms to enhance AI safety and
reliability.
| [
{
"version": "v1",
"created": "Sun, 20 Apr 2025 16:18:06 GMT"
}
] | 2025-04-22T00:00:00 | [
[
"deVadoss",
"John",
""
],
[
"Artzt",
"Matthias",
""
]
] |
2504.14669 | Wei Zou | Wei Zou, Sen Yang, Yu Bao, Shujian Huang, Jiajun Chen, Shanbo Cheng | Trans-Zero: Self-Play Incentivizes Large Language Models for
Multilingual Translation Without Parallel Data | 11 pages, 4 figures | null | null | null | cs.CL | http://creativecommons.org/licenses/by-nc-sa/4.0/ | The rise of Large Language Models (LLMs) has reshaped machine translation
(MT), but multilingual MT still relies heavily on parallel data for supervised
fine-tuning (SFT), facing challenges like data scarcity for low-resource
languages and catastrophic forgetting. To address these issues, we propose
TRANS-ZERO, a self-play framework that leverages only monolingual data and the
intrinsic multilingual knowledge of LLM. TRANS-ZERO combines Genetic
Monte-Carlo Tree Search (G-MCTS) with preference optimization, achieving strong
translation performance that rivals supervised methods. Experiments demonstrate
that this approach not only matches the performance of models trained on
large-scale parallel data but also excels in non-English translation
directions. Further analysis reveals that G-MCTS itself significantly enhances
translation quality by exploring semantically consistent candidates through
iterative translations, providing a robust foundation for the framework's
succuss.
| [
{
"version": "v1",
"created": "Sun, 20 Apr 2025 16:20:30 GMT"
}
] | 2025-04-22T00:00:00 | [
[
"Zou",
"Wei",
""
],
[
"Yang",
"Sen",
""
],
[
"Bao",
"Yu",
""
],
[
"Huang",
"Shujian",
""
],
[
"Chen",
"Jiajun",
""
],
[
"Cheng",
"Shanbo",
""
]
] |
2504.14670 | Tuan Anh Pham | Tuan Anh Pham | The orbit method for the Virasoro algebra | 47 pages | null | null | null | math.RA math.RT | http://arxiv.org/licenses/nonexclusive-distrib/1.0/ | Let $W = \mathbb{C}[t, t^{-1}]\partial_t$ be the Witt algebra of algebraic
vector fields on $\mathbb{C}^\times$ and let $V\!ir$ be the Virasoro algebra,
the unique nontrivial central extension of $W$. In 2023, Petukhov and Sierra
showed that Poisson primitive ideals of $\mathrm{S}(W)$ and $\mathrm{S}(V\!ir)$
can be constructed from elements of $W^*$ and $V\!ir^*$ of a particular form,
called local functions. In this paper, we show how to use a local function on
$W$ or $V\!ir$ to construct a representation of the Lie algebra. We further
show that the annihilators of these representations are new completely prime
primitive ideals of $\mathrm{U}(W)$ and $\mathrm{U}(V\!ir)$. We use this to
define a Dixmier map from the Poisson primitive spectrum of
$\mathrm{S}(V\!ir)$, respectively $\mathrm{S}(W)$, to the primitive spectrum of
$\mathrm{U}(V\!ir)$, respectively $\mathrm{U}(W)$, successfully extending the
orbit method from finite-dimensional solvable Lie algebras to our
countable-dimensional setting.
Our method involves new ring homomorphisms from $\mathrm{U}(W)$ to the tensor
product of a localized Weyl algebra and the enveloping algebra of a
finite-dimensional solvable subquotient of $W$. We further show that the
kernels of these homomorphisms are intersections of the primitive ideals
constructed from natural subsets of $W^*$. As a corollary, we disprove the
conjecture that any primitive ideal of $\mathrm{U}(W)$ is the kernel of some
map from $\mathrm{U}(W)$ to the first Weyl algebra.
| [
{
"version": "v1",
"created": "Sun, 20 Apr 2025 16:24:59 GMT"
}
] | 2025-04-22T00:00:00 | [
[
"Pham",
"Tuan Anh",
""
]
] |
2504.14671 | Debasish Borah | Debasish Borah, Devabrat Mahanta, Indrajit Saha | Gravitational wave signatures of dark sector portal leptogenesis | 23 pages, 6 captioned figures | null | null | null | hep-ph astro-ph.CO | http://arxiv.org/licenses/nonexclusive-distrib/1.0/ | We study the possibility of probing leptogenesis via stochastic gravitational
waves (GW) arising from a dark sector assisted first-order electroweak phase
transition. The same dark sector, with non-trivial transformation under an
unbroken $Z_2$ symmetry is also responsible for providing the only source of CP
asymmetry via one-loop interference with the tree level decay of a heavy
right-handed neutrino into lepton and Higgs doublets. The new Yukawa and scalar
portal couplings enhance the CP asymmetry allowing TeV scale leptogenesis
without any resonant enhancement. Light neutrino masses arise from a
combination of type-I and one-loop contributions with vanishing lightest
neutrino mass. While the new degrees of freedom in sub-TeV range keep the
detection prospects at terrestrial experiments promising, the new scalars
enhance the strength of the electroweak phase transition keeping the GW signals
within reach of near future experiments like LISA.
| [
{
"version": "v1",
"created": "Sun, 20 Apr 2025 16:25:47 GMT"
}
] | 2025-04-22T00:00:00 | [
[
"Borah",
"Debasish",
""
],
[
"Mahanta",
"Devabrat",
""
],
[
"Saha",
"Indrajit",
""
]
] |
2504.14672 | Anton Seleznev | Dmitry I. Tagaev and Anton F. Seleznev | Search for the possible members of the open cluster NGC 3532 with poor
astrometric solutions of Gaia DR3 | 14 pages, 14 figures, 1 table. Accepted to Astronomy Reports | null | null | null | astro-ph.GA | http://creativecommons.org/licenses/by-nc-nd/4.0/ | We performed star counts in the region of the open cluster NGC 3532. The
ranges of trigonometric parallaxes and proper motions containing all the stars
of the cluster were determined using the stars with 5- and 6-parameter Gaia DR3
solutions. The estimated radius of the cluster was R_c=178+-3 arcminutes and
the number of cluster stars was N_c=2200+-40. We estimate the number of stars
with poor astrometric solutions that may be members of the cluster. For this
purpose, we analyze the surface density distribution of stars with
two-parameter Gaia DR3 solutions, stars with the parameter RUWE>1.4, and stars
with large relative errors of trigonometric parallaxes in the vicinity of the
cluster. We are looking for stars that fall within the area of the
color-magnitude diagram occupied by probable members of the NGC 3532 cluster
from the Hant&Reffert sample. The radial surface density profile plotted with
such stars shows the concentration of stars toward the cluster center. An
analysis of the profile yields an estimate of 2150+-230 stars that may be
cluster members. Thus, nearly one half of cluster members can be lost when the
probable members are selected only by exact astrometric data of Gaia DR3. Among
these lost stars, there may be a significant number of unresolved binary and
multiple systems.
| [
{
"version": "v1",
"created": "Sun, 20 Apr 2025 16:31:19 GMT"
}
] | 2025-04-22T00:00:00 | [
[
"Tagaev",
"Dmitry I.",
""
],
[
"Seleznev",
"Anton F.",
""
]
] |
2504.14673 | Hoang Anh Tran | Hoang Anh Tran, Binh Tuan Nguyen, Yong Sheng Soh | Moment Sum-of-Squares Hierarchy for Gromov Wasserstein: Continuous
Extensions and Sample Complexity | null | null | null | null | math.OC | http://creativecommons.org/licenses/by-nc-nd/4.0/ | The Gromov-Wasserstein (GW) problem is an extension of the classical optimal
transport problem to settings where the source and target distributions reside
in incomparable spaces, and for which a cost function that attributes the price
of moving resources is not available. The sum-of-squares (SOS) hierarchy is a
principled method for deriving tractable semidefinite relaxations to generic
polynomial optimization problems. In this work, we apply ideas from the
moment-SOS hierarchy to solve the GW problem.
More precisely, we identify extensions of the moment-SOS hierarchy,
previously introduced for the discretized GW problem, such that they remain
valid for general probability distributions. This process requires a suitable
generalization of positive semidefiniteness over finite-dimensional vector
spaces to the space of probability distributions. We prove the following
properties concerning these continuous extensions: First, these relaxations
form a genuine hierarchy in that the optimal value converges to the GW
distance. Second, each of these relaxations induces a pseudo-metric over the
collection of metric measure spaces. Crucially, unlike the GW problem, these
induced instances are tractable to compute -- the discrete analogs are
expressible as semidefinite programs and hence are tractable to solve.
Separately from these properties, we also establish a statistical consistency
result arising from sampling the source and target distributions. Our work
suggests fascinating applications of the SOS hierarchy to optimization problems
over probability distributions in settings where the objective and constraint
depend on these distributions in a polynomial way.
| [
{
"version": "v1",
"created": "Sun, 20 Apr 2025 16:33:09 GMT"
}
] | 2025-04-22T00:00:00 | [
[
"Tran",
"Hoang Anh",
""
],
[
"Nguyen",
"Binh Tuan",
""
],
[
"Soh",
"Yong Sheng",
""
]
] |
2504.14674 | Mrinal Kanti Bose | Mrinal Kanti Bose, Udaya Parampalli and Abhay Kumar Singh | Binary cyclic codes from permutation polynomials over $\mathbb{F}_{2^m}$ | null | null | null | null | cs.IT math.IT | http://creativecommons.org/licenses/by/4.0/ | Binary cyclic codes having large dimensions and minimum distances close to
the square-root bound are highly valuable in applications where high-rate
transmission and robust error correction are both essential. They provide an
optimal trade-off between these two factors, making them suitable for demanding
communication and storage systems, post-quantum cryptography, radar and sonar
systems, wireless sensor networks, and space communications. This paper aims to
investigate cyclic codes by an efficient approach introduced by Ding
\cite{SETA5} from several known classes of permutation monomials and trinomials
over $\mathbb{F}_{2^m}$. We present several infinite families of binary cyclic
codes of length $2^m-1$ with dimensions larger than $(2^m-1)/2$. By applying
the Hartmann-Tzeng bound, some of the lower bounds on the minimum distances of
these cyclic codes are relatively close to the square root bound. Moreover, we
obtain a new infinite family of optimal binary cyclic codes with parameters
$[2^m-1,2^m-2-3m,8]$, where $m\geq 5$ is odd, according to the sphere-packing
bound.
| [
{
"version": "v1",
"created": "Sun, 20 Apr 2025 16:33:57 GMT"
}
] | 2025-04-22T00:00:00 | [
[
"Bose",
"Mrinal Kanti",
""
],
[
"Parampalli",
"Udaya",
""
],
[
"Singh",
"Abhay Kumar",
""
]
] |
2504.14675 | Tamoghna Ray | Tamoghna Ray, Abhishek Dhar, and Manas Kulkarni | Page curve like dynamics in Interacting Quantum Systems | 18 pages, 12 figures | null | null | null | quant-ph cond-mat.mes-hall cond-mat.quant-gas cond-mat.stat-mech hep-th | http://arxiv.org/licenses/nonexclusive-distrib/1.0/ | We study the dynamics of entanglement in a one-dimensional $XXZ$ spin-$1/2$
chain, with and without integrability-breaking interactions, that is connected
to a bath. We start from a state where the system and bath are completely
unentangled, and the bath is polarized spin-down. We consider two different
initial states for the system - (i) a polarized spin-up state, and (ii) an
infinite temperature state. In the particle representation of the spin chain,
the polarized spin-up state corresponds to a filled state, while the polarized
spin-down state corresponds to an empty state. Starting from these
inhomogeneous quenches, in all the above-mentioned cases we obtain the Page
curve like behavior in the entanglement. We report different power-law behavior
in the growth of entanglement for different initial states and different kinds
of baths (interacting and non-interacting). In an attempt to explore plausible
deep connections between entanglement and Boltzmann entropy, we investigate the
latter in both the filled and the infinite temperature case, for the system and
the bath. For the filled case, the Boltzmann entropy of the system has the form
of a Page curve but quantitatively deviates from the entanglement. On the other
hand, the entropy of the bath keeps increasing. Remarkably, for the infinite
temperature case, we find that the system and bath Boltzmann entropies agree
with the entanglement entropy, after and before the Page time, respectively.
Our findings are expected to hold for generic interacting quantum systems and
could be of relevance to black hole physics.
| [
{
"version": "v1",
"created": "Sun, 20 Apr 2025 16:36:38 GMT"
}
] | 2025-04-22T00:00:00 | [
[
"Ray",
"Tamoghna",
""
],
[
"Dhar",
"Abhishek",
""
],
[
"Kulkarni",
"Manas",
""
]
] |
2504.14676 | Felix Brandt | Felix Brandt, Arnab Roy | Analysis of stochastic fluid-rigid body dynamics: an approach by
stochastic maximal regularity | null | null | null | null | math.AP | http://creativecommons.org/licenses/by-sa/4.0/ | This paper investigates for the first time a rigorous mathematical framework
for a stochastic fluid-rigid body interaction problem. It is shown that the
coupled problem of a rigid ball immersed in a Newtonian fluid filling the
entire $\mathbb{R}^3$ and subject to additive noise in the fluid and rigid body
equation as well as transport noise in the fluid equation is locally strongly
well-posed. In addition, blow-up criteria for the stochastic fluid-rigid body
interaction problem are provided. For these results, the boundedness of the
$\mathcal{H}^\infty$-calculus of the so-called fluid-structure operator is
established by means of a decoupling approach, paving the way for stochastic
maximal regularity. The local strong well-posedness is then obtained by the
theory of stochastic evolution equations in weighted spaces.
| [
{
"version": "v1",
"created": "Sun, 20 Apr 2025 16:37:08 GMT"
}
] | 2025-04-22T00:00:00 | [
[
"Brandt",
"Felix",
""
],
[
"Roy",
"Arnab",
""
]
] |
2504.14677 | Jia Liu | Jia Liu, Cheng Jinguo, Xia Fang, Zhenyuan Ma, Yuankai Wu | Evaluating Temporal Plasticity in Foundation Time Series Models for
Incremental Fine-tuning | Accepted at IJCNN 2025 | null | null | null | cs.LG cs.AI | http://arxiv.org/licenses/nonexclusive-distrib/1.0/ | Time series foundation models excel at diverse time series forecasting tasks,
but their capacity for continuous improvement through incremental learning
remains unexplored. We present the first comprehensive study investigating
these models' temporal plasticity - their ability to progressively enhance
performance through continual learning while maintaining existing capabilities.
Through experiments on real-world datasets exhibiting distribution shifts, we
evaluate both conventional deep learning models and foundation models using a
novel continual learning framework. Our findings reveal that while traditional
models struggle with performance deterioration during incremental fine-tuning,
foundation models like Time-MoE and Chronos demonstrate sustained improvement
in predictive accuracy. This suggests that optimizing foundation model
fine-tuning strategies may be more valuable than developing domain-specific
small models. Our research introduces new evaluation methodologies and insights
for developing foundation time series models with robust continuous learning
capabilities.
| [
{
"version": "v1",
"created": "Sun, 20 Apr 2025 16:43:01 GMT"
}
] | 2025-04-22T00:00:00 | [
[
"Liu",
"Jia",
""
],
[
"Jinguo",
"Cheng",
""
],
[
"Fang",
"Xia",
""
],
[
"Ma",
"Zhenyuan",
""
],
[
"Wu",
"Yuankai",
""
]
] |
2504.14678 | Arturo Mart\'inez-Celis | Piotr Borodulin-Nadzieja, Arturo Mart\'inez-Celis, Adam Morawski,
Jadwiga \'Swierczy\'nska | On the Rudin-Blass Ordering of Measures | 26 pages | null | null | null | math.LO math.CA | http://arxiv.org/licenses/nonexclusive-distrib/1.0/ | We study the Rudin-Blass (and the Rudin-Keisler) ordering on the finite
additive measures on $\omega$. We propose a generalization of the notion of
Q-point and selective ultrafilter to measures: Q-measures and selective
measures. We show some symmetries between Q-points and Q-measures but also we
show where those symmetries break up. In particular we present an example of a
measure which is minimal in the sense of Rudin-Blass but which is not a
Q-measure.
| [
{
"version": "v1",
"created": "Sun, 20 Apr 2025 16:45:17 GMT"
}
] | 2025-04-22T00:00:00 | [
[
"Borodulin-Nadzieja",
"Piotr",
""
],
[
"Martínez-Celis",
"Arturo",
""
],
[
"Morawski",
"Adam",
""
],
[
"Świerczyńska",
"Jadwiga",
""
]
] |
2504.14679 | Mark Elin | Mark Elin | Nonlinear resolvents: distortion and order of starlikeness | null | null | null | null | math.CV | http://arxiv.org/licenses/nonexclusive-distrib/1.0/ | This paper presents a new approach to studying nonlinear resolvents of
holomorphically accretive mappings on the open unit ball of a complex Banach
space. We establish a distortion theorem and apply it to address problems in
geometric function theory concerning the class of resolvents. Specifically, we
prove the accretivity of resolvents and provide estimates for the squeezing
ratio. Further, we show that nonlinear resolvents are starlike mappings of
certain order and determine lower bounds for this order.
| [
{
"version": "v1",
"created": "Sun, 20 Apr 2025 16:48:11 GMT"
}
] | 2025-04-22T00:00:00 | [
[
"Elin",
"Mark",
""
]
] |
2504.14681 | Zeyu Wang | Zeyu Wang, Frank P.-W. Lo, Qian Chen, Yongqi Zhang, Chen Lin, Xu Chen,
Zhenhua Yu, Alexander J. Thompson, Eric M. Yeatman and Benny P. L. Lo | An LLM-enabled Multi-Agent Autonomous Mechatronics Design Framework | Accepted by CVPR 2025 Workshop | null | null | null | cs.RO cs.AI | http://arxiv.org/licenses/nonexclusive-distrib/1.0/ | Existing LLM-enabled multi-agent frameworks are predominantly limited to
digital or simulated environments and confined to narrowly focused knowledge
domain, constraining their applicability to complex engineering tasks that
require the design of physical embodiment, cross-disciplinary integration, and
constraint-aware reasoning. This work proposes a multi-agent autonomous
mechatronics design framework, integrating expertise across mechanical design,
optimization, electronics, and software engineering to autonomously generate
functional prototypes with minimal direct human design input. Operating
primarily through a language-driven workflow, the framework incorporates
structured human feedback to ensure robust performance under real-world
constraints. To validate its capabilities, the framework is applied to a
real-world challenge involving autonomous water-quality monitoring and
sampling, where traditional methods are labor-intensive and ecologically
disruptive. Leveraging the proposed system, a fully functional autonomous
vessel was developed with optimized propulsion, cost-effective electronics, and
advanced control. The design process was carried out by specialized agents,
including a high-level planning agent responsible for problem abstraction and
dedicated agents for structural, electronics, control, and software
development. This approach demonstrates the potential of LLM-based multi-agent
systems to automate real-world engineering workflows and reduce reliance on
extensive domain expertise.
| [
{
"version": "v1",
"created": "Sun, 20 Apr 2025 16:57:45 GMT"
}
] | 2025-04-22T00:00:00 | [
[
"Wang",
"Zeyu",
""
],
[
"Lo",
"Frank P. -W.",
""
],
[
"Chen",
"Qian",
""
],
[
"Zhang",
"Yongqi",
""
],
[
"Lin",
"Chen",
""
],
[
"Chen",
"Xu",
""
],
[
"Yu",
"Zhenhua",
""
],
[
"Thompson",
"Alexander J.",
""
],
[
"Yeatman",
"Eric M.",
""
],
[
"Lo",
"Benny P. L.",
""
]
] |
2504.14682 | Yang Sun | Yang Sun, Zhiyuan Ji, George H. Rieke, Francesco D'Eugenio, Yongda
Zhu, Fengwu Sun, Xiaojing Lin, Andrew J. Bunker, Jianwei Lyu, Pierluigi
Rinaldi, Christopher N. A. Willmer | Extreme Neutral Outflow in an Inactive Quenching Galaxy at z$\sim$1.3 | 21 pages, 11 figure, submitted to ApJ, comments are welcome | null | null | null | astro-ph.GA | http://creativecommons.org/licenses/by/4.0/ | We have discovered a substantial sodium doublet (Na D $\lambda\lambda$5890,
5896\AA)-traced neutral outflow in a quenching galaxy JADES-GS-206183 at
$z=1.317$ in GOODS-S field. Its JWST NIRSpec/MSA spectrum shows a significantly
blueshifted and deep Na D absorption, revealing a neutral outflow with a
velocity of $v_{\rm out}=828^{+79}_{-49}\,\mathrm{km\,s^{-1}}$ and a mass
outflow rate of $\log(\dot{M}_{\rm
out}/\mathrm{M_{\odot}\,yr^{-1}})=2.40^{+0.11}_{-0.16}$. The mass outflow rate
of this outflow is higher than any of the neutral outflows identified
previously beyond $z\sim1$ by the same line diagnostic and is comparable with
those in local galaxies with extremely strong star formation activities or
luminous AGN. Nonetheless, the best-fit SED modeling of JADES-GS-206183, based
on its multi-band photometry from HST/ACS to JWST/NIRCam, suggests that the
host galaxy now is quenched, and the Paschen $\alpha$ (Pa$\alpha$) emission in
the FRESCO NIRCam grism spectrum confirms its current low star formation rate
($10.78\pm 0.55\,\mathrm{M_{\odot}\,yr^{-1}}$). More surprisingly, optical line
ratio diagnostics indicate that the current AGN activity of JADES-GS-206183, if
present, is weak. Even though we tentatively detect a broad component of the
H$\alpha$ line, it is more likely tracing the ionized outflow than an AGN. The
results demonstrate that the Na D outflow in JADES-GS-206183 is highly unlikely
to be driven by current star formation or nuclear activity. Instead, we propose
that the outflow that we are witnessing in JADES-GS-206183 may be a
long-lasting fossil outflow, powered by previous AGN activity that has recently
shut down.
| [
{
"version": "v1",
"created": "Sun, 20 Apr 2025 17:01:38 GMT"
}
] | 2025-04-22T00:00:00 | [
[
"Sun",
"Yang",
""
],
[
"Ji",
"Zhiyuan",
""
],
[
"Rieke",
"George H.",
""
],
[
"D'Eugenio",
"Francesco",
""
],
[
"Zhu",
"Yongda",
""
],
[
"Sun",
"Fengwu",
""
],
[
"Lin",
"Xiaojing",
""
],
[
"Bunker",
"Andrew J.",
""
],
[
"Lyu",
"Jianwei",
""
],
[
"Rinaldi",
"Pierluigi",
""
],
[
"Willmer",
"Christopher N. A.",
""
]
] |
2504.14683 | Sina Bagheri Nezhad | Sina Bagheri Nezhad, Sayan Bandyapadhyay, Tianzhi Chen | Polynomial-Time Constant-Approximation for Fair Sum-of-Radii Clustering | null | null | null | null | cs.DS | http://creativecommons.org/licenses/by/4.0/ | In a seminal work, Chierichetti et al. introduced the $(t,k)$-fair clustering
problem: Given a set of red points and a set of blue points in a metric space,
a clustering is called fair if the number of red points in each cluster is at
most $t$ times and at least $1/t$ times the number of blue points in that
cluster. The goal is to compute a fair clustering with at most $k$ clusters
that optimizes certain objective function. Considering this problem, they
designed a polynomial-time $O(1)$- and $O(t)$-approximation for the $k$-center
and the $k$-median objective, respectively. Recently, Carta et al. studied this
problem with the sum-of-radii objective and obtained a
$(6+\epsilon)$-approximation with running time
$O((k\log_{1+\epsilon}(k/\epsilon))^kn^{O(1)})$, i.e., fixed-parameter
tractable in $k$. Here $n$ is the input size. In this work, we design the first
polynomial-time $O(1)$-approximation for $(t,k)$-fair clustering with the
sum-of-radii objective, improving the result of Carta et al. Our result places
sum-of-radii in the same group of objectives as $k$-center, that admit
polynomial-time $O(1)$-approximations. This result also implies a
polynomial-time $O(1)$-approximation for the Euclidean version of the problem,
for which an $f(k)\cdot n^{O(1)}$-time $(1+\epsilon)$-approximation was known
due to Drexler et al.. Here $f$ is an exponential function of $k$. We are also
able to extend our result to any arbitrary $\ell\ge 2$ number of colors when
$t=1$. This matches known results for the $k$-center and $k$-median objectives
in this case. The significant disparity of sum-of-radii compared to $k$-center
and $k$-median presents several complex challenges, all of which we
successfully overcome in our work. Our main contribution is a novel
cluster-merging-based analysis technique for sum-of-radii that helps us achieve
the constant-approximation bounds.
| [
{
"version": "v1",
"created": "Sun, 20 Apr 2025 17:19:03 GMT"
}
] | 2025-04-22T00:00:00 | [
[
"Nezhad",
"Sina Bagheri",
""
],
[
"Bandyapadhyay",
"Sayan",
""
],
[
"Chen",
"Tianzhi",
""
]
] |
2504.14684 | Dipendra Prasad | Santosh Nadimpalli, Santosha Pattanayak, Dipendra Prasad | Character theory at a torsion element | null | null | null | null | math.RT | http://arxiv.org/licenses/nonexclusive-distrib/1.0/ | The paper relates character value of an irreducible representation of a
compact connected Lie group at certain elements of finite order with the
dimension of a representation on another group, up to some precise constants,
which all have significance. An important input is to analyse torsion elements
of order d in an adjoint group with minimal dimensional centraliser, and to
prove that in most cases when d divides the Coxeter number of G, this gives
rise to a unique conjugacy class.
| [
{
"version": "v1",
"created": "Sun, 20 Apr 2025 17:20:05 GMT"
}
] | 2025-04-22T00:00:00 | [
[
"Nadimpalli",
"Santosh",
""
],
[
"Pattanayak",
"Santosha",
""
],
[
"Prasad",
"Dipendra",
""
]
] |
2504.14686 | Jos\'e Su\'arez-Varela | Jos\'e Su\'arez-Varela and Andra Lutu | Uncovering Issues in the Radio Access Network by Looking at the
Neighbors | 7 pages | IEEE International Conference on Machine Learning for
Communication and Networking, 2025 | null | null | cs.NI cs.AI cs.LG | http://creativecommons.org/licenses/by/4.0/ | Mobile network operators (MNOs) manage Radio Access Networks (RANs) with
massive amounts of cells over multiple radio generations (2G-5G). To handle
such complexity, operations teams rely on monitoring systems, including anomaly
detection tools that identify unexpected behaviors. In this paper, we present
c-ANEMON, a Contextual ANomaly dEtection MONitor for the RAN based on Graph
Neural Networks (GNNs). Our solution captures spatio-temporal variations by
analyzing the behavior of individual cells in relation to their local
neighborhoods, enabling the detection of anomalies that are independent of
external mobility factors. This, in turn, allows focusing on anomalies
associated with network issues (e.g., misconfigurations, equipment failures).
We evaluate c-ANEMON using real-world data from a large European metropolitan
area (7,890 cells; 3 months). First, we show that the GNN model within our
solution generalizes effectively to cells from previously unseen areas,
suggesting the possibility of using a single model across extensive deployment
regions. Then, we analyze the anomalies detected by c-ANEMON through manual
inspection and define several categories of long-lasting anomalies (6+ hours).
Notably, 45.95% of these anomalies fall into a category that is more likely to
require intervention by operations teams.
| [
{
"version": "v1",
"created": "Sun, 20 Apr 2025 17:36:52 GMT"
}
] | 2025-04-22T00:00:00 | [
[
"Suárez-Varela",
"José",
""
],
[
"Lutu",
"Andra",
""
]
] |
2504.14687 | Seokju Cho | Seokju Cho and Jiahui Huang and Seungryong Kim and Joon-Young Lee | Seurat: From Moving Points to Depth | CVPR 2025 Highlight. Project page: https://seurat-cvpr.github.io | null | null | null | cs.CV | http://creativecommons.org/licenses/by/4.0/ | Accurate depth estimation from monocular videos remains challenging due to
ambiguities inherent in single-view geometry, as crucial depth cues like
stereopsis are absent. However, humans often perceive relative depth
intuitively by observing variations in the size and spacing of objects as they
move. Inspired by this, we propose a novel method that infers relative depth by
examining the spatial relationships and temporal evolution of a set of tracked
2D trajectories. Specifically, we use off-the-shelf point tracking models to
capture 2D trajectories. Then, our approach employs spatial and temporal
transformers to process these trajectories and directly infer depth changes
over time. Evaluated on the TAPVid-3D benchmark, our method demonstrates robust
zero-shot performance, generalizing effectively from synthetic to real-world
datasets. Results indicate that our approach achieves temporally smooth,
high-accuracy depth predictions across diverse domains.
| [
{
"version": "v1",
"created": "Sun, 20 Apr 2025 17:37:02 GMT"
}
] | 2025-04-22T00:00:00 | [
[
"Cho",
"Seokju",
""
],
[
"Huang",
"Jiahui",
""
],
[
"Kim",
"Seungryong",
""
],
[
"Lee",
"Joon-Young",
""
]
] |
2504.14688 | Michiki Taniwaki | Michiki Taniwaki and Thomas Brian Winkler and Jan Roth\"orl and
Raphael Gruber and Chiharu Mitsumata and Masato Kotsugi and Mathias Kl\"aui | Persistent Homology-Based Descriptor of Topological Ordering in
Two-Dimensional Quasi-Particle Systems with Application to Skyrmion Lattices | 18 pages, 6 figures, 1 table, 1 supplemental table, and 3
supplemental figures | null | null | null | cond-mat.stat-mech | http://arxiv.org/licenses/nonexclusive-distrib/1.0/ | Two-dimensional (2D) quasi-particles systems, such as magnetic
skyrmions, can exhibit a rich variety of topological phase transitions.
However, the methodology for capturing the configurational properties of
the lattice ordering and constructing an appropriate descriptor that can
be easily calculated is not obvious. Here, we present a topological
descriptor, "persistent diagram", and propose an indicator for
topological phase transitions using persistent homology (PH). PH offers
novel insights beyond conventional indicators by capturing topological
features derived from the configurational properties of the lattice. The
proposed persistent-homology-based topological indicator, which
selectively counts stable features in the persistence diagram,
effectively traces the lattice's topological changes, as confirmed by
comparisons with the conventionally used measure of the ordering
$\langle|\Psi_6|\rangle$, typically used
to identify lattice phases. While our method is demonstrated in the
context of skyrmion lattice systems, the approach is general and can be
extended to other two-dimensional systems composed of repulsively
interacting quasi-particles. Moreover, our indicator offers lower
computational complexity than the conventionally used methods.
| [
{
"version": "v1",
"created": "Sun, 20 Apr 2025 17:38:38 GMT"
}
] | 2025-04-22T00:00:00 | [
[
"Taniwaki",
"Michiki",
""
],
[
"Winkler",
"Thomas Brian",
""
],
[
"Rothörl",
"Jan",
""
],
[
"Gruber",
"Raphael",
""
],
[
"Mitsumata",
"Chiharu",
""
],
[
"Kotsugi",
"Masato",
""
],
[
"Kläui",
"Mathias",
""
]
] |
2504.14689 | Katelyn Xiaoying Mei | Katelyn Xiaoying Mei, and Nic Weber | Designing AI Systems that Augment Human Performed vs. Demonstrated
Critical Thinking | Presented at the 2025 ACM Workshop on Human-AI Interaction for
Augmented Reasoning, | Proceedings of the 2025 ACM CHI Workshop on Human-AI Interaction
for Augmented Reasoning | null | CHI25-WS-AUGMENTED-REASONING | cs.HC | http://creativecommons.org/licenses/by-nc-sa/4.0/ | The recent rapid advancement of LLM-based AI systems has accelerated our
search and production of information. While the advantages brought by these
systems seemingly improve the performance or efficiency of human activities,
they do not necessarily enhance human capabilities. Recent research has started
to examine the impact of generative AI on individuals' cognitive abilities,
especially critical thinking. Based on definitions of critical thinking across
psychology and education, this position paper proposes the distinction between
demonstrated and performed critical thinking in the era of generative AI and
discusses the implication of this distinction in research and development of AI
systems that aim to augment human critical thinking.
| [
{
"version": "v1",
"created": "Sun, 20 Apr 2025 17:40:28 GMT"
}
] | 2025-04-22T00:00:00 | [
[
"Mei",
"Katelyn Xiaoying",
""
],
[
"Weber",
"Nic",
""
]
] |
2504.14690 | Mehrnoush ShamsFard | Mehrnoush Shamsfard, Zahra Saaberi, Mostafa Karimi manesh, Seyed
Mohammad Hossein Hashemi, Zahra Vatankhah, Motahareh Ramezani, Niki Pourazin,
Tara Zare, Maryam Azimi, Sarina Chitsaz, Sama Khoraminejad, Morteza Mahdavi
Mortazavi, Mohammad Mahdi Chizari, Sahar Maleki, Seyed Soroush Majd, Mostafa
Masumi, Sayed Ali Musavi Khoeini, Amir Mohseni, Sogol Alipour | FarsEval-PKBETS: A new diverse benchmark for evaluating Persian large
language models | 24 pages, 3 figures, 3 tables | null | null | null | cs.CL cs.AI | http://creativecommons.org/publicdomain/zero/1.0/ | Research on evaluating and analyzing large language models (LLMs) has been
extensive for resource-rich languages such as English, yet their performance in
languages such as Persian has received considerably less attention. This paper
introduces FarsEval-PKBETS benchmark, a subset of FarsEval project for
evaluating large language models in Persian. This benchmark consists of 4000
questions and answers in various formats, including multiple choice, short
answer and descriptive responses. It covers a wide range of domains and
tasks,including medicine, law, religion, Persian language, encyclopedic
knowledge, human preferences, social knowledge, ethics and bias, text
generation, and respecting others' rights. This bechmark incorporates
linguistics, cultural, and local considerations relevant to the Persian
language and Iran. To ensure the questions are challenging for current LLMs,
three models -- Llama3-70B, PersianMind, and Dorna -- were evaluated using this
benchmark. Their average accuracy was below 50%, meaning they provided fully
correct answers to fewer than half of the questions. These results indicate
that current language models are still far from being able to solve this
benchmark
| [
{
"version": "v1",
"created": "Sun, 20 Apr 2025 17:43:47 GMT"
}
] | 2025-04-22T00:00:00 | [
[
"Shamsfard",
"Mehrnoush",
""
],
[
"Saaberi",
"Zahra",
""
],
[
"manesh",
"Mostafa Karimi",
""
],
[
"Hashemi",
"Seyed Mohammad Hossein",
""
],
[
"Vatankhah",
"Zahra",
""
],
[
"Ramezani",
"Motahareh",
""
],
[
"Pourazin",
"Niki",
""
],
[
"Zare",
"Tara",
""
],
[
"Azimi",
"Maryam",
""
],
[
"Chitsaz",
"Sarina",
""
],
[
"Khoraminejad",
"Sama",
""
],
[
"Mortazavi",
"Morteza Mahdavi",
""
],
[
"Chizari",
"Mohammad Mahdi",
""
],
[
"Maleki",
"Sahar",
""
],
[
"Majd",
"Seyed Soroush",
""
],
[
"Masumi",
"Mostafa",
""
],
[
"Khoeini",
"Sayed Ali Musavi",
""
],
[
"Mohseni",
"Amir",
""
],
[
"Alipour",
"Sogol",
""
]
] |
2504.14691 | Breno Ferraz de Oliveira | D. Bazeia and M. Bongestab and M. J. B. Ferreira and B. F. de Oliveira
and J. E. B. Santos | Quantifying scale-free behaviors in Rock-Paper-Scissors Models as a
function of Mobility | 7 pages, 6 figures, Version to appear in Physica A: Statistical
Mechanics and its Applications | null | null | null | q-bio.PE | http://creativecommons.org/licenses/by-sa/4.0/ | We investigate the scale-free behavior of the spatial rock-paper-scissors
model with May-Leonard dynamics, analyzing specific quantifiers that engender
the power-law feature. The main results show that an important parameter that
drives the scale-free behavior is the mobility, which can be used to
quantitatively describe several scale-free aspects of the model, such as the
number of clusters, the characteristic length, the individuals' lifespan and
its corresponding mean traveled distance. All of these are novel quantifiers of
current practical interest for the study of biodiversity.
| [
{
"version": "v1",
"created": "Sun, 20 Apr 2025 17:43:54 GMT"
}
] | 2025-04-22T00:00:00 | [
[
"Bazeia",
"D.",
""
],
[
"Bongestab",
"M.",
""
],
[
"Ferreira",
"M. J. B.",
""
],
[
"de Oliveira",
"B. F.",
""
],
[
"Santos",
"J. E. B.",
""
]
] |
2504.14692 | Songtao Jiang | Songtao Jiang, Yuan Wang, Sibo Song, Yan Zhang, Zijie Meng, Bohan Lei,
Jian Wu, Jimeng Sun, Zuozhu Liu | OmniV-Med: Scaling Medical Vision-Language Model for Universal Visual
Understanding | null | null | null | null | cs.CL | http://arxiv.org/licenses/nonexclusive-distrib/1.0/ | The practical deployment of medical vision-language models (Med-VLMs)
necessitates seamless integration of textual data with diverse visual
modalities, including 2D/3D images and videos, yet existing models typically
employ separate encoders for different modalities. To address this limitation,
we present OmniV-Med, a unified framework for multimodal medical understanding.
Our technical contributions are threefold: First, we construct
OmniV-Med-Instruct, a comprehensive multimodal medical dataset containing 252K
instructional samples spanning 14 medical image modalities and 11 clinical
tasks. Second, we devise a rotary position-adaptive encoder that processes
multi-resolution 2D/3D images and videos within a unified architecture,
diverging from conventional modality-specific encoders. Third, we introduce a
medical-aware token pruning mechanism that exploits spatial-temporal redundancy
in volumetric data (e.g., consecutive CT slices) and medical videos,
effectively reducing 60\% of visual tokens without performance degradation.
Empirical evaluations demonstrate that OmniV-Med-7B achieves state-of-the-art
performance on 7 benchmarks spanning 2D/3D medical imaging and video
understanding tasks. Notably, our lightweight variant (OmniV-Med-1.5B) attains
comparable performance while requiring only 8 RTX3090 GPUs for training and
supporting efficient long-video inference. Data, code and model will be
released.
| [
{
"version": "v1",
"created": "Sun, 20 Apr 2025 17:53:56 GMT"
}
] | 2025-04-22T00:00:00 | [
[
"Jiang",
"Songtao",
""
],
[
"Wang",
"Yuan",
""
],
[
"Song",
"Sibo",
""
],
[
"Zhang",
"Yan",
""
],
[
"Meng",
"Zijie",
""
],
[
"Lei",
"Bohan",
""
],
[
"Wu",
"Jian",
""
],
[
"Sun",
"Jimeng",
""
],
[
"Liu",
"Zuozhu",
""
]
] |
2504.14693 | Enxin Song | Enxin Song, Wenhao Chai, Weili Xu, Jianwen Xie, Yuxuan Liu, and Gaoang
Wang | Video-MMLU: A Massive Multi-Discipline Lecture Understanding Benchmark | Code, docs, and benchmark are all avaliable at
https://enxinsong.com/Video-MMLU-web/ | null | null | null | cs.CV cs.AI | http://arxiv.org/licenses/nonexclusive-distrib/1.0/ | Recent advancements in language multimodal models (LMMs) for video have
demonstrated their potential for understanding video content, yet the task of
comprehending multi-discipline lectures remains largely unexplored. We
introduce Video-MMLU, a massive benchmark designed to evaluate the capabilities
of LMMs in understanding Multi-Discipline Lectures. We evaluate over 90
open-source and proprietary models, ranging from 0.5B to 40B parameters. Our
results highlight the limitations of current models in addressing the cognitive
challenges presented by these lectures, especially in tasks requiring both
perception and reasoning. Additionally, we explore how the number of visual
tokens and the large language models influence performance, offering insights
into the interplay between multimodal perception and reasoning in lecture
comprehension.
| [
{
"version": "v1",
"created": "Sun, 20 Apr 2025 17:58:46 GMT"
}
] | 2025-04-22T00:00:00 | [
[
"Song",
"Enxin",
""
],
[
"Chai",
"Wenhao",
""
],
[
"Xu",
"Weili",
""
],
[
"Xie",
"Jianwen",
""
],
[
"Liu",
"Yuxuan",
""
],
[
"Wang",
"Gaoang",
""
]
] |
2504.14694 | Yi-Hua Huang | Yuting He and Yiqiang Chen and XiaoDong Yang and Hanchao Yu and Yi-Hua
Huang and Yang Gu | Learning Critically: Selective Self Distillation in Federated Learning
on Non-IID Data | null | null | null | null | cs.LG cs.AI | http://arxiv.org/licenses/nonexclusive-distrib/1.0/ | Federated learning (FL) enables multiple clients to collaboratively train a
global model while keeping local data decentralized. Data heterogeneity
(non-IID) across clients has imposed significant challenges to FL, which makes
local models re-optimize towards their own local optima and forget the global
knowledge, resulting in performance degradation and convergence slowdown. Many
existing works have attempted to address the non-IID issue by adding an extra
global-model-based regularizing item to the local training but without an
adaption scheme, which is not efficient enough to achieve high performance with
deep learning models. In this paper, we propose a Selective Self-Distillation
method for Federated learning (FedSSD), which imposes adaptive constraints on
the local updates by self-distilling the global model's knowledge and
selectively weighting it by evaluating the credibility at both the class and
sample level. The convergence guarantee of FedSSD is theoretically analyzed and
extensive experiments are conducted on three public benchmark datasets, which
demonstrates that FedSSD achieves better generalization and robustness in fewer
communication rounds, compared with other state-of-the-art FL methods.
| [
{
"version": "v1",
"created": "Sun, 20 Apr 2025 18:06:55 GMT"
}
] | 2025-04-22T00:00:00 | [
[
"He",
"Yuting",
""
],
[
"Chen",
"Yiqiang",
""
],
[
"Yang",
"XiaoDong",
""
],
[
"Yu",
"Hanchao",
""
],
[
"Huang",
"Yi-Hua",
""
],
[
"Gu",
"Yang",
""
]
] |
2504.14695 | Weirui Peng | Weirui Peng, Yinuo Yang, Zheng Zhang, Toby Jia-Jun Li | GLITTER: An AI-assisted Platform for Material-Grounded Asynchronous
Discussion in Flipped Learning | null | null | null | null | cs.HC | http://creativecommons.org/licenses/by/4.0/ | Flipped classrooms promote active learning by having students engage with
materials independently before class, allowing in-class time for collaborative
problem-solving. During this pre-class phase, asynchronous online discussions
help students build knowledge and clarify concepts with peers. However, it
remains difficult to engage with temporally dispersed peer contributions,
connect discussions with static learning materials, and prepare for in-class
sessions based on their self-learning outcome. Our formative study identified
cognitive challenges students encounter, including navigation barriers,
reflection gaps, and contribution difficulty and anxiety. We present GLITTER,
an AI-assisted discussion platform for pre-class learning in flipped
classrooms. GLITTER helps students identify posts with shared conceptual
dimensions, scaffold knowledge integration through conceptual blending, and
enhance metacognition via personalized reflection reports. A lab study within
subjects (n = 12) demonstrates that GLITTER improves discussion engagement,
sparks new ideas, supports reflection, and increases preparedness for in-class
activities.
| [
{
"version": "v1",
"created": "Sun, 20 Apr 2025 18:06:55 GMT"
}
] | 2025-04-22T00:00:00 | [
[
"Peng",
"Weirui",
""
],
[
"Yang",
"Yinuo",
""
],
[
"Zhang",
"Zheng",
""
],
[
"Li",
"Toby Jia-Jun",
""
]
] |
2504.14696 | Naima Tasnim | Naima Tasnim, Atefeh Gilani, Lalitha Sankar and Oliver Kosut | Reveal-or-Obscure: A Differentially Private Sampling Algorithm for
Discrete Distributions | 8 pages, 3 figures | null | null | null | cs.IT cs.CR cs.DS cs.LG math.IT | http://creativecommons.org/licenses/by/4.0/ | We introduce a differentially private (DP) algorithm called reveal-or-obscure
(ROO) to generate a single representative sample from a dataset of $n$
observations drawn i.i.d. from an unknown discrete distribution $P$. Unlike
methods that add explicit noise to the estimated empirical distribution, ROO
achieves $\epsilon$-differential privacy by randomly choosing whether to
"reveal" or "obscure" the empirical distribution. While ROO is structurally
identical to Algorithm 1 proposed by Cheu and Nayak (arXiv:2412.10512), we
prove a strictly better bound on the sampling complexity than that established
in Theorem 12 of (arXiv:2412.10512). To further improve the privacy-utility
trade-off, we propose a novel generalized sampling algorithm called
Data-Specific ROO (DS-ROO), where the probability of obscuring the empirical
distribution of the dataset is chosen adaptively. We prove that DS-ROO
satisfies $\epsilon$-DP, and provide empirical evidence that DS-ROO can achieve
better utility under the same privacy budget of vanilla ROO.
| [
{
"version": "v1",
"created": "Sun, 20 Apr 2025 18:20:11 GMT"
}
] | 2025-04-22T00:00:00 | [
[
"Tasnim",
"Naima",
""
],
[
"Gilani",
"Atefeh",
""
],
[
"Sankar",
"Lalitha",
""
],
[
"Kosut",
"Oliver",
""
]
] |
2504.14697 | Zhengjiang Lin | Shi Chen, Zhengjiang Lin, Yury Polyanskiy, Philippe Rigollet | Quantitative Clustering in Mean-Field Transformer Models | 47 pages, 4 figures | null | null | null | cs.LG math.AP math.DS stat.ML | http://arxiv.org/licenses/nonexclusive-distrib/1.0/ | The evolution of tokens through a deep transformer models can be modeled as
an interacting particle system that has been shown to exhibit an asymptotic
clustering behavior akin to the synchronization phenomenon in Kuramoto models.
In this work, we investigate the long-time clustering of mean-field transformer
models. More precisely, we establish exponential rates of contraction to a
Dirac point mass for any suitably regular initialization under some assumptions
on the parameters of transformer models, any suitably regular mean-field
initialization synchronizes exponentially fast with some quantitative rates.
| [
{
"version": "v1",
"created": "Sun, 20 Apr 2025 18:21:34 GMT"
}
] | 2025-04-22T00:00:00 | [
[
"Chen",
"Shi",
""
],
[
"Lin",
"Zhengjiang",
""
],
[
"Polyanskiy",
"Yury",
""
],
[
"Rigollet",
"Philippe",
""
]
] |
2504.14698 | Svetlana Kotochigova | Jacek K{\l}os, Eite Tiesinga, Lan Cheng, Svetlana Kotochigova | Unconventional Chemical Bonding of Lanthanide-OH Molecules | 5 figures | null | null | null | physics.atom-ph quant-ph | http://creativecommons.org/licenses/by-nc-sa/4.0/ | We present a theoretical study of the low lying adiabatic relativistic
electronic states of lanthanide monohydroxide (Ln-OH) molecules near their
linear equilibrium geometries. We focus on heavy, magnetic DyOH and ErOH
relevant to fundamental symmetry tests. We use a restricted-active-space
self-consistent field method combined with spin-orbit coupling as well as a
relativistic coupled-cluster method. In addition, electric dipole and magnetic
moments are computed with the self-consistent field method. Analysis of the
results from both methods shows that the dominant molecular configuration of
the ground state is one where an electron from the partially filled and
submerged 4f orbital of the lanthanide atom moves to the hydroxyl group,
leaving the closed outer-most 6s$^2$ lone electron pair of the lanthanide atom
intact in sharp contrast to the bonding in alkaline-earth monohydroxides and
YbOH, where an electron from the outer-most s shell moves to the hydroxyl
group. For linear molecules the projection of the total electron angular
momentum on the symmetry axis is a conserved quantity with quantum number
$\Omega$ and we study the polynomial $\Omega$ dependence of the energies of the
ground states as well as their electric and magnetic moments. We find that the
lowest energy states have $|\Omega|=15/2$ and 1/2 for DyOH and ErOH,
respectively. The zero field splittings among these $\Omega$ states is
approximately $hc\times 1\,000$~cm$^{-1}$. We find that the permanent dipole
moments for both triatomics are fairly small at 0.23 atomic units. The magnetic
moments are closely related to that of the corresponding atomic Ln$^+$ ion in
an excited electronic state. We also realize that the total electron angular
momentum is to good approximation conserved and has a quantum number of 15/2
for both triatomic molecules.
| [
{
"version": "v1",
"created": "Sun, 20 Apr 2025 18:22:03 GMT"
}
] | 2025-04-22T00:00:00 | [
[
"Kłos",
"Jacek",
""
],
[
"Tiesinga",
"Eite",
""
],
[
"Cheng",
"Lan",
""
],
[
"Kotochigova",
"Svetlana",
""
]
] |
2504.14699 | Sascha Jecklin | Sascha Jecklin, Aidana Massalimova, Ruyi Zha, Lilian Calvet, Christoph
J. Laux, Mazda Farshad, and Philipp F\"urnstahl | IXGS-Intraoperative 3D Reconstruction from Sparse, Arbitrarily Posed
Real X-rays | null | null | null | null | cs.CV cs.AI | http://creativecommons.org/licenses/by-nc-nd/4.0/ | Spine surgery is a high-risk intervention demanding precise execution, often
supported by image-based navigation systems. Recently, supervised learning
approaches have gained attention for reconstructing 3D spinal anatomy from
sparse fluoroscopic data, significantly reducing reliance on
radiation-intensive 3D imaging systems. However, these methods typically
require large amounts of annotated training data and may struggle to generalize
across varying patient anatomies or imaging conditions. Instance-learning
approaches like Gaussian splatting could offer an alternative by avoiding
extensive annotation requirements. While Gaussian splatting has shown promise
for novel view synthesis, its application to sparse, arbitrarily posed real
intraoperative X-rays has remained largely unexplored. This work addresses this
limitation by extending the $R^2$-Gaussian splatting framework to reconstruct
anatomically consistent 3D volumes under these challenging conditions. We
introduce an anatomy-guided radiographic standardization step using style
transfer, improving visual consistency across views, and enhancing
reconstruction quality. Notably, our framework requires no pretraining, making
it inherently adaptable to new patients and anatomies. We evaluated our
approach using an ex-vivo dataset. Expert surgical evaluation confirmed the
clinical utility of the 3D reconstructions for navigation, especially when
using 20 to 30 views, and highlighted the standardization's benefit for
anatomical clarity. Benchmarking via quantitative 2D metrics (PSNR/SSIM)
confirmed performance trade-offs compared to idealized settings, but also
validated the improvement gained from standardization over raw inputs. This
work demonstrates the feasibility of instance-based volumetric reconstruction
from arbitrary sparse-view X-rays, advancing intraoperative 3D imaging for
surgical navigation.
| [
{
"version": "v1",
"created": "Sun, 20 Apr 2025 18:28:13 GMT"
}
] | 2025-04-22T00:00:00 | [
[
"Jecklin",
"Sascha",
""
],
[
"Massalimova",
"Aidana",
""
],
[
"Zha",
"Ruyi",
""
],
[
"Calvet",
"Lilian",
""
],
[
"Laux",
"Christoph J.",
""
],
[
"Farshad",
"Mazda",
""
],
[
"Fürnstahl",
"Philipp",
""
]
] |
2504.14700 | Anderson Rodriguez | Anderson M. Rodriguez | Curvature-Induced Saturation in Catalytic Reaction Networks: A
Differential Geometrical Framework for Modeling Chemical Complexity | 10 pages | null | null | null | q-bio.MN | http://creativecommons.org/licenses/by/4.0/ | The evolution of chemical reaction networks is often analyzed through kinetic
models and energy landscapes, but these approaches fail to capture the deeper
structural constraints governing complexity growth. In chemical reaction
networks, emergent constraints dictate the organization of reaction pathways,
limiting combinatorial expansion and determining stability conditions. This
paper introduces a novel approach to modeling chemical reaction networks by
incorporating differential geometry into the classical framework of reaction
kinetics. By utilizing the Riemannian metric, Christoffel symbols, and a
system-specific entropy-like term, we provide a new method for understanding
the evolution of complex reaction systems. The approach captures the
interdependence between species, the curvature of the reaction network's
configuration space, and the tendency of the system to evolve toward more
probable states. The interaction topology constrains the accessible reaction
trajectories and the introduced differential geometrical approach allows
analysis of curvature constraints which help us to understand pathway
saturation and transition dynamics. Rather than treating reaction space as an
unconstrained combinatorial landscape, we frame it as a structured manifold
with higher order curvature describing a geodesic for system evolution under
intrinsic constraints. This geometrical perspective offers a unique insight
into pathway saturation, self-interruption, and emergent behavior in reaction
networks, and provides a scalable framework for modeling large biochemical or
catalytic systems.
| [
{
"version": "v1",
"created": "Sun, 20 Apr 2025 18:28:46 GMT"
}
] | 2025-04-22T00:00:00 | [
[
"Rodriguez",
"Anderson M.",
""
]
] |
2504.14701 | Andres Fernandez Rodriguez | Andres Fernandez, Frank Schneider, Maren Mahsereci, Philipp Hennig | Connecting Parameter Magnitudes and Hessian Eigenspaces at Scale using
Sketched Methods | Accepted at TMLR 2025 | null | null | null | cs.LG stat.ML | http://creativecommons.org/licenses/by/4.0/ | Recently, it has been observed that when training a deep neural net with SGD,
the majority of the loss landscape's curvature quickly concentrates in a tiny
*top* eigenspace of the loss Hessian, which remains largely stable thereafter.
Independently, it has been shown that successful magnitude pruning masks for
deep neural nets emerge early in training and remain stable thereafter. In this
work, we study these two phenomena jointly and show that they are connected: We
develop a methodology to measure the similarity between arbitrary parameter
masks and Hessian eigenspaces via Grassmannian metrics. We identify *overlap*
as the most useful such metric due to its interpretability and stability. To
compute *overlap*, we develop a matrix-free algorithm based on sketched SVDs
that allows us to compute over 1000 Hessian eigenpairs for nets with over 10M
parameters --an unprecedented scale by several orders of magnitude. Our
experiments reveal an *overlap* between magnitude parameter masks and top
Hessian eigenspaces consistently higher than chance-level, and that this effect
gets accentuated for larger network sizes. This result indicates that *top
Hessian eigenvectors tend to be concentrated around larger parameters*, or
equivalently, that *larger parameters tend to align with directions of larger
loss curvature*. Our work provides a methodology to approximate and analyze
deep learning Hessians at scale, as well as a novel insight on the structure of
their eigenspace.
| [
{
"version": "v1",
"created": "Sun, 20 Apr 2025 18:29:39 GMT"
}
] | 2025-04-22T00:00:00 | [
[
"Fernandez",
"Andres",
""
],
[
"Schneider",
"Frank",
""
],
[
"Mahsereci",
"Maren",
""
],
[
"Hennig",
"Philipp",
""
]
] |
2504.14702 | Alex Iosevich | K. Aldahleh, A. Iosevich, J. Iosevich, J. Jaimangal, A. Mayeli, and S.
Pack | Additive energy, uncertainty principle and signal recovery mechanisms | null | null | null | null | math.CA cs.IT math.IT | http://arxiv.org/licenses/nonexclusive-distrib/1.0/ | Given a signal $f:G\to\mathbb{C}$, where $G$ is a finite abelian group, under
what reasonable assumptions can we guarantee the exact recovery of $f$ from a
proper subset of its Fourier coefficients? In 1989, Donoho and Stark
established a result \cite{DS89} using the classical uncertainty principle,
which states that $|\text{supp}(f)|\cdot|\text{supp}(\hat{f})|\geq |G|$ for any
nonzero signal $f$. Another result, first proven by Santose and Symes
\cite{SS86}, was based on the Logan phenomenon \cite{L65}. In particular, the
result showcases how the $L^1$ and $L^2$ minimizing signals with matching
Fourier frequencies often recovers the original signal.
The purpose of this paper is to relate these recovery mechanisms to additive
energy, a combinatorial measure denoted and defined by $$\Lambda(A)=\left|
\left\{ (x_1, x_2, x_3, x_4) \in A^4 \mid x_1 + x_2 = x_3 + x_4 \right\}
\right|,$$ where $A\subset\mathbb{Z}_N^d$. In the first part of this paper, we
use combinatorial techniques to establish an improved variety of the
uncertainty principle in terms of additive energy. In a similar fashion as the
Donoho-Stark argument, we use this principle to establish an often stronger
recovery condition. In the latter half of the paper, we invoke these
combinatorial methods to demonstrate two $L^p$ minimizing recovery results.
| [
{
"version": "v1",
"created": "Sun, 20 Apr 2025 18:31:25 GMT"
}
] | 2025-04-22T00:00:00 | [
[
"Aldahleh",
"K.",
""
],
[
"Iosevich",
"A.",
""
],
[
"Iosevich",
"J.",
""
],
[
"Jaimangal",
"J.",
""
],
[
"Mayeli",
"A.",
""
],
[
"Pack",
"S.",
""
]
] |
2504.14703 | Natalia Maslova | Mingzhu Chen, Natalia V. Maslova, Marianna R. Zinov'eva | On characterization of groups by isomorphism type of Gruenberg-Kegel
graph | 15 pages, 3 figures | null | null | null | math.GR | http://creativecommons.org/licenses/by/4.0/ | The Gruenberg-Kegel graph (or the prime graph) $\Gamma(G)$ of a finite group
$G$ is the graph whose vertex set is the set of prime divisors of $|G|$ and in
which two distinct vertices $r$ and $s$ are adjacent if and only if there
exists an element of order $rs$ in $G$. A group $G$ is recognizable by
isomorphism type of Gruenberg--Kegel graph if for every group $H$ the
isomorphism between $\Gamma(H)$ and $\Gamma(G)$ as abstract graphs (i.\,e.
unlabeled graphs) implies that $G\cong H$. In this paper, we prove that finite
simple exceptional groups of Lie type ${^2}E_6(2)$ and $E_8(q)$ for $q \in \{3,
4, 5, 7, 8, 9, 17\}$ are recognizable by isomorphism type of Gruenberg-Kegel
graph.
| [
{
"version": "v1",
"created": "Sun, 20 Apr 2025 18:32:51 GMT"
}
] | 2025-04-22T00:00:00 | [
[
"Chen",
"Mingzhu",
""
],
[
"Maslova",
"Natalia V.",
""
],
[
"Zinov'eva",
"Marianna R.",
""
]
] |
2504.14704 | Hong Yang | Hong Yang, Qi Yu, Travis Desel | Can We Ignore Labels In Out of Distribution Detection? | null | null | null | null | cs.LG cs.AI cs.SY eess.SY | http://creativecommons.org/licenses/by/4.0/ | Out-of-distribution (OOD) detection methods have recently become more
prominent, serving as a core element in safety-critical autonomous systems. One
major purpose of OOD detection is to reject invalid inputs that could lead to
unpredictable errors and compromise safety. Due to the cost of labeled data,
recent works have investigated the feasibility of self-supervised learning
(SSL) OOD detection, unlabeled OOD detection, and zero shot OOD detection. In
this work, we identify a set of conditions for a theoretical guarantee of
failure in unlabeled OOD detection algorithms from an information-theoretic
perspective. These conditions are present in all OOD tasks dealing with
real-world data: I) we provide theoretical proof of unlabeled OOD detection
failure when there exists zero mutual information between the learning
objective and the in-distribution labels, a.k.a. 'label blindness', II) we
define a new OOD task - Adjacent OOD detection - that tests for label blindness
and accounts for a previously ignored safety gap in all OOD detection
benchmarks, and III) we perform experiments demonstrating that existing
unlabeled OOD methods fail under conditions suggested by our label blindness
theory and analyze the implications for future research in unlabeled OOD
methods.
| [
{
"version": "v1",
"created": "Sun, 20 Apr 2025 18:37:51 GMT"
}
] | 2025-04-22T00:00:00 | [
[
"Yang",
"Hong",
""
],
[
"Yu",
"Qi",
""
],
[
"Desel",
"Travis",
""
]
] |
2504.14705 | Rewa Bush | Rewa Clark Bush, Samantha C. Wu, Rosa Wallace Everson, Ricardo Yarza,
Ariadna Murguia-Berthier, Enrico Ramirez-Ruiz | Black Hole Survival Guide: Searching for Stars in the Galactic Center
That Endure Partial Tidal Disruption | 14 pages, 6 figures, submitted to ApJL | null | null | null | astro-ph.SR astro-ph.GA astro-ph.HE | http://creativecommons.org/licenses/by/4.0/ | Once per 10,000-100,000 years, an unlucky star may experience a close
encounter with a supermassive black hole (SMBH), partially or fully tearing
apart the star in an exceedingly brief, bright interaction called a tidal
disruption event (TDE). Remnants of partial TDEs are expected to be plentiful
in our Galactic Center, where at least six unexplained, diffuse, star-like "G
objects" have already been detected which may have formed via interactions
between stars and the SMBH. Using numerical simulations, this work aims to
identify the characteristics of TDE remnants. We take 3D hydrodynamic FLASH
models of partially disrupted stars and map them into the 1D stellar evolution
code MESA to examine the properties of these remnants from tens to billions of
years after the TDE. The remnants initially exhibit a brief, highly luminous
phase, followed by an extended cooling period as they return to stable hydrogen
burning. During the initial stage (< 10,000 yr) their luminosities increase by
orders of magnitude, making them intriguing candidates to explain a fraction of
the mysterious G objects. Notably, mild TDEs are the most common and result in
the brightest remnants during this initial phase. However, most remnants exist
in a long-lived stage where they are only modestly offset in temperature and
luminosity compared to main-sequence stars of equivalent mass. Nonetheless, our
results indicate remnants will sustain abnormal, metal-enriched envelopes that
may be discernible through spectroscopic analysis. Identifying TDE survivors
within the Milky Way could further illuminate some of the most gravitationally
intense encounters in the Universe.
| [
{
"version": "v1",
"created": "Sun, 20 Apr 2025 18:43:19 GMT"
}
] | 2025-04-22T00:00:00 | [
[
"Bush",
"Rewa Clark",
""
],
[
"Wu",
"Samantha C.",
""
],
[
"Everson",
"Rosa Wallace",
""
],
[
"Yarza",
"Ricardo",
""
],
[
"Murguia-Berthier",
"Ariadna",
""
],
[
"Ramirez-Ruiz",
"Enrico",
""
]
] |
2504.14707 | Ratna Kandala | Ratna Kandala, Katie Hoemann | Evaluating BERTopic on Open-Ended Data: A Case Study with Belgian Dutch
Daily Narratives | null | null | null | null | cs.CL | http://creativecommons.org/licenses/by/4.0/ | This study explores BERTopic's potential for modeling open-ended Belgian
Dutch daily narratives, contrasting its performance with Latent Dirichlet
Allocation (LDA) and KMeans. Although LDA scores well on certain automated
metrics, human evaluations reveal semantically irrelevant co-occurrences,
highlighting the limitations of purely statistic-based methods. In contrast,
BERTopic's reliance on contextual embeddings yields culturally resonant themes,
underscoring the importance of hybrid evaluation frameworks that account for
morphologically rich languages. KMeans performed less coherently than prior
research suggested, pointing to the unique challenges posed by personal
narratives. Our findings emphasize the need for robust generalization in NLP
models, especially in underrepresented linguistic contexts.
| [
{
"version": "v1",
"created": "Sun, 20 Apr 2025 18:51:08 GMT"
}
] | 2025-04-22T00:00:00 | [
[
"Kandala",
"Ratna",
""
],
[
"Hoemann",
"Katie",
""
]
] |
2504.14708 | Parshuram Aarotale | Parshuram N. Aarotale and Ajita Rattani | Time Frequency Analysis of EMG Signal for Gesture Recognition using Fine
grained Features | null | null | null | null | cs.CV cs.AI cs.LG | http://creativecommons.org/licenses/by-nc-nd/4.0/ | Electromyography (EMG) based hand gesture recognition converts forearm muscle
activity into control commands for prosthetics, rehabilitation, and human
computer interaction. This paper proposes a novel approach to EMG-based hand
gesture recognition that uses fine-grained classification and presents XMANet,
which unifies low-level local and high level semantic cues through cross layer
mutual attention among shallow to deep CNN experts. Using stacked spectrograms
and scalograms derived from the Short Time Fourier Transform (STFT) and Wavelet
Transform (WT), we benchmark XMANet against ResNet50, DenseNet-121,
MobileNetV3, and EfficientNetB0. Experimental results on the Grabmyo dataset
indicate that, using STFT, the proposed XMANet model outperforms the baseline
ResNet50, EfficientNetB0, MobileNetV3, and DenseNet121 models with improvement
of approximately 1.72%, 4.38%, 5.10%, and 2.53%, respectively. When employing
the WT approach, improvements of around 1.57%, 1.88%, 1.46%, and 2.05% are
observed over the same baselines. Similarly, on the FORS EMG dataset, the
XMANet(ResNet50) model using STFT shows an improvement of about 5.04% over the
baseline ResNet50. In comparison, the XMANet(DenseNet121) and
XMANet(MobileNetV3) models yield enhancements of approximately 4.11% and 2.81%,
respectively. Moreover, when using WT, the proposed XMANet achieves gains of
around 4.26%, 9.36%, 5.72%, and 6.09% over the baseline ResNet50, DenseNet121,
MobileNetV3, and EfficientNetB0 models, respectively. These results confirm
that XMANet consistently improves performance across various architectures and
signal processing techniques, demonstrating the strong potential of fine
grained features for accurate and robust EMG classification.
| [
{
"version": "v1",
"created": "Sun, 20 Apr 2025 18:51:10 GMT"
}
] | 2025-04-22T00:00:00 | [
[
"Aarotale",
"Parshuram N.",
""
],
[
"Rattani",
"Ajita",
""
]
] |
2504.14709 | Hui Zhou | Hui Zhou and Shaoshuai Shi and Hongsheng Li | Exposing the Copycat Problem of Imitation-based Planner: A Novel
Closed-Loop Simulator, Causal Benchmark and Joint IL-RL Baseline | null | null | null | null | cs.CV cs.AI | http://arxiv.org/licenses/nonexclusive-distrib/1.0/ | Machine learning (ML)-based planners have recently gained significant
attention. They offer advantages over traditional optimization-based planning
algorithms. These advantages include fewer manually selected parameters and
faster development. Within ML-based planning, imitation learning (IL) is a
common algorithm. It primarily learns driving policies directly from supervised
trajectory data. While IL has demonstrated strong performance on many open-loop
benchmarks, it remains challenging to determine if the learned policy truly
understands fundamental driving principles, rather than simply extrapolating
from the ego-vehicle's initial state. Several studies have identified this
limitation and proposed algorithms to address it. However, these methods often
use original datasets for evaluation. In these datasets, future trajectories
are heavily dependent on initial conditions. Furthermore, IL often overfits to
the most common scenarios. It struggles to generalize to rare or unseen
situations.
To address these challenges, this work proposes: 1) a novel closed-loop
simulator supporting both imitation and reinforcement learning, 2) a causal
benchmark derived from the Waymo Open Dataset to rigorously assess the impact
of the copycat problem, and 3) a novel framework integrating imitation learning
and reinforcement learning to overcome the limitations of purely imitative
approaches. The code for this work will be released soon.
| [
{
"version": "v1",
"created": "Sun, 20 Apr 2025 18:51:26 GMT"
}
] | 2025-04-22T00:00:00 | [
[
"Zhou",
"Hui",
""
],
[
"Shi",
"Shaoshuai",
""
],
[
"Li",
"Hongsheng",
""
]
] |
2504.14710 | Fidel F. Villase\~nor | Miguel S\'anchez and Fidel F. Villase\~nor | The ladder of Finsler-type objects and their variational problems on
spacetimes | This preprint has not undergone peer review or any post-submission
improvements or corrections. The Version of Record of this contribution will
be published by within Springer Proceedings in Mathematics & Statistics | null | null | null | math.DG gr-qc math-ph math.MP | http://creativecommons.org/licenses/by/4.0/ | The space of anisotropic $r$-contravariant $s$-covariant $\alpha$-homogeneous
tensors on a manifold admits a functorial structure where vertical derivatives
$\dot{\partial}$ and contractions $\imath_{\mathbb{C}}$ by the Liouville vector
field $\mathbb{C}$ are operators which maintain $s+\alpha$ constant. In
(semi-)Finsler geometry, this structure is transmitted faithfully to
connection-type elements yielding the following ladder: geodesic sprays /
nonlinear connections / anisotropic connections / linear (Finslerian)
connections. However, it is more loosely transmitted to metric-type ones:
Finslerian Lagrangians / Legendre transformations / anisotropic metrics. We
will study this structure in depth and apply it to discuss the recent
variational proposals (Einstein-Hilbert, Einstein-Palatini, Einstein-Cartan)
for generalizing Einstein equations to the Finsler setting.
| [
{
"version": "v1",
"created": "Sun, 20 Apr 2025 18:54:49 GMT"
}
] | 2025-04-22T00:00:00 | [
[
"Sánchez",
"Miguel",
""
],
[
"Villaseñor",
"Fidel F.",
""
]
] |
2504.14711 | Zimu Xiang | H. A. Kierstead, Alexandr Kostochka, Zimu Xiang | Results and Problems on Equitable Coloring of Graphs | null | null | null | null | math.CO | http://creativecommons.org/licenses/by-nc-sa/4.0/ | A proper coloring of vertices of a graph is equitable if the sizes of any two
color classes differ by at most 1. Such colorings have many applications and
are interesting by themselves. In this paper, we discuss the state of art and
unsolved problems on equitable coloring and its list versions.
| [
{
"version": "v1",
"created": "Sun, 20 Apr 2025 18:56:19 GMT"
}
] | 2025-04-22T00:00:00 | [
[
"Kierstead",
"H. A.",
""
],
[
"Kostochka",
"Alexandr",
""
],
[
"Xiang",
"Zimu",
""
]
] |
2504.14712 | Zhengyang Kris Weng | Zhengyang Kris Weng | BiDexHand: Design and Evaluation of an Open-Source 16-DoF Biomimetic
Dexterous Hand | This work has been submitted to ICRA 2025 Dexterity Workshop for
possible publication | null | null | null | cs.RO | http://creativecommons.org/licenses/by-nc-sa/4.0/ | Achieving human-level dexterity in robotic hands remains a fundamental
challenge for enabling versatile manipulation across diverse applications. This
extended abstract presents BiDexHand, a cable-driven biomimetic robotic hand
that combines human-like dexterity with accessible and efficient mechanical
design. The robotic hand features 16 independently actuated degrees of freedom
and 5 mechanically coupled joints through novel phalange designs that replicate
natural finger motion. Performance validation demonstrated success across all
33 grasp types in the GRASP Taxonomy, 9 of 11 positions in the Kapandji thumb
opposition test, a measured fingertip force of 2.14\,N, and the capability to
lift a 10\,lb weight. As an open-source platform supporting multiple control
modes including vision-based teleoperation, BiDexHand aims to democratize
access to advanced manipulation capabilities for the broader robotics research
community.
| [
{
"version": "v1",
"created": "Sun, 20 Apr 2025 18:56:20 GMT"
}
] | 2025-04-22T00:00:00 | [
[
"Weng",
"Zhengyang Kris",
""
]
] |
2504.14713 | Mark Shattuck | Toufik Mansour and Mark Shattuck | Avoidance of vincular patterns by flattened derangements | null | null | null | null | math.CO | http://creativecommons.org/licenses/by/4.0/ | In this paper, we consider the problem of avoiding a single vincular pattern
of length three by derangements in the flattened sense and find explicit
formulas for the generating functions enumerating members of each corresponding
avoidance class according to the number of cycles. We make frequent use of the
kernel method in solving the functional equations that arise which are
satisfied by these (ordinary) generating functions. In the case of avoiding
23-1, which is equivalent to 32-1 in the flattened sense, it is more convenient
to consider the exponential generating function instead due to the form of the
recurrence. This leads to an explicit expression for the distribution of the
number of cycles in terms of Stirling numbers of the second kind and the
determinant of a certain tridiagonal matrix. Finally, the cases of 3-12 and
3-21 are perhaps the most difficult of all, and here we make use of a pair of
auxiliary statistics in order to find a system of recurrences that enumerate
each avoidance class.
| [
{
"version": "v1",
"created": "Sun, 20 Apr 2025 19:01:51 GMT"
}
] | 2025-04-22T00:00:00 | [
[
"Mansour",
"Toufik",
""
],
[
"Shattuck",
"Mark",
""
]
] |
2504.14714 | Leonid Prigozhin | Vladimir Sokolovsky and Leonid Prigozhin | Electromagnetic analysis of coated conductors with ferromagnetic
substrates: Novel insights | 16 pages, 8 figures | null | null | null | cond-mat.supr-con | http://creativecommons.org/licenses/by/4.0/ | Ferromagnetic substrates can significantly influence the electromagnetic
response of a coated conductor to an external magnetic field and transport
current. This study analyzes this response theoretically using the thin shell
integrodifferential model. First, assuming the substrate is strongly magnetic
and the superconductor is in the Meissner state, we present the analytical
solution in a convenient explicit form. This helps us to analyze the
superconducting current density distributions, highlighting their differences
from those in conductors with non-magnetic substrates. Secondly, for a
superconducting layer characterized by a nonlinear current-voltage relation and
a substrate with a finite field-independent magnetic permeability, we use an
effective spectral numerical method to study the unique features of this hybrid
superconductor/ferromagnet system, such as magnetization in a parallel external
field and the peculiar nonmonotonic variation of loss observed when alternating
transport current and parallel field are applied simultaneously. Dynamic losses
for the case of a direct transport current and an alternating parallel field
are also investigated. It is shown that tuning the phase and amplitude of the
applied parallel field relative to those of the transport current can minimize
AC losses.
| [
{
"version": "v1",
"created": "Sun, 20 Apr 2025 19:03:42 GMT"
}
] | 2025-04-22T00:00:00 | [
[
"Sokolovsky",
"Vladimir",
""
],
[
"Prigozhin",
"Leonid",
""
]
] |
2504.14715 | Lameya Sabrin | Md. Sanaullah Chowdhury, Salauddin Tapu, Noyon Kumar Sarkar, Ferdous
Bin Ali, Lameya Sabrin | Med-2D SegNet: A Light Weight Deep Neural Network for Medical 2D Image
Segmentation | null | null | null | null | cs.CV | http://arxiv.org/licenses/nonexclusive-distrib/1.0/ | Accurate and efficient medical image segmentation is crucial for advancing
clinical diagnostics and surgical planning, yet remains a complex challenge due
to the variability in anatomical structures and the demand for low-complexity
models. In this paper, we introduced Med-2D SegNet, a novel and highly
efficient segmentation architecture that delivers outstanding accuracy while
maintaining a minimal computational footprint. Med-2D SegNet achieves
state-of-the-art performance across multiple benchmark datasets, including
KVASIR-SEG, PH2, EndoVis, and GLAS, with an average Dice similarity coefficient
(DSC) of 89.77% across 20 diverse datasets. Central to its success is the
compact Med Block, a specialized encoder design that incorporates dimension
expansion and parameter reduction, enabling precise feature extraction while
keeping model parameters to a low count of just 2.07 million. Med-2D SegNet
excels in cross-dataset generalization, particularly in polyp segmentation,
where it was trained on KVASIR-SEG and showed strong performance on unseen
datasets, demonstrating its robustness in zero-shot learning scenarios, even
though we acknowledge that further improvements are possible. With top-tier
performance in both binary and multi-class segmentation, Med-2D SegNet
redefines the balance between accuracy and efficiency, setting a new benchmark
for medical image analysis. This work paves the way for developing accessible,
high-performance diagnostic tools suitable for clinical environments and
resource-constrained settings, making it a step forward in the democratization
of advanced medical technology.
| [
{
"version": "v1",
"created": "Sun, 20 Apr 2025 19:04:43 GMT"
}
] | 2025-04-22T00:00:00 | [
[
"Chowdhury",
"Md. Sanaullah",
""
],
[
"Tapu",
"Salauddin",
""
],
[
"Sarkar",
"Noyon Kumar",
""
],
[
"Ali",
"Ferdous Bin",
""
],
[
"Sabrin",
"Lameya",
""
]
] |
2504.14716 | Tuhina Tripathi | Tuhina Tripathi, Manya Wadhwa, Greg Durrett, Scott Niekum | Pairwise or Pointwise? Evaluating Feedback Protocols for Bias in
LLM-Based Evaluation | null | null | null | null | cs.LG | http://creativecommons.org/licenses/by/4.0/ | Large Language Models (LLMs) are widely used as proxies for human labelers in
both training (Reinforcement Learning from AI Feedback) and large-scale
response evaluation (LLM-as-a-judge). Alignment and evaluation are critical
components in the development of reliable LLMs, and the choice of feedback
protocol plays a central role in both but remains understudied. In this work,
we show that the choice of feedback protocol (absolute scores versus relative
preferences) can significantly affect evaluation reliability and induce
systematic biases. In particular, we show that pairwise evaluation protocols
are more vulnerable to distracted evaluation. Generator models can exploit
spurious attributes (or distractor features) favored by the LLM judge,
resulting in inflated scores for lower-quality outputs and misleading training
signals. We find that absolute scoring is more robust to such manipulation,
producing judgments that better reflect response quality and are less
influenced by distractor features. Our results demonstrate that generator
models can flip preferences by embedding distractor features, skewing
LLM-as-a-judge comparisons and leading to inaccurate conclusions about model
quality in benchmark evaluations. Pairwise preferences flip in about 35% of the
cases, compared to only 9% for absolute scores. We offer recommendations for
choosing feedback protocols based on dataset characteristics and evaluation
objectives.
| [
{
"version": "v1",
"created": "Sun, 20 Apr 2025 19:05:59 GMT"
}
] | 2025-04-22T00:00:00 | [
[
"Tripathi",
"Tuhina",
""
],
[
"Wadhwa",
"Manya",
""
],
[
"Durrett",
"Greg",
""
],
[
"Niekum",
"Scott",
""
]
] |
2504.14717 | Lei Ke | Bowei Zhang, Lei Ke, Adam W. Harley and Katerina Fragkiadaki | TAPIP3D: Tracking Any Point in Persistent 3D Geometry | Long-term feed-forward 3D point tracking in persistent 3D point maps.
Code:https://github.com/zbw001/TAPIP3D | null | null | null | cs.CV cs.LG | http://arxiv.org/licenses/nonexclusive-distrib/1.0/ | We introduce TAPIP3D, a novel approach for long-term 3D point tracking in
monocular RGB and RGB-D videos. TAPIP3D represents videos as camera-stabilized
spatio-temporal feature clouds, leveraging depth and camera motion information
to lift 2D video features into a 3D world space where camera motion is
effectively canceled. TAPIP3D iteratively refines multi-frame 3D motion
estimates within this stabilized representation, enabling robust tracking over
extended periods. To manage the inherent irregularities of 3D point
distributions, we propose a Local Pair Attention mechanism. This 3D
contextualization strategy effectively exploits spatial relationships in 3D,
forming informative feature neighborhoods for precise 3D trajectory estimation.
Our 3D-centric approach significantly outperforms existing 3D point tracking
methods and even enhances 2D tracking accuracy compared to conventional 2D
pixel trackers when accurate depth is available. It supports inference in both
camera coordinates (i.e., unstabilized) and world coordinates, and our results
demonstrate that compensating for camera motion improves tracking performance.
Our approach replaces the conventional 2D square correlation neighborhoods used
in prior 2D and 3D trackers, leading to more robust and accurate results across
various 3D point tracking benchmarks. Project Page: https://tapip3d.github.io
| [
{
"version": "v1",
"created": "Sun, 20 Apr 2025 19:09:43 GMT"
}
] | 2025-04-22T00:00:00 | [
[
"Zhang",
"Bowei",
""
],
[
"Ke",
"Lei",
""
],
[
"Harley",
"Adam W.",
""
],
[
"Fragkiadaki",
"Katerina",
""
]
] |
2504.14718 | Hossam Farag | Hossam Farag, Mohamed Ragab, Gilberto Berardinelli, and Cedomir
Stefanovic | Proactive Radio Resource Allocation for 6G In-Factory Subnetworks | null | null | null | null | eess.SY cs.SY | http://creativecommons.org/licenses/by/4.0/ | 6G In-Factory Subnetworks (InF-S) have recently been introduced as
short-range, low-power radio cells installed in robots and production modules
to support the strict requirements of modern control systems. Information
freshness, characterized by the Age of Information (AoI), is crucial to
guarantee the stability and accuracy of the control loop in these systems.
However, achieving strict AoI performance poses significant challenges
considering the limited resources and the high dynamic environment of InF-S. In
this work, we introduce a proactive radio resource allocation approach to
minimize the AoI violation probability. The proposed approach adopts a
decentralized learning framework using Bayesian Ridge Regression (BRR) to
predict the future AoI by actively learning the system dynamics. Based on the
predicted AoI value, radio resources are proactively allocated to minimize the
probability of AoI exceeding a predefined threshold, hence enhancing the
reliability and accuracy of the control loop. The conducted simulation results
prove the effectiveness of our proposed approach to improve the AoI performance
where a reduction of 98% is achieved in the AoI violation probability compared
to relevant baseline methods.
| [
{
"version": "v1",
"created": "Sun, 20 Apr 2025 19:14:27 GMT"
}
] | 2025-04-22T00:00:00 | [
[
"Farag",
"Hossam",
""
],
[
"Ragab",
"Mohamed",
""
],
[
"Berardinelli",
"Gilberto",
""
],
[
"Stefanovic",
"Cedomir",
""
]
] |
2504.14719 | Konstantinos Kartas | Konstantinos Kartas | Perfectoid $C_i$ transfer | 20 pages | null | null | MPIM-Bonn-2025 | math.NT math.AG math.LO | http://creativecommons.org/licenses/by/4.0/ | We prove a perfectoid analogue of the Ax-Kochen theorem on zeros of $p$-adic
forms: Given $d\in \mathbb{N}$, there is a finite totally ramified extension
$E/\mathbb{Q}_p$ such that every untilt of
$\mathbb{F}_p(\!(t^{1/p^{\infty}})\!)$ containing $E$ is $C_2(d)$. We also
prove a similar result for the existence of rational points in rationally
connected varieties over perfectoid field extensions of $\mathbb{Q}_p^{ur}$.
| [
{
"version": "v1",
"created": "Sun, 20 Apr 2025 19:17:54 GMT"
}
] | 2025-04-22T00:00:00 | [
[
"Kartas",
"Konstantinos",
""
]
] |
2504.14720 | Tamir Berger | Tamir Berger, Jonathan Sterenson, Raz Birman, and Ofer Hadar | Video QoE Metrics from Encrypted Traffic: Application-agnostic
Methodology | 8 pages, 7 figures | null | null | null | cs.NI cs.LG cs.MM | http://creativecommons.org/licenses/by/4.0/ | Instant Messaging-Based Video Call Applications (IMVCAs) and Video
Conferencing Applications (VCAs) have become integral to modern communication.
Ensuring a high Quality of Experience (QoE) for users in this context is
critical for network operators, as network conditions significantly impact user
QoE. However, network operators lack access to end-device QoE metrics due to
encrypted traffic. Existing solutions estimate QoE metrics from encrypted
traffic traversing the network, with the most advanced approaches leveraging
machine learning models. Subsequently, the need for ground truth QoE metrics
for training and validation poses a challenge, as not all video applications
provide these metrics. To address this challenge, we propose an
application-agnostic approach for objective QoE estimation from encrypted
traffic. Independent of the video application, we obtained key video QoE
metrics, enabling broad applicability to various proprietary IMVCAs and VCAs.
To validate our solution, we created a diverse dataset from WhatsApp video
sessions under various network conditions, comprising 25,680 seconds of traffic
data and QoE metrics. Our evaluation shows high performance across the entire
dataset, with 85.2% accuracy for FPS predictions within an error margin of two
FPS, and 90.2% accuracy for PIQE-based quality rating classification.
| [
{
"version": "v1",
"created": "Sun, 20 Apr 2025 19:18:13 GMT"
}
] | 2025-04-22T00:00:00 | [
[
"Berger",
"Tamir",
""
],
[
"Sterenson",
"Jonathan",
""
],
[
"Birman",
"Raz",
""
],
[
"Hadar",
"Ofer",
""
]
] |
2504.14721 | Anqi Dong | Shenghan Mei, Ziqin He, Yidan Mei, Xin Mao, Anqi Dong, Ren Wang, and
Can Chen | Data-driven model order reduction for T-Product-Based dynamical systems | 12 pages, 1 figure | null | null | null | eess.SY cs.NA cs.SY math.DS math.NA | http://creativecommons.org/licenses/by/4.0/ | Model order reduction plays a crucial role in simplifying complex systems
while preserving their essential dynamic characteristics, making it an
invaluable tool in a wide range of applications, including robotic systems,
signal processing, and fluid dynamics. However, traditional model order
reduction techniques like balanced truncation are not designed to handle tensor
data directly and instead require unfolding the data, which may lead to the
loss of important higher-order structural information. In this article, we
introduce a novel framework for data-driven model order reduction of
T-product-based dynamical systems (TPDSs), which are often used to capture the
evolution of third-order tensor data such as images and videos through the
T-product. Specifically, we develop advanced T-product-based techniques,
including T-balanced truncation, T-balanced proper orthogonal decomposition,
and the T-eigensystem realization algorithm for input-output TPDSs by
leveraging the unique properties of T-singular value decomposition. We
demonstrate that these techniques offer significant memory and computational
savings while achieving reduction errors that are comparable to those of
conventional methods. The effectiveness of the proposed framework is further
validated through synthetic and real-world examples.
| [
{
"version": "v1",
"created": "Sun, 20 Apr 2025 19:28:20 GMT"
}
] | 2025-04-22T00:00:00 | [
[
"Mei",
"Shenghan",
""
],
[
"He",
"Ziqin",
""
],
[
"Mei",
"Yidan",
""
],
[
"Mao",
"Xin",
""
],
[
"Dong",
"Anqi",
""
],
[
"Wang",
"Ren",
""
],
[
"Chen",
"Can",
""
]
] |
2504.14722 | An Ghysels | Wouter Vervust, Daniel T. Zhang, Enrico Riccardi, Titus S. van Erp, An
Ghysels | Path sampling challenges in large biomolecular systems: RETIS and
REPPTIS for ABL-imatinib kinetics | Preprint | null | null | null | physics.bio-ph cond-mat.soft cond-mat.stat-mech physics.chem-ph physics.comp-ph | http://creativecommons.org/licenses/by-sa/4.0/ | Predicting the kinetics of drug-protein interactions is crucial for
understanding drug efficacy, particularly in personalized medicine, where
protein mutations can significantly alter drug residence times. This study
applies Replica Exchange Transition Interface Sampling (RETIS) and its Partial
Path variant (REPPTIS) to investigate the dissociation kinetics of imatinib
from Abelson nonreceptor tyrosine kinase (ABL) and mutants relevant to chronic
myeloid leukemia therapy. These path-sampling methods offer a bias-free
alternative to conventional approaches requiring qualitative predefined
reaction coordinates. Nevertheless, the complex free-energy landscape of
ABL-imatinib dissociation presents significant challenges. Multiple metastable
states and orthogonal barriers lead to parallel unbinding pathways,
complicating convergence in TIS-based methods. Despite employing computational
efficiency strategies such as asynchronous replica exchange, full convergence
remained elusive. This work provides a critical assessment of path sampling in
high-dimensional biological systems, discussing the need for enhanced
initialization strategies, advanced Monte Carlo path generation moves, and
machine learning-derived reaction coordinates to improve kinetic predictions of
drug dissociation with minimal prior knowledge.
| [
{
"version": "v1",
"created": "Sun, 20 Apr 2025 19:35:24 GMT"
}
] | 2025-04-22T00:00:00 | [
[
"Vervust",
"Wouter",
""
],
[
"Zhang",
"Daniel T.",
""
],
[
"Riccardi",
"Enrico",
""
],
[
"van Erp",
"Titus S.",
""
],
[
"Ghysels",
"An",
""
]
] |
2504.14723 | Andrzej Lingas | Miroslaw Kowaluk and Andrzej Lingas and Mia Persson | Approximate all-pairs Hamming distances and 0-1 matrix multiplication | 18 pages | null | null | null | cs.DS cs.CC | http://arxiv.org/licenses/nonexclusive-distrib/1.0/ | Arslan showed that computing all-pairs Hamming distances is easily
reducible to arithmetic 0-1 matrix multiplication (IPL 2018). We
provide a reverse, linear-time reduction of arithmetic 0-1 matrix
multiplication to computing all-pairs distances in a Hamming space.
On the other hand, we present a fast randomized algorithm for
approximate all-pairs distances in a Hamming space. By combining it
with our reduction, we obtain also a fast randomized algorithm for
approximate 0-1 matrix multiplication. Next, we present an
output-sensitive randomized algorithm for a minimum spanning tree of
a set of points in a generalized Hamming space, the lower is the
cost of the minimum spanning tree the faster is our
algorithm. Finally, we provide $(2+\epsilon)$- approximation
algorithms for the $\ell$-center clustering and minimum-diameter
$\ell$-clustering problems in a Hamming space $\{0,1\}^d$ that are
substantially faster than the known $2$-approximation ones when both
$\ell$ and $d$ are super-logarithmic.
| [
{
"version": "v1",
"created": "Sun, 20 Apr 2025 19:36:04 GMT"
}
] | 2025-04-22T00:00:00 | [
[
"Kowaluk",
"Miroslaw",
""
],
[
"Lingas",
"Andrzej",
""
],
[
"Persson",
"Mia",
""
]
] |
2504.14724 | Harilaos Sandalidis G. | Aikaterini A.Tsantari, Konstantinos K. Delibasis, Harilaos G.
Sandalidis, Nestor D. Chatzidiamantis | Parametrization of microbial survival models under UVC exposure | null | null | null | null | physics.med-ph q-bio.CB | http://creativecommons.org/licenses/by-nc-sa/4.0/ | This study aims to identify and parameterize the optimal survival curves for
33 fundamental microorganisms subject to UVC exposure through experimental
measurements. We compile published data on UVC doses and corresponding survival
fractions for these microorganisms to estimate parameters for four prominent
survival models: Single-target (ST), Multi-target (MT), Linear Quadratic (LQ),
and Two-Stage Decay (TSD). The best-fitting model for each microorganism is
determined by selecting the one with the lowest mean squared error (MSE)
compared to the experimental data. Our analysis indicates that the MT model is
the most frequently appropriate, accurately fitting 21 of the 33
microorganisms. The TSD model is the best fit for only three, while the LQ
model, though occasionally suitable at lower doses, is often excluded due to
unreliable performance at higher doses. The assessed models, particularly the
MT model, demonstrate strong predictive capabilities for UVC surface
sterilization of microorganisms. However, caution is warranted with the LQ
model at higher doses due to its potential limitations.
| [
{
"version": "v1",
"created": "Sun, 20 Apr 2025 19:37:11 GMT"
}
] | 2025-04-22T00:00:00 | [
[
"Tsantari",
"Aikaterini A.",
""
],
[
"Delibasis",
"Konstantinos K.",
""
],
[
"Sandalidis",
"Harilaos G.",
""
],
[
"Chatzidiamantis",
"Nestor D.",
""
]
] |
2504.14725 | Jayanth Bhargav | Jayanth Bhargav, Shreyas Sundaram, Mahsa Ghasemi | Sensor Scheduling in Intrusion Detection Games with Uncertain Payoffs | 19 pages, 1 figure | null | null | null | eess.SY cs.GT cs.SY math.OC | http://creativecommons.org/licenses/by/4.0/ | We study the problem of sensor scheduling for an intrusion detection task. We
model this as a two-player zero-sum game over a graph, where the defender
(Player 1) seeks to identify the optimal strategy for scheduling sensor
orientations to minimize the probability of missed detection at minimal cost,
while the intruder (Player 2) aims to identify the optimal path selection
strategy to maximize missed detection probability at minimal cost. The
defender's strategy space grows exponentially with the number of sensors,
making direct computation of the Nash Equilibrium (NE) strategies
computationally expensive. To tackle this, we propose a distributed variant of
the Weighted Majority algorithm that exploits the structure of the game's
payoff matrix, enabling efficient computation of the NE strategies with
provable convergence guarantees. Next, we consider a more challenging scenario
where the defender lacks knowledge of the true sensor models and, consequently,
the game's payoff matrix. For this setting, we develop online learning
algorithms that leverage bandit feedback from sensors to estimate the NE
strategies. By building on existing results from perturbation theory and online
learning in matrix games, we derive high-probability order-optimal regret
bounds for our algorithms. Finally, through simulations, we demonstrate the
empirical performance of our proposed algorithms in both known and unknown
payoff scenarios.
| [
{
"version": "v1",
"created": "Sun, 20 Apr 2025 19:39:37 GMT"
}
] | 2025-04-22T00:00:00 | [
[
"Bhargav",
"Jayanth",
""
],
[
"Sundaram",
"Shreyas",
""
],
[
"Ghasemi",
"Mahsa",
""
]
] |
2504.14726 | Maxim Dvornikov | Maxim Dvornikov (IZMIRAN) | Quantum field theory treatment of neutrino spin-flavor oscillations in a
magnetic field | 15 pages, 5 pdf figures | null | null | null | hep-ph hep-th | http://creativecommons.org/licenses/by/4.0/ | We study spin-flavor oscillations of neutrinos in a magnetic field within the
quantum field theory approach in which neutrinos are virtual particles.
Neutrinos are taken to be Majorana particles having a nonzero transition
magnetic moment. We derive the dressed propagators of the neutrino mass
eigenstates exactly accounting for the magnetic field contribution. The matrix
element and the transition probability for spin-flavor oscillations are
obtained. The leading term in the transition probability is shown to coincide
with the result of the standard quantum mechanical description of neutrino
oscillations. We also derive the correction to the transition probability which
has the quantum field theory origin.
| [
{
"version": "v1",
"created": "Sun, 20 Apr 2025 19:52:15 GMT"
}
] | 2025-04-22T00:00:00 | [
[
"Dvornikov",
"Maxim",
"",
"IZMIRAN"
]
] |
2504.14727 | Geng Liu | Geng Liu, Fei Zhu, Rong Feng, Zhiqiang Yi, Shiqi Wang, Gaofeng Meng,
Zhaoxiang Zhang | Semi-parametric Memory Consolidation: Towards Brain-like Deep Continual
Learning | null | null | null | null | cs.LG cs.AI cs.CV | http://arxiv.org/licenses/nonexclusive-distrib/1.0/ | Humans and most animals inherently possess a distinctive capacity to
continually acquire novel experiences and accumulate worldly knowledge over
time. This ability, termed continual learning, is also critical for deep neural
networks (DNNs) to adapt to the dynamically evolving world in open
environments. However, DNNs notoriously suffer from catastrophic forgetting of
previously learned knowledge when trained on sequential tasks. In this work,
inspired by the interactive human memory and learning system, we propose a
novel biomimetic continual learning framework that integrates semi-parametric
memory and the wake-sleep consolidation mechanism. For the first time, our
method enables deep neural networks to retain high performance on novel tasks
while maintaining prior knowledge in real-world challenging continual learning
scenarios, e.g., class-incremental learning on ImageNet. This study
demonstrates that emulating biological intelligence provides a promising path
to enable deep neural networks with continual learning capabilities.
| [
{
"version": "v1",
"created": "Sun, 20 Apr 2025 19:53:13 GMT"
}
] | 2025-04-22T00:00:00 | [
[
"Liu",
"Geng",
""
],
[
"Zhu",
"Fei",
""
],
[
"Feng",
"Rong",
""
],
[
"Yi",
"Zhiqiang",
""
],
[
"Wang",
"Shiqi",
""
],
[
"Meng",
"Gaofeng",
""
],
[
"Zhang",
"Zhaoxiang",
""
]
] |
Subsets and Splits
No saved queries yet
Save your SQL queries to embed, download, and access them later. Queries will appear here once saved.