id
string | submitter
string | authors
string | title
string | comments
string | journal-ref
string | doi
string | report-no
string | categories
string | license
string | abstract
string | versions
list | update_date
timestamp[s] | authors_parsed
sequence |
|---|---|---|---|---|---|---|---|---|---|---|---|---|---|
2505.20302 | Patrick Yubeaton | Patrick Yubeaton and Andre Nakkab and Weihua Xiao and Luca Collini and Ramesh Karri and Chinmay Hegde and Siddharth Garg | VeriThoughts: Enabling Automated Verilog Code Generation using Reasoning and Formal Verification | null | null | null | null | cs.PL cs.AI cs.LO | http://arxiv.org/licenses/nonexclusive-distrib/1.0/ | This paper introduces VeriThoughts, a novel dataset designed for reasoning-based Verilog code generation. We establish a new benchmark framework grounded in formal verification methods to evaluate the quality and correctness of generated hardware descriptions. Additionally, we present a suite of specialized small-scale models optimized specifically for Verilog generation. Our work addresses the growing need for automated hardware design tools that can produce verifiably correct implementations from high-level specifications, potentially accelerating the hardware development process while maintaining rigorous correctness guarantees. Our code and data are available at \href{https://github.com/wilyub/VeriThoughts}{this URL}. | [
{
"version": "v1",
"created": "Fri, 16 May 2025 21:33:14 GMT"
}
] | 2025-05-28T00:00:00 | [
[
"Yubeaton",
"Patrick",
""
],
[
"Nakkab",
"Andre",
""
],
[
"Xiao",
"Weihua",
""
],
[
"Collini",
"Luca",
""
],
[
"Karri",
"Ramesh",
""
],
[
"Hegde",
"Chinmay",
""
],
[
"Garg",
"Siddharth",
""
]
] |
2505.20303 | David Hanson | David Hanson | Future of Code with Generative AI: Transparency and Safety in the Era of AI Generated Software | null | null | null | null | cs.SE cs.AI | http://creativecommons.org/licenses/by/4.0/ | As artificial intelligence becomes increasingly integrated into software development processes, the prevalence and sophistication of AI-generated code continue to expand rapidly. This study addresses the critical need for transparency and safety in AI generated code by examining the current landscape, identifying potential risks, and exploring future implications. We analyze market opportunities for detecting AI-generated code, discuss the challenges associated with managing increasing complexity, and propose solutions to enhance transparency and functionality analysis. Furthermore, this study investigates the longterm implications of AI generated code, including its potential role in the development of artificial general intelligence and its impact on human AI interaction. In conclusion, we emphasize the importance of proactive measures for ensuring the responsible development and deployment of AI in software engineering. | [
{
"version": "v1",
"created": "Sun, 18 May 2025 05:01:41 GMT"
}
] | 2025-05-28T00:00:00 | [
[
"Hanson",
"David",
""
]
] |
2505.20304 | Francisco Herrera | Francisco Herrera, Reyes Calder\'on | Opacity as a Feature, Not a Flaw: The LoBOX Governance Ethic for Role-Sensitive Explainability and Institutional Trust in AI | 18 pages, 1 figure | null | null | null | cs.CY | http://arxiv.org/licenses/nonexclusive-distrib/1.0/ | This paper introduces LoBOX (Lack of Belief: Opacity \& eXplainability) governance ethic structured framework for managing artificial intelligence (AI) opacity when full transparency is infeasible. Rather than treating opacity as a design flaw, LoBOX defines it as a condition that can be ethically governed through role-calibrated explanation and institutional accountability. The framework comprises a three-stage pathway: reduce accidental opacity, bound irreducible opacity, and delegate trust through structured oversight. Integrating the RED/BLUE XAI model for stakeholder-sensitive explanation and aligned with emerging legal instruments such as the EU AI Act, LoBOX offers a scalable and context-aware alternative to transparency-centric approaches. Reframe trust not as a function of complete system explainability, but as an outcome of institutional credibility, structured justification, and stakeholder-responsive accountability. A governance loop cycles back to ensure that LoBOX remains responsive to evolving technological contexts and stakeholder expectations, to ensure the complete opacity governance. We move from transparency ideals to ethical governance, emphasizing that trustworthiness in AI must be institutionally grounded and contextually justified. We also discuss how cultural or institutional trust varies in different contexts. This theoretical framework positions opacity not as a flaw but as a feature that must be actively governed to ensure responsible AI systems. | [
{
"version": "v1",
"created": "Sun, 18 May 2025 16:59:45 GMT"
}
] | 2025-05-28T00:00:00 | [
[
"Herrera",
"Francisco",
""
],
[
"Calderón",
"Reyes",
""
]
] |
2505.20305 | Francisco Herrera | Francisco Herrera | Making Sense of the Unsensible: Reflection, Survey, and Challenges for XAI in Large Language Models Toward Human-Centered AI | 30 pages, 1 figure | null | null | null | cs.CY | http://arxiv.org/licenses/nonexclusive-distrib/1.0/ | As large language models (LLMs) are increasingly deployed in sensitive domains such as healthcare, law, and education, the demand for transparent, interpretable, and accountable AI systems becomes more urgent. Explainable AI (XAI) acts as a crucial interface between the opaque reasoning of LLMs and the diverse stakeholders who rely on their outputs in high-risk decisions. This paper presents a comprehensive reflection and survey of XAI for LLMs, framed around three guiding questions: Why is explainability essential? What technical and ethical dimensions does it entail? And how can it fulfill its role in real-world deployment?
We highlight four core dimensions central to explainability in LLMs, faithfulness, truthfulness, plausibility, and contrastivity, which together expose key design tensions and guide the development of explanation strategies that are both technically sound and contextually appropriate. The paper discusses how XAI can support epistemic clarity, regulatory compliance, and audience-specific intelligibility across stakeholder roles and decision settings.
We further examine how explainability is evaluated, alongside emerging developments in audience-sensitive XAI, mechanistic interpretability, causal reasoning, and adaptive explanation systems. Emphasizing the shift from surface-level transparency to governance-ready design, we identify critical challenges and future research directions for ensuring the responsible use of LLMs in complex societal contexts. We argue that explainability must evolve into a civic infrastructure fostering trust, enabling contestability, and aligning AI systems with institutional accountability and human-centered decision-making. | [
{
"version": "v1",
"created": "Sun, 18 May 2025 17:30:10 GMT"
}
] | 2025-05-28T00:00:00 | [
[
"Herrera",
"Francisco",
""
]
] |
2505.20306 | Xueqiang Ouyang | Xueqiang Ouyang and Jia Wei | Multi-Modal Artificial Intelligence of Embryo Grading and Pregnancy Prediction in Assisted Reproductive Technology: A Review | null | null | null | null | cs.AI eess.IV q-bio.QM | http://arxiv.org/licenses/nonexclusive-distrib/1.0/ | As a global disease, infertility has always affected human beings. The development of assisted reproductive technology can effectively solve this disease. However, the traditional in vitro fertilization-embryo transfer technology still faces many challenges in improving the success rate of pregnancy, such as the subjectivity of embryo grading and the inefficiency of integrating multi-modal data. Therefore, the introduction of artificial intelligence-based technologies is particularly crucial. This article reviews the application progress of multi-modal artificial intelligence in embryo grading and pregnancy prediction based on different data modalities (including static images, time-lapse videos and structured table data) from a new perspective, and discusses the main challenges in current research, such as the complexity of multi-modal information fusion and data scarcity. | [
{
"version": "v1",
"created": "Mon, 19 May 2025 07:07:13 GMT"
}
] | 2025-05-28T00:00:00 | [
[
"Ouyang",
"Xueqiang",
""
],
[
"Wei",
"Jia",
""
]
] |
2505.20308 | Muhammad Tayyab Khan | Muhammad Tayyab Khan, Lequn Chen, Wenhe Feng and Seung Ki Moon | Large Language Model-Powered Decision Support for a Metal Additive Manufacturing Knowledge Graph | null | null | null | null | cs.IR cs.AI | http://creativecommons.org/licenses/by/4.0/ | Metal additive manufacturing (AM) involves complex interdependencies among processes, materials, feedstock, and post-processing steps. However, the underlying relationships and domain knowledge remain fragmented across literature and static databases that often demand expert-level queries, limiting their applicability in design and planning. To address these gaps, we develop a novel and queryable knowledge graph (KG) in Neo4j, encoding 53 distinct metals and alloys across seven material families, nine AM processes, four feedstock types, and associated post-processing requirements. A large language model (LLM) interface, guided by a few-shot prompting strategy, enables natural language querying without the need for formal query syntax. The system supports a range of tasks, including compatibility checks, multi-constraint filtering, and design for AM (DfAM) guidance. User natural language queries are normalized, translated into Cypher, and executed over the KG, with results reformatted into structured responses. This work presents the first real-time, interactive system that integrates a domain-specific metal AM KG with an LLM interface, offering accessible, explainable decision support for engineers and advancing human-centric tools in manufacturing intelligence. | [
{
"version": "v1",
"created": "Tue, 20 May 2025 18:27:22 GMT"
}
] | 2025-05-28T00:00:00 | [
[
"Khan",
"Muhammad Tayyab",
""
],
[
"Chen",
"Lequn",
""
],
[
"Feng",
"Wenhe",
""
],
[
"Moon",
"Seung Ki",
""
]
] |
2505.20309 | Amr Hegazy | Amr Hegazy, Mostafa Elhoushi, Amr Alanwar | Guiding Giants: Lightweight Controllers for Weighted Activation Steering in LLMs | null | null | null | null | cs.CL cs.LG | http://creativecommons.org/licenses/by/4.0/ | Controlling undesirable Large Language Model (LLM) behaviors, such as the generation of unsafe content or failing to adhere to safety guidelines, often relies on costly fine-tuning. Activation steering provides an alternative for inference-time control, but existing methods typically lack fine-grained, adaptive mechanisms. We introduce a novel approach using a lightweight, trainable controller network integrated during inference. This controller network observes specific intermediate LLM activations and predicts both a global scaling factor and layer-specific weights. The predicted global scaling factor and layer-specific weights then dynamically modulate the intensity of a steering patch, derived from a pre-computed "refusal direction" vector, applied across the LLM's layers during generation. Trained on activations from both harmful and benign prompts, our controller learns to discriminatively apply nuanced, layer-aware interventions, activating steering primarily for harmful inputs. Experiments using safety benchmarks like ToxicChat & In-The-Wild Jailbreak Prompts demonstrate that our weighted steering controller significantly increases refusal rates compared to the base LLM, achieving targeted behavioral modification without altering the original model parameters. Our experiments with Llama-3.1-8B, Llama-3.2-1B & Mistral-7B show our approach outperforms existing methods, presenting an efficient and adaptive method for fine-grained control over LLM behavior at inference time. | [
{
"version": "v1",
"created": "Thu, 22 May 2025 01:48:38 GMT"
}
] | 2025-05-28T00:00:00 | [
[
"Hegazy",
"Amr",
""
],
[
"Elhoushi",
"Mostafa",
""
],
[
"Alanwar",
"Amr",
""
]
] |
2505.20310 | Wanghan Xu | Wanghan Xu, Wenlong Zhang, Fenghua Ling, Ben Fei, Yusong Hu, Fangxuan Ren, Jintai Lin, Wanli Ouyang, Lei Bai | Manalyzer: End-to-end Automated Meta-analysis with Multi-agent System | null | null | null | null | cs.AI cs.MA | http://arxiv.org/licenses/nonexclusive-distrib/1.0/ | Meta-analysis is a systematic research methodology that synthesizes data from multiple existing studies to derive comprehensive conclusions. This approach not only mitigates limitations inherent in individual studies but also facilitates novel discoveries through integrated data analysis. Traditional meta-analysis involves a complex multi-stage pipeline including literature retrieval, paper screening, and data extraction, which demands substantial human effort and time. However, while LLM-based methods can accelerate certain stages, they still face significant challenges, such as hallucinations in paper screening and data extraction. In this paper, we propose a multi-agent system, Manalyzer, which achieves end-to-end automated meta-analysis through tool calls. The hybrid review, hierarchical extraction, self-proving, and feedback checking strategies implemented in Manalyzer significantly alleviate these two hallucinations. To comprehensively evaluate the performance of meta-analysis, we construct a new benchmark comprising 729 papers across 3 domains, encompassing text, image, and table modalities, with over 10,000 data points. Extensive experiments demonstrate that Manalyzer achieves significant performance improvements over the LLM baseline in multi meta-analysis tasks. Project page: https://black-yt.github.io/meta-analysis-page/ . | [
{
"version": "v1",
"created": "Thu, 22 May 2025 07:25:31 GMT"
}
] | 2025-05-28T00:00:00 | [
[
"Xu",
"Wanghan",
""
],
[
"Zhang",
"Wenlong",
""
],
[
"Ling",
"Fenghua",
""
],
[
"Fei",
"Ben",
""
],
[
"Hu",
"Yusong",
""
],
[
"Ren",
"Fangxuan",
""
],
[
"Lin",
"Jintai",
""
],
[
"Ouyang",
"Wanli",
""
],
[
"Bai",
"Lei",
""
]
] |
2505.20311 | Anton Hummel | Anton Hummel, H{\aa}kan Burden, Susanne Stenberg, Jan-Philipp Stegh\"ofer, Niklas K\"uhl | The EU AI Act, Stakeholder Needs, and Explainable AI: Aligning Regulatory Compliance in a Clinical Decision Support System | 17 pages, 2 figures | null | null | null | cs.CY cs.HC | http://creativecommons.org/licenses/by/4.0/ | Explainable AI (XAI) is a promising solution to ensure compliance with the EU AI Act, the first multi-national regulation for AI. XAI aims to enhance transparency and human oversight of AI systems, particularly ``black-box models'', which are criticized as incomprehensible. However, the discourse around the main stakeholders in the AI Act and XAI appears disconnected. While XAI prioritizes the end user's needs as the primary goal, the AI Act focuses on the obligations of the provider and deployer of the AI system. We aim to bridge this divide and provide guidance on how these two worlds are related. By fostering an interdisciplinary discussion in a cross-functional team with XAI, AI Act, legal, and requirements engineering experts, we walk through the steps necessary to analyze an AI-based clinical decision support system to clarify the end-user needs and assess AI Act applicability. By analyzing our justified understanding using an AI system under development as a case, we show that XAI techniques can fill a gap between stakeholder needs and the requirements of the AI Act. We look at the similarities and contrasts between the legal requirements and the needs of stakeholders. In doing so, we encourage researchers and practitioners from the XAI community to reflect on their role towards the AI Act by achieving a mutual understanding of the implications of XAI and the AI Act within different disciplines. | [
{
"version": "v1",
"created": "Thu, 22 May 2025 07:39:04 GMT"
}
] | 2025-05-28T00:00:00 | [
[
"Hummel",
"Anton",
""
],
[
"Burden",
"Håkan",
""
],
[
"Stenberg",
"Susanne",
""
],
[
"Steghöfer",
"Jan-Philipp",
""
],
[
"Kühl",
"Niklas",
""
]
] |
2505.20312 | Aditya Bhattacharya | Aditya Bhattacharya and Katrien Verbert | Let's Get You Hired: A Job Seeker's Perspective on Multi-Agent Recruitment Systems for Explaining Hiring Decisions | Pre-print version only. Please check the published version for any reference or citation | null | null | null | cs.CY cs.AI cs.MA | http://creativecommons.org/licenses/by/4.0/ | During job recruitment, traditional applicant selection methods often lack transparency. Candidates are rarely given sufficient justifications for recruiting decisions, whether they are made manually by human recruiters or through the use of black-box Applicant Tracking Systems (ATS). To address this problem, our work introduces a multi-agent AI system that uses Large Language Models (LLMs) to guide job seekers during the recruitment process. Using an iterative user-centric design approach, we first conducted a two-phased exploratory study with four active job seekers to inform the design and development of the system. Subsequently, we conducted an in-depth, qualitative user study with 20 active job seekers through individual one-to-one interviews to evaluate the developed prototype. The results of our evaluation demonstrate that participants perceived our multi-agent recruitment system as significantly more actionable, trustworthy, and fair compared to traditional methods. Our study further helped us uncover in-depth insights into factors contributing to these perceived user experiences. Drawing from these insights, we offer broader design implications for building user-aligned, multi-agent explainable AI systems across diverse domains. | [
{
"version": "v1",
"created": "Thu, 22 May 2025 10:33:42 GMT"
}
] | 2025-05-28T00:00:00 | [
[
"Bhattacharya",
"Aditya",
""
],
[
"Verbert",
"Katrien",
""
]
] |
2505.20313 | Artur Garcez | Son Tran and Edjard Mota and Artur d'Avila Garcez | Reasoning in Neurosymbolic AI | 50 pages, 13 figures, 56 references. Keywords: Neurosymbolic AI, Restricted Boltzmann Machines, Logical Reasoning, SAT solving, MaxSAT, Energy-based Learning, Constrained Optimization, Modular Deep Learning | null | null | null | cs.AI cs.LO | http://creativecommons.org/licenses/by/4.0/ | Knowledge representation and reasoning in neural networks have been a long-standing endeavor which has attracted much attention recently. The principled integration of reasoning and learning in neural networks is a main objective of the area of neurosymbolic Artificial Intelligence (AI). In this chapter, a simple energy-based neurosymbolic AI system is described that can represent and reason formally about any propositional logic formula. This creates a powerful combination of learning from data and knowledge and logical reasoning. We start by positioning neurosymbolic AI in the context of the current AI landscape that is unsurprisingly dominated by Large Language Models (LLMs). We identify important challenges of data efficiency, fairness and safety of LLMs that might be addressed by neurosymbolic reasoning systems with formal reasoning capabilities. We then discuss the representation of logic by the specific energy-based system, including illustrative examples and empirical evaluation of the correspondence between logical reasoning and energy minimization using Restricted Boltzmann Machines (RBM). Learning from data and knowledge is also evaluated empirically and compared with a symbolic, neural and a neurosymbolic system. Results reported in this chapter in an accessible way are expected to reignite the research on the use of neural networks as massively-parallel models for logical reasoning and promote the principled integration of reasoning and learning in deep networks. We conclude the chapter with a discussion of the importance of positioning neurosymbolic AI within a broader framework of formal reasoning and accountability in AI, discussing the challenges for neurosynbolic AI to tackle the various known problems of reliability of deep learning. | [
{
"version": "v1",
"created": "Thu, 22 May 2025 11:57:04 GMT"
}
] | 2025-05-28T00:00:00 | [
[
"Tran",
"Son",
""
],
[
"Mota",
"Edjard",
""
],
[
"Garcez",
"Artur d'Avila",
""
]
] |
2505.20315 | Zhewei Yao | Zhewei Yao, Guoheng Sun, Lukasz Borchmann, Zheyu Shen, Minghang Deng, Bohan Zhai, Hao Zhang, Ang Li, Yuxiong He | Arctic-Text2SQL-R1: Simple Rewards, Strong Reasoning in Text-to-SQL | 22 pages, 2 figures | null | null | null | cs.CL cs.AI | http://creativecommons.org/licenses/by/4.0/ | Translating natural language into SQL (Test2SQL) is a longstanding challenge at the intersection of natural language understanding and structured data access. While large language models (LLMs) have significantly improved fluency in SQL generation, producing correct and executable SQL--particularly for complex queries--remains a bottleneck. We present Arctic-Text2SQL-R1, a reinforcement learning (RL) framework and model family designed to generate accurate, executable SQL using a lightweight reward signal based solely on execution correctness. Our approach avoids brittle intermediate supervision and complex reward shaping, promoting stable training and alignment with the end task. Combined with carefully curated data, strong supervised initialization, and effective training practices, Arctic-Text2SQL-R1 achieves state-of-the-art execution accuracy across six diverse Test2SQL benchmarks, including the top position on the BIRD leaderboard. Notably, our 7B model outperforms prior 70B-class systems, highlighting the framework's scalability and efficiency. We further demonstrate inference-time robustness through simple extensions like value retrieval and majority voting. Extensive experiments and ablation studies offer both positive and negative insights, providing practical guidance for future Test2SQL research. | [
{
"version": "v1",
"created": "Thu, 22 May 2025 23:33:47 GMT"
}
] | 2025-05-28T00:00:00 | [
[
"Yao",
"Zhewei",
""
],
[
"Sun",
"Guoheng",
""
],
[
"Borchmann",
"Lukasz",
""
],
[
"Shen",
"Zheyu",
""
],
[
"Deng",
"Minghang",
""
],
[
"Zhai",
"Bohan",
""
],
[
"Zhang",
"Hao",
""
],
[
"Li",
"Ang",
""
],
[
"He",
"Yuxiong",
""
]
] |
2505.20316 | Yingpeng Du | Yingpeng Du, Tianjun Wei, Zhu Sun, Jie Zhang | Reinforcement Speculative Decoding for Fast Ranking | 21 pages, 5 figures, 5 table | null | null | null | cs.AI | http://arxiv.org/licenses/nonexclusive-distrib/1.0/ | Large Language Models (LLMs) have been widely adopted in ranking systems such as information retrieval (IR) systems and recommender systems (RSs). To alleviate the latency of auto-regressive decoding, some studies explore the single (first) token decoding for ranking approximation, but they suffer from severe degradation in tail positions. Although speculative decoding (SD) methods can be a remedy with verification at different positions, they face challenges in ranking systems due to their left-to-right decoding paradigm. Firstly, ranking systems require strict latency constraints, but verification rounds in SD methods remain agnostic; Secondly, SD methods usually discard listwise ranking knowledge about unaccepted items in previous rounds, hindering future multi-token prediction, especially when candidate tokens are the unaccepted items. In this paper, we propose a Reinforcement Speculative Decoding method for fast ranking inference of LLMs. To meet the ranking systems' latency requirement, we propose an up-to-down decoding paradigm that employs an agent to iteratively modify the ranking sequence under a constrained budget. Specifically, we design a ranking-tailored policy optimization, actively exploring optimal multi-round ranking modification policy verified by LLMs via reinforcement learning (RL). To better approximate the target LLM under the constrained budget, we trigger the agent fully utilizing the listwise ranking knowledge about all items verified by LLMs across different rounds in RL, enhancing the modification policy of the agent. More importantly, we demonstrate the theoretical robustness and advantages of our paradigm and implementation. Experiments on both IR and RS tasks show the effectiveness of our proposed method. | [
{
"version": "v1",
"created": "Fri, 23 May 2025 02:25:26 GMT"
}
] | 2025-05-28T00:00:00 | [
[
"Du",
"Yingpeng",
""
],
[
"Wei",
"Tianjun",
""
],
[
"Sun",
"Zhu",
""
],
[
"Zhang",
"Jie",
""
]
] |
2505.20317 | Roni Dey | Roni Dey and Dipanwita Mondal and Sudipta Das and Varchaswi K. S. Kashyap and Bedangadas Mohanty | Measurement of cosmic muon-induced events in an HPGe detector using time-coincidence technique | null | null | null | null | physics.ins-det astro-ph.IM nucl-ex | http://creativecommons.org/licenses/by-nc-sa/4.0/ | Detailed understanding and suppression of backgrounds are among the key challenges faced by Coherent Elastic Neutrino-Nucleus Scattering (CE\ensuremath{\nu}NS) experiments. The sensitivity of these experiments is largely determined by the background levels arising from various sources. Above-ground and shallow-overburden neutrino experiments typically employ passive shielding, primarily composed of lead (Pb), to suppress environmental $\gamma$ background. However, such shielding can introduce additional backgrounds that are particularly challenging for CE\ensuremath{\nu}NS experiments. These backgrounds arise mainly from $\gamma$ and neutrons produced by cosmic muon interactions in the shielding, and their contribution can become significant depending on the amount of Pb shielding used. In the current work, we measure the yield of secondary particles originating from Pb as a result of high-energy cosmic muon interaction, using a high-purity germanium (HPGe) detector and plastic scintillators. A time-coincidence technique is used to identify and reject these secondary background events from the experimental data. The obtained mean characteristic time of these residual background events is 11 $\pm$ 4 $\mu$s, which is consistent with the Geant4-based MC simulation result of 11 $\pm$ 1 $\mu$s. The measured efficiency-corrected rate of muon-induced events in the HPGe detector is 34 $\pm$ 1 (stat.) $\pm$ 3 (sys.) day$^{-1}$kg$^{-1}$ within the energy range of 30 keV to 2000 keV. The yield of muon-induced secondary backgrounds in 10 cm thick Pb shielding is evaluated to be $(11 \pm 1 (\text{stat.}) \pm 1 (\text{sys.}))$ secondary events$\thinspace\text{kg}^{-1}\thinspace\mathrm{m^{-2}}\thinspace\text{muon}^{-1}$ at sea level. | [
{
"version": "v1",
"created": "Fri, 23 May 2025 11:33:43 GMT"
}
] | 2025-05-28T00:00:00 | [
[
"Dey",
"Roni",
""
],
[
"Mondal",
"Dipanwita",
""
],
[
"Das",
"Sudipta",
""
],
[
"Kashyap",
"Varchaswi K. S.",
""
],
[
"Mohanty",
"Bedangadas",
""
]
] |
2505.20318 | Wang Cai | Wang Cai, Hsiu-Yuan Huang, Zhixiang Wang, Yunfang Wu | Beyond Demonstrations: Dynamic Vector Construction from Latent Representations | null | null | null | null | cs.CL cs.AI | http://creativecommons.org/licenses/by/4.0/ | In-Context derived Vector (ICV) methods extract task-relevant representations from large language models (LLMs) and reinject them during inference, achieving comparable performance to few-shot In-Context Learning (ICL) without repeated demonstration processing. However, existing ICV methods remain sensitive to ICL-specific factors, often use coarse or semantically fragmented representations as the source of the vector, and rely on heuristic-based injection positions, limiting their applicability.
To address these issues, we propose Dynamic Vector (DyVec), which incorporates an Exhaustive Query Rotation (EQR) strategy to extract robust semantically aggregated latent representations by mitigating variance introduced by ICL. It then applies Dynamic Latent Segmentation and Injection to adaptively partition representations based on task complexity and leverages REINFORCE-based optimization to learn optimal injection positions for each segment.
Experiments results show that DyVec outperforms few-shot ICL, LoRA, and prior ICV baselines. Further analysis highlights the effectiveness of dynamically segmenting and injecting semantically aggregated latent representations. DyVec provides a lightweight and data-efficient solution for inference-time task adaptation. | [
{
"version": "v1",
"created": "Fri, 23 May 2025 12:13:50 GMT"
}
] | 2025-05-28T00:00:00 | [
[
"Cai",
"Wang",
""
],
[
"Huang",
"Hsiu-Yuan",
""
],
[
"Wang",
"Zhixiang",
""
],
[
"Wu",
"Yunfang",
""
]
] |
2505.20319 | Julia Gersey | Joseph Maffetone, Julia Gersey, Pei Zhang | ZV-Sim: Probabilistic Simulation Framework for Pre-emergent Novel Zoonose Tracking | 5 pages | null | null | null | q-bio.QM cs.SY eess.SY | http://creativecommons.org/licenses/by/4.0/ | ZV-Sim is an open-source, modular Python framework for probabilistic simulation and analysis of pre-emergent novel zoonotic diseases using pervasive sensing data. It incorporates customizable Human and Animal Presence agents that leverage known and simulated location data, contact networks, and illness reports to assess and predict disease origins and spread. The framework supports Monte Carlo experiments to analyze outcomes with various user-defined movement and probability models. Although initial models are basic and illustrative, ZV-Sim's extensible design facilitates the integration of more sophisticated models as richer data become available, enhancing future capabilities in zoonotic disease tracking. The source code is publicly available \href{https://github.com/jmaff/zv-sim}{\underline{\textit{here}}}. | [
{
"version": "v1",
"created": "Fri, 23 May 2025 15:14:00 GMT"
}
] | 2025-05-28T00:00:00 | [
[
"Maffetone",
"Joseph",
""
],
[
"Gersey",
"Julia",
""
],
[
"Zhang",
"Pei",
""
]
] |
2505.20320 | Dhavalkumar Patel | Satya Narayana Cheetirala, Ganesh Raut, Dhavalkumar Patel, Fabio Sanatana, Robert Freeman, Matthew A Levin, Girish N. Nadkarni, Omar Dawkins, Reba Miller, Randolph M. Steinhagen, Eyal Klang, Prem Timsina | Less Context, Same Performance: A RAG Framework for Resource-Efficient LLM-Based Clinical NLP | null | null | null | null | cs.CL cs.AI | http://creativecommons.org/licenses/by/4.0/ | Long text classification is challenging for Large Language Models (LLMs) due to token limits and high computational costs. This study explores whether a Retrieval Augmented Generation (RAG) approach using only the most relevant text segments can match the performance of processing entire clinical notes with large context LLMs. We begin by splitting clinical documents into smaller chunks, converting them into vector embeddings, and storing these in a FAISS index. We then retrieve the top 4,000 words most pertinent to the classification query and feed these consolidated segments into an LLM. We evaluated three LLMs (GPT4o, LLaMA, and Mistral) on a surgical complication identification task. Metrics such as AUC ROC, precision, recall, and F1 showed no statistically significant differences between the RAG based approach and whole-text processing (p > 0.05p > 0.05). These findings indicate that RAG can significantly reduce token usage without sacrificing classification accuracy, providing a scalable and cost effective solution for analyzing lengthy clinical documents. | [
{
"version": "v1",
"created": "Fri, 23 May 2025 16:13:08 GMT"
}
] | 2025-05-28T00:00:00 | [
[
"Cheetirala",
"Satya Narayana",
""
],
[
"Raut",
"Ganesh",
""
],
[
"Patel",
"Dhavalkumar",
""
],
[
"Sanatana",
"Fabio",
""
],
[
"Freeman",
"Robert",
""
],
[
"Levin",
"Matthew A",
""
],
[
"Nadkarni",
"Girish N.",
""
],
[
"Dawkins",
"Omar",
""
],
[
"Miller",
"Reba",
""
],
[
"Steinhagen",
"Randolph M.",
""
],
[
"Klang",
"Eyal",
""
],
[
"Timsina",
"Prem",
""
]
] |
2505.20321 | Mathew Koretsky | Mathew J. Koretsky, Maya Willey, Adi Asija, Owen Bianchi, Chelsea X. Alvarado, Tanay Nayak, Nicole Kuznetsov, Sungwon Kim, Mike A. Nalls, Daniel Khashabi, Faraz Faghri | BiomedSQL: Text-to-SQL for Scientific Reasoning on Biomedical Knowledge Bases | Under Review | null | null | null | cs.CL cs.AI cs.LG | http://creativecommons.org/licenses/by/4.0/ | Biomedical researchers increasingly rely on large-scale structured databases for complex analytical tasks. However, current text-to-SQL systems often struggle to map qualitative scientific questions into executable SQL, particularly when implicit domain reasoning is required. We introduce BiomedSQL, the first benchmark explicitly designed to evaluate scientific reasoning in text-to-SQL generation over a real-world biomedical knowledge base. BiomedSQL comprises 68,000 question/SQL query/answer triples grounded in a harmonized BigQuery knowledge base that integrates gene-disease associations, causal inference from omics data, and drug approval records. Each question requires models to infer domain-specific criteria, such as genome-wide significance thresholds, effect directionality, or trial phase filtering, rather than rely on syntactic translation alone. We evaluate a range of open- and closed-source LLMs across prompting strategies and interaction paradigms. Our results reveal a substantial performance gap: GPT-o3-mini achieves 59.0% execution accuracy, while our custom multi-step agent, BMSQL, reaches 62.6%, both well below the expert baseline of 90.0%. BiomedSQL provides a new foundation for advancing text-to-SQL systems capable of supporting scientific discovery through robust reasoning over structured biomedical knowledge bases. Our dataset is publicly available at https://huggingface.co/datasets/NIH-CARD/BiomedSQL, and our code is open-source at https://github.com/NIH-CARD/biomedsql. | [
{
"version": "v1",
"created": "Fri, 23 May 2025 17:58:07 GMT"
}
] | 2025-05-28T00:00:00 | [
[
"Koretsky",
"Mathew J.",
""
],
[
"Willey",
"Maya",
""
],
[
"Asija",
"Adi",
""
],
[
"Bianchi",
"Owen",
""
],
[
"Alvarado",
"Chelsea X.",
""
],
[
"Nayak",
"Tanay",
""
],
[
"Kuznetsov",
"Nicole",
""
],
[
"Kim",
"Sungwon",
""
],
[
"Nalls",
"Mike A.",
""
],
[
"Khashabi",
"Daniel",
""
],
[
"Faghri",
"Faraz",
""
]
] |
2505.20322 | Mengru Wang | Mengru Wang, Ziwen Xu, Shengyu Mao, Shumin Deng, Zhaopeng Tu, Huajun Chen, Ningyu Zhang | Beyond Prompt Engineering: Robust Behavior Control in LLMs via Steering Target Atoms | null | null | null | null | cs.CL cs.AI cs.CV cs.IR cs.LG | http://arxiv.org/licenses/nonexclusive-distrib/1.0/ | Precise control over language model generation is vital for ensuring both safety and reliability. Although prompt engineering and steering are commonly used to intervene in model behaviors, the vast number of parameters in models often results in highly intertwined internal representations. This interdependency can limit control precision and sometimes lead to unintended side effects. Recent research has explored the use of sparse autoencoders (SAE) to disentangle knowledge in high-dimensional spaces for steering. However, these applications have been limited to toy tasks owing to the nontrivial issue of locating atomic knowledge components. In this paper, we propose Steering Target Atoms (STA), a novel method that isolates and manipulates disentangled knowledge components to enhance safety. Comprehensive experiments demonstrate the effectiveness of our approach. Further analysis reveals that steering exhibits superior robustness and flexibility, particularly in adversarial scenarios. We also apply the steering strategy to the large reasoning model, confirming its effectiveness in precise reasoning control. | [
{
"version": "v1",
"created": "Fri, 23 May 2025 17:59:18 GMT"
}
] | 2025-05-28T00:00:00 | [
[
"Wang",
"Mengru",
""
],
[
"Xu",
"Ziwen",
""
],
[
"Mao",
"Shengyu",
""
],
[
"Deng",
"Shumin",
""
],
[
"Tu",
"Zhaopeng",
""
],
[
"Chen",
"Huajun",
""
],
[
"Zhang",
"Ningyu",
""
]
] |
2505.20323 | Shahriar Noroozizadeh | Shahriar Noroozizadeh, Sayantan Kumar, George H. Chen, Jeremy C. Weiss | PMOA-TTS: Introducing the PubMed Open Access Textual Times Series Corpus | null | null | null | null | cs.CL cs.AI cs.LG | http://creativecommons.org/licenses/by/4.0/ | Understanding temporal dynamics in clinical narratives is essential for modeling patient trajectories, yet large-scale temporally annotated resources remain limited. We present PMOA-TTS, the first openly available dataset of 124,699 PubMed Open Access (PMOA) case reports, each converted into structured (event, time) timelines via a scalable LLM-based pipeline. Our approach combines heuristic filtering with Llama 3.3 to identify single-patient case reports, followed by prompt-driven extraction using Llama 3.3 and DeepSeek R1, resulting in over 5.6 million timestamped clinical events. To assess timeline quality, we evaluate against a clinician-curated reference set using three metrics: (i) event-level matching (80% match at a cosine similarity threshold of 0.1), (ii) temporal concordance (c-index > 0.90), and (iii) Area Under the Log-Time CDF (AULTC) for timestamp alignment. Corpus-level analysis shows wide diagnostic and demographic coverage. In a downstream survival prediction task, embeddings from extracted timelines achieve time-dependent concordance indices up to 0.82 $\pm$ 0.01, demonstrating the predictive value of temporally structured narratives. PMOA-TTS provides a scalable foundation for timeline extraction, temporal reasoning, and longitudinal modeling in biomedical NLP. The dataset is available at: https://huggingface.co/datasets/snoroozi/pmoa-tts . | [
{
"version": "v1",
"created": "Fri, 23 May 2025 18:01:09 GMT"
}
] | 2025-05-28T00:00:00 | [
[
"Noroozizadeh",
"Shahriar",
""
],
[
"Kumar",
"Sayantan",
""
],
[
"Chen",
"George H.",
""
],
[
"Weiss",
"Jeremy C.",
""
]
] |
2505.20324 | Tevfik Kosar | Md Arman Islam, Devi Varaprasad Jonnala, Ritika Rekhi, Pratik Pokharel, Siddharth Cilamkoti, Asif Imran, Tevfik Kosar, Bekir Turkkan | Evaluating the Energy-Efficiency of the Code Generated by LLMs | null | null | null | null | cs.SE cs.AI | http://creativecommons.org/licenses/by/4.0/ | As the quality of code generated by Large Language Models (LLMs) improves, their adoption in the software industry for automated code generation continues to grow. Researchers primarily focus on enhancing the functional correctness of the generated code while commonly overlooking its energy efficiency and environmental impact. This paper investigates the energy efficiency of the code generated by 20 popular LLMs for 878 programming problems of varying difficulty levels and diverse algorithmic categories selected from the LeetCode platform by comparing them against canonical human-written solutions. Although LLMs can produce functionally correct results in most cases, our findings show that the performance and energy efficiency of LLM-produced solutions are often far below those of human-written solutions. Among the studied LLMs, DeepSeek-v3 and GPT-4o generate the most energy-efficient code, whereas Grok-2 and Gemini-1.5-Pro are among the least energy-efficient models. On average, human-generated canonical solutions are approximately 1.17 times more energy efficient than DeepSeek-v3, 1.21 times more energy efficient than GPT-4o, and over 2 times more energy efficient than Grok-2 and Gemini-1.5-Pro. For specific algorithmic groups such as dynamic programming, backtracking, and bit manipulation, LLM-generated code can consume up to 450 times more energy than human-generated canonical solutions. | [
{
"version": "v1",
"created": "Fri, 23 May 2025 18:13:27 GMT"
}
] | 2025-05-28T00:00:00 | [
[
"Islam",
"Md Arman",
""
],
[
"Jonnala",
"Devi Varaprasad",
""
],
[
"Rekhi",
"Ritika",
""
],
[
"Pokharel",
"Pratik",
""
],
[
"Cilamkoti",
"Siddharth",
""
],
[
"Imran",
"Asif",
""
],
[
"Kosar",
"Tevfik",
""
],
[
"Turkkan",
"Bekir",
""
]
] |
2505.20325 | Amirhosein Ghasemabadi | Amirhosein Ghasemabadi, Keith G. Mills, Baochun Li, Di Niu | Guided by Gut: Efficient Test-Time Scaling with Reinforced Intrinsic Confidence | null | null | null | null | cs.CL cs.AI | http://creativecommons.org/licenses/by/4.0/ | Test-Time Scaling (TTS) methods for enhancing Large Language Model (LLM) reasoning often incur substantial computational costs, primarily due to extensive reliance on external Process Reward Models (PRMs) or sampling methods like Best-of-N (BoN). This paper introduces Guided by Gut (GG), an efficient self-guided TTS framework that achieves PRM-level performance without costly external verifier models. Our method employs a lightweight tree search guided solely by intrinsic LLM signals, token-level confidence and step novelty. One critical innovation is improving the reliability of internal confidence estimates via a targeted reinforcement learning fine-tuning phase. Empirical evaluations on challenging mathematical reasoning benchmarks demonstrate that GG enables smaller models (e.g., 1.5B parameters) to achieve accuracy matching or surpassing significantly larger models (e.g., 32B-70B parameters), while reducing GPU memory usage by up to 10x. Compared to PRM-based methods, GG achieves comparable accuracy with 8x faster inference speeds and 4-5x lower memory usage. Additionally, GG reduces KV cache memory usage by approximately 50% compared to the BoN strategy, facilitating more efficient and practical deployment of TTS techniques. | [
{
"version": "v1",
"created": "Fri, 23 May 2025 18:19:09 GMT"
}
] | 2025-05-28T00:00:00 | [
[
"Ghasemabadi",
"Amirhosein",
""
],
[
"Mills",
"Keith G.",
""
],
[
"Li",
"Baochun",
""
],
[
"Niu",
"Di",
""
]
] |
2505.20326 | Avinash Madasu | Avinash Madasu, Vasudev Lal, Phillip Howard | Cultural Awareness in Vision-Language Models: A Cross-Country Exploration | null | null | null | null | cs.CY cs.AI cs.CL cs.CV | http://creativecommons.org/licenses/by/4.0/ | Vision-Language Models (VLMs) are increasingly deployed in diverse cultural contexts, yet their internal biases remain poorly understood. In this work, we propose a novel framework to systematically evaluate how VLMs encode cultural differences and biases related to race, gender, and physical traits across countries. We introduce three retrieval-based tasks: (1) Race to Country retrieval, which examines the association between individuals from specific racial groups (East Asian, White, Middle Eastern, Latino, South Asian, and Black) and different countries; (2) Personal Traits to Country retrieval, where images are paired with trait-based prompts (e.g., Smart, Honest, Criminal, Violent) to investigate potential stereotypical associations; and (3) Physical Characteristics to Country retrieval, focusing on visual attributes like skinny, young, obese, and old to explore how physical appearances are culturally linked to nations. Our findings reveal persistent biases in VLMs, highlighting how visual representations may inadvertently reinforce societal stereotypes. | [
{
"version": "v1",
"created": "Fri, 23 May 2025 18:47:52 GMT"
}
] | 2025-05-28T00:00:00 | [
[
"Madasu",
"Avinash",
""
],
[
"Lal",
"Vasudev",
""
],
[
"Howard",
"Phillip",
""
]
] |
2505.20327 | Kateryna Morozovska | Aurora Poggi, Giuseppe Alessio D'Inverno, Hjalmar Brismar, Ozan \"Oktem, Matthieu Barreau, Kateryna Morozovska | Data-driven multi-agent modelling of calcium interactions in cell culture: PINN vs Regularized Least-squares | null | null | null | null | q-bio.QM cs.AI | http://creativecommons.org/licenses/by-nc-nd/4.0/ | Data-driven discovery of dynamics in biological systems allows for better observation and characterization of processes, such as calcium signaling in cell culture. Recent advancements in techniques allow the exploration of previously unattainable insights of dynamical systems, such as the Sparse Identification of Non-Linear Dynamics (SINDy), overcoming the limitations of more classic methodologies. The latter requires some prior knowledge of an effective library of candidate terms, which is not realistic for a real case study. Using inspiration from fields like traffic density estimation and control theory, we propose a methodology for characterization and performance analysis of calcium delivery in a family of cells. In this work, we compare the performance of the Constrained Regularized Least-Squares Method (CRLSM) and Physics-Informed Neural Networks (PINN) for system identification and parameter discovery for governing ordinary differential equations (ODEs). The CRLSM achieves a fairly good parameter estimate and a good data fit when using the learned parameters in the Consensus problem. On the other hand, despite the initial hypothesis, PINNs fail to match the CRLSM performance and, under the current configuration, do not provide fair parameter estimation. However, we have only studied a limited number of PINN architectures, and it is expected that additional hyperparameter tuning, as well as uncertainty quantification, could significantly improve the performance in future works. | [
{
"version": "v1",
"created": "Fri, 23 May 2025 19:41:12 GMT"
}
] | 2025-05-28T00:00:00 | [
[
"Poggi",
"Aurora",
""
],
[
"D'Inverno",
"Giuseppe Alessio",
""
],
[
"Brismar",
"Hjalmar",
""
],
[
"Öktem",
"Ozan",
""
],
[
"Barreau",
"Matthieu",
""
],
[
"Morozovska",
"Kateryna",
""
]
] |
2505.20328 | Junfang Wang | J.-F. Wang and G. Qin | A Cosmic Ray Acceleration Mechanism Based on Background Flow Velocity Inhomogeneities Yielding Power-Law Spectra | null | null | null | null | astro-ph.HE astro-ph.SR | http://creativecommons.org/licenses/by/4.0/ | In this article, momentum transport generated by the combined effects of pitch-angle diffusion and Background Flow Velocity Inhomogeneities (BFVIs) is proposed to obtain a cosmic rays acceleration mechanism, starting from the well-known focusing equation describing particle diffusion and acceleration. The inhomogeneities of background flow velocity is ubiquitous in astrophysical environment. The isotropic distribution function equation of charged energetic particles is derived, and its solution is obtained, demonstrating the form of momentum power laws of cosmic rays. In addition, if it is assumed that cosmic rays penetrate compressible plasma waves or turbulence, for quasi-steady states, the spectral index $\delta$ of the momentum power law spectrum of cosmic rays is found to be in the range $[-5, -3]$, which includes the observed power law indices of galactic cosmic rays. The results obtained in this article demonstrate that the mechanism proposed in this article, along with shock acceleration, may also contribute to the acceleration of galactic cosmic rays. Furthermore, when momentum convection effect and higher-order momentum derivative terms are considered, the indices of power laws should be smaller than $-5$. This may explain the power laws of solar energetic particle events. | [
{
"version": "v1",
"created": "Sat, 24 May 2025 01:04:58 GMT"
}
] | 2025-05-28T00:00:00 | [
[
"Wang",
"J. -F.",
""
],
[
"Qin",
"G.",
""
]
] |
2505.20329 | Ian McCulloh | Ian McCulloh, Pedro Rodriguez, Srivaths Kumar, Manu Gupta, Viplove Raj Sharma, Benjamin Johnson, and Anthony N. Johnson | Generative AI in Computer Science Education: Accelerating Python Learning with ChatGPT | null | null | null | null | cs.CY | http://creativecommons.org/licenses/by/4.0/ | The increasing demand for digital literacy and artificial intelligence (AI) fluency in the workforce has highlighted the need for scalable, efficient programming instruction. This study evaluates the effectiveness of integrating generative AI, specifically OpenAIs ChatGPT, into a self-paced Python programming module embedded within a sixteen-week professional training course on applied generative AI. A total of 86 adult learners with varying levels of programming experience completed asynchronous Python instruction in Weeks three and four, using ChatGPT to generate, interpret, and debug code. Python proficiency and general coding knowledge was assessed across 30 different assessments during the first 13 weeks of the course through timed, code-based evaluations. A mixed-design ANOVA revealed that learners without prior programming experience scored significantly lower than their peers on early assessments. However, following the completion of the accelerated Python instruction module, these group differences were no longer statistically significant,, indicating that the intervention effectively closed initial performance gaps and supported proficiency gains across all learner groups. These findings suggest that generative AI can support accelerated learning outcomes and reduce entry barriers for learners with no prior coding background. While ChatGPT effectively facilitated foundational skill acquisition, the study also highlights the importance of balancing AI assistance with opportunities for independent problem-solving. The results support the potential of AI-augmented instruction as a scalable model for reskilling in the digital economy. | [
{
"version": "v1",
"created": "Sat, 24 May 2025 03:13:46 GMT"
}
] | 2025-05-28T00:00:00 | [
[
"McCulloh",
"Ian",
""
],
[
"Rodriguez",
"Pedro",
""
],
[
"Kumar",
"Srivaths",
""
],
[
"Gupta",
"Manu",
""
],
[
"Sharma",
"Viplove Raj",
""
],
[
"Johnson",
"Benjamin",
""
],
[
"Johnson",
"Anthony N.",
""
]
] |
2505.20330 | Zhijian Ou | Yunfu Song, Zhijian Ou | Joint-stochastic-approximation Random Fields with Application to Semi-supervised Learning | ICML 2018 submission. arXiv admin note: text overlap with arXiv:1808.01630, arXiv:2505.18558 | null | null | null | cs.LG stat.ML | http://arxiv.org/licenses/nonexclusive-distrib/1.0/ | Our examination of deep generative models (DGMs) developed for semi-supervised learning (SSL), mainly GANs and VAEs, reveals two problems. First, mode missing and mode covering phenomenons are observed in genertion with GANs and VAEs. Second, there exists an awkward conflict between good classification and good generation in SSL by employing directed generative models. To address these problems, we formally present joint-stochastic-approximation random fields (JRFs) -- a new family of algorithms for building deep undirected generative models, with application to SSL. It is found through synthetic experiments that JRFs work well in balancing mode covering and mode missing, and match the empirical data distribution well. Empirically, JRFs achieve good classification results comparable to the state-of-art methods on widely adopted datasets -- MNIST, SVHN, and CIFAR-10 in SSL, and simultaneously perform good generation. | [
{
"version": "v1",
"created": "Sat, 24 May 2025 07:04:32 GMT"
}
] | 2025-05-28T00:00:00 | [
[
"Song",
"Yunfu",
""
],
[
"Ou",
"Zhijian",
""
]
] |
2505.20331 | Gamal G.L. Nashed | G.G.L. Nashed and Waleed El Hanafy | Stable soliton dark matter wormhole in non-minimally coupled $f({\cal Q},{\cal T})$ gravity | 23 pages, 12 figures | null | null | null | gr-qc hep-th | http://creativecommons.org/licenses/by/4.0/ | We show that non-minimal coupling between matter and geometry can indeed help in constructing stable, traversable, wormholes (WHs) without requiring exotic matter under certain conditions. In models like $f({\cal Q},{\cal T})={\cal Q}+\beta {\cal T}$ gravity, where ${\cal Q}$ is the non-metricity scalar, and ${\cal T}$ is the trace of the energy-momentum tensor, the coupling between matter and geometry introduces additional degrees of freedom in terms of the parameter $\beta$. These can mimic the effects of exotic matter or even replace it entirely under specific parameter choice. The analysis involves deriving WH shape functions based on two dark matter (DM) density profiles: a solitonic core at the center of DM halos, and the outer halo follows the universal Navarro-Frenk-White (NFW) density profile of cold DM (CDM). The wormhole solutions derived in these models satisfy important geometric conditions like: Flaring-out condition (necessary for traversability) and asymptotic flatness condition. For large positive coupling parameter, the null energy condition (NEC) can be satisfied at the wormhole throat, meaning exotic matter is not needed, while the wormhole is no longer Lorentzian and the flaring-out condition is broken. However, for large negative coupling parameter, the NEC can be satisfied, allowing for healthy wormholes without exotic matter, provided the coupling strength remains within certain bounds. In the latter case, the NEC is broken only effectively. We investigate the stability of the obtained wormhole solutions by virtue of a modified version of Tolman-Oppenheimer-Volkoff (TOV) equation, which includes a new force due to matter-geometry non-minimal, showing that these wormholes can be dynamically stable. | [
{
"version": "v1",
"created": "Sat, 24 May 2025 09:01:13 GMT"
}
] | 2025-05-28T00:00:00 | [
[
"Nashed",
"G. G. L.",
""
],
[
"Hanafy",
"Waleed El",
""
]
] |
2505.20332 | Neil Chaudhary | Neil Chaudhary, Zaynah Dhunny | An Artificial Intelligence Model for Early Stage Breast Cancer Detection from Biopsy Images | null | null | null | null | eess.IV cs.LG | http://creativecommons.org/licenses/by/4.0/ | Accurate identification of breast cancer types plays a critical role in guiding treatment decisions and improving patient outcomes. This paper presents an artificial intelligence enabled tool designed to aid in the identification of breast cancer types using histopathological biopsy images. Traditionally additional tests have to be done on women who are detected with breast cancer to find out the types of cancer it is to give the necessary cure. Those tests are not only invasive but also delay the initiation of treatment and increase patient burden. The proposed model utilizes a convolutional neural network (CNN) architecture to distinguish between benign and malignant tissues as well as accurate subclassification of breast cancer types. By preprocessing the images to reduce noise and enhance features, the model achieves reliable levels of classification performance. Experimental results on such datasets demonstrate the model's effectiveness, outperforming several existing solutions in terms of accuracy, precision, recall, and F1-score. The study emphasizes the potential of deep learning techniques in clinical diagnostics and offers a promising tool to assist pathologists in breast cancer classification. | [
{
"version": "v1",
"created": "Sat, 24 May 2025 09:11:50 GMT"
}
] | 2025-05-28T00:00:00 | [
[
"Chaudhary",
"Neil",
""
],
[
"Dhunny",
"Zaynah",
""
]
] |
2505.20333 | Yukun Zhang | Yukun Zhang, Qi Dong | Multi-Scale Manifold Alignment: A Unified Framework for Enhanced Explainability of Large Language Models | null | null | null | null | cs.CL cs.AI | http://creativecommons.org/licenses/by/4.0/ | Recent advances in Large Language Models (LLMs) have achieved strong performance, yet their internal reasoning remains opaque, limiting interpretability and trust in critical applications. We propose a novel Multi_Scale Manifold Alignment framework that decomposes the latent space into global, intermediate, and local semantic manifolds capturing themes, context, and word-level details. Our method introduces cross_scale mapping functions that jointly enforce geometric alignment (e.g., Procrustes analysis) and information preservation (via mutual information constraints like MINE or VIB). We further incorporate curvature regularization and hyperparameter tuning for stable optimization. Theoretical analysis shows that alignment error, measured by KL divergence, can be bounded under mild assumptions. This framework offers a unified explanation of how LLMs structure multi-scale semantics, advancing interpretability and enabling applications such as bias detection and robustness enhancement. | [
{
"version": "v1",
"created": "Sat, 24 May 2025 10:25:58 GMT"
}
] | 2025-05-28T00:00:00 | [
[
"Zhang",
"Yukun",
""
],
[
"Dong",
"Qi",
""
]
] |
2505.20334 | Yixuan Wang | Yixuan Wang, Shiyu Ji, Yijun Liu, Yuzhuang Xu, Yang Xu, Qingfu Zhu, Wanxiang Che | Lookahead Q-Cache: Achieving More Consistent KV Cache Eviction via Pseudo Query | 16 pages, 10 figures | null | null | null | cs.CL cs.AI | http://creativecommons.org/licenses/by-sa/4.0/ | Large language models (LLMs) rely on key-value cache (KV cache) to accelerate decoding by reducing redundant computations. However, the KV cache memory usage grows substantially with longer text sequences, posing challenges for efficient deployment. Existing KV cache eviction methods prune tokens using prefilling-stage attention scores, causing inconsistency with actual inference queries, especially under tight memory budgets. In this paper, we propose Lookahead Q-Cache (LAQ), a novel eviction framework that generates low-cost pseudo lookahead queries to better approximate the true decoding-stage queries. By using these lookahead queries as the observation window for importance estimation, LAQ achieves more consistent and accurate KV cache eviction aligned with real inference scenarios. Experimental results on LongBench and Needle-in-a-Haystack benchmarks show that LAQ outperforms existing methods across various budget levels, achieving a 1 $\sim$ 4 point improvement on LongBench under limited cache budget. Moreover, LAQ is complementary to existing approaches and can be flexibly combined to yield further improvements. | [
{
"version": "v1",
"created": "Sat, 24 May 2025 10:34:38 GMT"
}
] | 2025-05-28T00:00:00 | [
[
"Wang",
"Yixuan",
""
],
[
"Ji",
"Shiyu",
""
],
[
"Liu",
"Yijun",
""
],
[
"Xu",
"Yuzhuang",
""
],
[
"Xu",
"Yang",
""
],
[
"Zhu",
"Qingfu",
""
],
[
"Che",
"Wanxiang",
""
]
] |
2505.20335 | Shangzhe Li | Zishun Yu, Shangzhe Li, Xinhua Zhang | Language Model Distillation: A Temporal Difference Imitation Learning Perspective | null | null | null | null | cs.CL cs.AI | http://creativecommons.org/licenses/by/4.0/ | Large language models have led to significant progress across many NLP tasks, although their massive sizes often incur substantial computational costs. Distillation has become a common practice to compress these large and highly capable models into smaller, more efficient ones. Many existing language model distillation methods can be viewed as behavior cloning from the perspective of imitation learning or inverse reinforcement learning. This viewpoint has inspired subsequent studies that leverage (inverse) reinforcement learning techniques, including variations of behavior cloning and temporal difference learning methods. Rather than proposing yet another specific temporal difference method, we introduce a general framework for temporal difference-based distillation by exploiting the distributional sparsity of the teacher model. Specifically, it is often observed that language models assign most probability mass to a small subset of tokens. Motivated by this observation, we design a temporal difference learning framework that operates on a reduced action space (a subset of vocabulary), and demonstrate how practical algorithms can be derived and the resulting performance improvements. | [
{
"version": "v1",
"created": "Sat, 24 May 2025 12:17:12 GMT"
}
] | 2025-05-28T00:00:00 | [
[
"Yu",
"Zishun",
""
],
[
"Li",
"Shangzhe",
""
],
[
"Zhang",
"Xinhua",
""
]
] |
2505.20336 | Wanli Jiang | Yu Zhang, Wanli Jiang, and Zhengyu Yang | MOSLIM:Align with diverse preferences in prompts through reward classification | null | null | null | null | cs.CL cs.AI | http://arxiv.org/licenses/nonexclusive-distrib/1.0/ | The multi-objective alignment of Large Language Models (LLMs) is essential for ensuring foundational models conform to diverse human preferences. Current research in this field typically involves either multiple policies or multiple reward models customized for various preferences, or the need to train a preference-specific supervised fine-tuning (SFT) model. In this work, we introduce a novel multi-objective alignment method, MOSLIM, which utilizes a single reward model and policy model to address diverse objectives. MOSLIM provides a flexible way to control these objectives through prompting and does not require preference training during SFT phase, allowing thousands of off-the-shelf models to be directly utilized within this training framework. MOSLIM leverages a multi-head reward model that classifies question-answer pairs instead of scoring them and then optimize policy model with a scalar reward derived from a mapping function that converts classification results from reward model into reward scores. We demonstrate the efficacy of our proposed method across several multi-objective benchmarks and conduct ablation studies on various reward model sizes and policy optimization methods. The MOSLIM method outperforms current multi-objective approaches in most results while requiring significantly fewer GPU computing resources compared with existing policy optimization methods. | [
{
"version": "v1",
"created": "Sat, 24 May 2025 12:22:21 GMT"
}
] | 2025-05-28T00:00:00 | [
[
"Zhang",
"Yu",
""
],
[
"Jiang",
"Wanli",
""
],
[
"Yang",
"Zhengyu",
""
]
] |
2505.20337 | Xin Wang | Xin Wang, Han-Xiao Tao, Re-Bing Wu | Predictive Performance of Deep Quantum Data Re-uploading Models | 34 pages, 15 figures | null | null | null | quant-ph cs.LG | http://arxiv.org/licenses/nonexclusive-distrib/1.0/ | Quantum machine learning models incorporating data re-uploading circuits have garnered significant attention due to their exceptional expressivity and trainability. However, their ability to generate accurate predictions on unseen data, referred to as the predictive performance, remains insufficiently investigated. This study reveals a fundamental limitation in predictive performance when deep encoding layers are employed within the data re-uploading model. Concretely, we theoretically demonstrate that when processing high-dimensional data with limited-qubit data re-uploading models, their predictive performance progressively degenerates to near random-guessing levels as the number of encoding layers increases. In this context, the repeated data uploading cannot mitigate the performance degradation. These findings are validated through experiments on both synthetic linearly separable datasets and real-world datasets. Our results demonstrate that when processing high-dimensional data, the quantum data re-uploading models should be designed with wider circuit architectures rather than deeper and narrower ones. | [
{
"version": "v1",
"created": "Sat, 24 May 2025 13:11:31 GMT"
}
] | 2025-05-28T00:00:00 | [
[
"Wang",
"Xin",
""
],
[
"Tao",
"Han-Xiao",
""
],
[
"Wu",
"Re-Bing",
""
]
] |
2505.20338 | Rohit Gheyi | Cayo Viegas, Rohit Gheyi, M\'arcio Ribeiro | Assessing the Capability of LLMs in Solving POSCOMP Questions | null | null | null | null | cs.CL cs.AI | http://creativecommons.org/licenses/by/4.0/ | Recent advancements in Large Language Models (LLMs) have significantly expanded the capabilities of artificial intelligence in natural language processing tasks. Despite this progress, their performance in specialized domains such as computer science remains relatively unexplored. Understanding the proficiency of LLMs in these domains is critical for evaluating their practical utility and guiding future developments. The POSCOMP, a prestigious Brazilian examination used for graduate admissions in computer science promoted by the Brazlian Computer Society (SBC), provides a challenging benchmark. This study investigates whether LLMs can match or surpass human performance on the POSCOMP exam. Four LLMs - ChatGPT-4, Gemini 1.0 Advanced, Claude 3 Sonnet, and Le Chat Mistral Large - were initially evaluated on the 2022 and 2023 POSCOMP exams. The assessments measured the models' proficiency in handling complex questions typical of the exam. LLM performance was notably better on text-based questions than on image interpretation tasks. In the 2022 exam, ChatGPT-4 led with 57 correct answers out of 69 questions, followed by Gemini 1.0 Advanced (49), Le Chat Mistral (48), and Claude 3 Sonnet (44). Similar trends were observed in the 2023 exam. ChatGPT-4 achieved the highest performance, surpassing all students who took the POSCOMP 2023 exam. LLMs, particularly ChatGPT-4, show promise in text-based tasks on the POSCOMP exam, although image interpretation remains a challenge. Given the rapid evolution of LLMs, we expanded our analysis to include more recent models - o1, Gemini 2.5 Pro, Claude 3.7 Sonnet, and o3-mini-high - evaluated on the 2022-2024 POSCOMP exams. These newer models demonstrate further improvements and consistently surpass both the average and top-performing human participants across all three years. | [
{
"version": "v1",
"created": "Sat, 24 May 2025 13:40:53 GMT"
}
] | 2025-05-28T00:00:00 | [
[
"Viegas",
"Cayo",
""
],
[
"Gheyi",
"Rohit",
""
],
[
"Ribeiro",
"Márcio",
""
]
] |
2505.20339 | Vincent C. M\"uller | Antoni Gomila and Vincent C. M\"uller | Challenges for artificial cognitive systems | null | Journal of Cognitive Science, 13 (4), 453-69 (2012) | null | null | cs.AI cs.HC | http://arxiv.org/licenses/nonexclusive-distrib/1.0/ | The declared goal of this paper is to fill this gap: "... cognitive systems research needs questions or challenges that define progress. The challenges are not (yet more) predictions of the future, but a guideline to what are the aims and what would constitute progress." -- the quotation being from the project description of EUCogII, the project for the European Network for Cognitive Systems within which this formulation of the 'challenges' was originally developed (http://www.eucognition.org). So, we stick out our neck and formulate the challenges for artificial cognitive systems. These challenges are articulated in terms of a definition of what a cognitive system is: a system that learns from experience and uses its acquired knowledge (both declarative and practical) in a flexible manner to achieve its own goals. | [
{
"version": "v1",
"created": "Sat, 24 May 2025 13:49:54 GMT"
}
] | 2025-05-28T00:00:00 | [
[
"Gomila",
"Antoni",
""
],
[
"Müller",
"Vincent C.",
""
]
] |
2505.20340 | Yukun Zhang | Yukun Zhang, Qi Dong | Dynamic Manifold Evolution Theory: Modeling and Stability Analysis of Latent Representations in Large Language Models | null | null | null | null | cs.CL cs.AI | http://creativecommons.org/licenses/by/4.0/ | We introduce Dynamic Manifold Evolution Theory (DMET),a unified framework that models large language model generation as a controlled dynamical system evolving on a low_dimensional semantic manifold. By casting latent_state updates as discrete time Euler approximations of continuous dynamics, we map intrinsic energy_driven flows and context_dependent forces onto Transformer components (residual connections, attention, feed-forward networks). Leveraging Lyapunov stability theory We define three empirical metrics (state continuity, clustering quality, topological persistence) that quantitatively link latent_trajectory properties to text fluency, grammaticality, and semantic coherence. Extensive experiments across decoding parameters validate DMET's predictions and yield principled guidelines for balancing creativity and consistency in text generation. | [
{
"version": "v1",
"created": "Sat, 24 May 2025 14:17:50 GMT"
}
] | 2025-05-28T00:00:00 | [
[
"Zhang",
"Yukun",
""
],
[
"Dong",
"Qi",
""
]
] |
2505.20341 | Rui Liu | Rui Liu, Pu Gao, Jiatian Xi, Berrak Sisman, Carlos Busso, Haizhou Li | Towards Emotionally Consistent Text-Based Speech Editing: Introducing EmoCorrector and The ECD-TSE Dataset | INTERSPEECH2025. Code and audio examples: https://github.com/AI-S2-Lab/EmoCorrector | null | null | null | eess.AS cs.AI cs.CL cs.SD | http://creativecommons.org/licenses/by/4.0/ | Text-based speech editing (TSE) modifies speech using only text, eliminating re-recording. However, existing TSE methods, mainly focus on the content accuracy and acoustic consistency of synthetic speech segments, and often overlook the emotional shifts or inconsistency issues introduced by text changes. To address this issue, we propose EmoCorrector, a novel post-correction scheme for TSE. EmoCorrector leverages Retrieval-Augmented Generation (RAG) by extracting the edited text's emotional features, retrieving speech samples with matching emotions, and synthesizing speech that aligns with the desired emotion while preserving the speaker's identity and quality. To support the training and evaluation of emotional consistency modeling in TSE, we pioneer the benchmarking Emotion Correction Dataset for TSE (ECD-TSE). The prominent aspect of ECD-TSE is its inclusion of $<$text, speech$>$ paired data featuring diverse text variations and a range of emotional expressions. Subjective and objective experiments and comprehensive analysis on ECD-TSE confirm that EmoCorrector significantly enhances the expression of intended emotion while addressing emotion inconsistency limitations in current TSE methods. Code and audio examples are available at https://github.com/AI-S2-Lab/EmoCorrector. | [
{
"version": "v1",
"created": "Sat, 24 May 2025 16:10:56 GMT"
}
] | 2025-05-28T00:00:00 | [
[
"Liu",
"Rui",
""
],
[
"Gao",
"Pu",
""
],
[
"Xi",
"Jiatian",
""
],
[
"Sisman",
"Berrak",
""
],
[
"Busso",
"Carlos",
""
],
[
"Li",
"Haizhou",
""
]
] |
2505.20342 | Paul de Font-Reaulx | Paul de Font-Reaulx | Machine Theory of Mind and the Structure of Human Values | This paper was originally submitted and accepted to the 2023 NeurIPS MP2 Workshop | null | null | null | cs.AI | http://arxiv.org/licenses/nonexclusive-distrib/1.0/ | Value learning is a crucial aspect of safe and ethical AI. This is primarily pursued by methods inferring human values from behaviour. However, humans care about much more than we are able to demonstrate through our actions. Consequently, an AI must predict the rest of our seemingly complex values from a limited sample. I call this the value generalization problem. In this paper, I argue that human values have a generative rational structure and that this allows us to solve the value generalization problem. In particular, we can use Bayesian Theory of Mind models to infer human values not only from behaviour, but also from other values. This has been obscured by the widespread use of simple utility functions to represent human values. I conclude that developing generative value-to-value inference is a crucial component of achieving a scalable machine theory of mind. | [
{
"version": "v1",
"created": "Sat, 24 May 2025 19:18:58 GMT"
}
] | 2025-05-28T00:00:00 | [
[
"de Font-Reaulx",
"Paul",
""
]
] |
2505.20343 | Sonal Prabhune | Sonal Prabhune, Balaji Padmanabhan, and Kaushik Dutta | Do LLMs have a Gender (Entropy) Bias? | 18 pages, 4 figures | null | null | null | cs.CL cs.AI | http://creativecommons.org/licenses/by/4.0/ | We investigate the existence and persistence of a specific type of gender bias in some of the popular LLMs and contribute a new benchmark dataset, RealWorldQuestioning (released on HuggingFace ), developed from real-world questions across four key domains in business and health contexts: education, jobs, personal financial management, and general health. We define and study entropy bias, which we define as a discrepancy in the amount of information generated by an LLM in response to real questions users have asked. We tested this using four different LLMs and evaluated the generated responses both qualitatively and quantitatively by using ChatGPT-4o (as "LLM-as-judge"). Our analyses (metric-based comparisons and "LLM-as-judge" evaluation) suggest that there is no significant bias in LLM responses for men and women at a category level. However, at a finer granularity (the individual question level), there are substantial differences in LLM responses for men and women in the majority of cases, which "cancel" each other out often due to some responses being better for males and vice versa. This is still a concern since typical users of these tools often ask a specific question (only) as opposed to several varied ones in each of these common yet important areas of life. We suggest a simple debiasing approach that iteratively merges the responses for the two genders to produce a final result. Our approach demonstrates that a simple, prompt-based debiasing strategy can effectively debias LLM outputs, thus producing responses with higher information content than both gendered variants in 78% of the cases, and consistently achieving a balanced integration in the remaining cases. | [
{
"version": "v1",
"created": "Sat, 24 May 2025 23:06:41 GMT"
}
] | 2025-05-28T00:00:00 | [
[
"Prabhune",
"Sonal",
""
],
[
"Padmanabhan",
"Balaji",
""
],
[
"Dutta",
"Kaushik",
""
]
] |
2505.20345 | Runze Wang | Runze Wang | Strong edge-coloring of graphs with maximum edge weight seven | null | null | null | null | math.CO | http://arxiv.org/licenses/nonexclusive-distrib/1.0/ | A strong edge-coloring of a graph $G$ is an edge-coloring such that any two edges of distance at most two receive distinct colors. The minimum number of colors we need in order to give $G$ a strong edge-coloring is called the strong chromatic index of $G$, denoted by $\chi_s'(G)$. The maximum edge weight of $G$ is defined to be $\max\{d(u)+d(v):\ uv\in E(G)\}$. In this paper, using the discharging method, we prove that if $G$ is a graph with maximum edge weight $7$ and maximum average degree less than $\frac{28}{9}$, then $\chi_s'(G)\le 13$. | [
{
"version": "v1",
"created": "Sun, 25 May 2025 05:18:16 GMT"
}
] | 2025-05-28T00:00:00 | [
[
"Wang",
"Runze",
""
]
] |
2505.20346 | Nuowei Liu | Jiahao Kuang, Nuowei Liu, Changzhi Sun, Tao Ji, Yuanbin Wu | PDFBench: A Benchmark for De novo Protein Design from Function | null | null | null | null | cs.LG cs.AI q-bio.BM | http://creativecommons.org/licenses/by-nc-nd/4.0/ | In recent years, while natural language processing and multimodal learning have seen rapid advancements, the field of de novo protein design has also experienced significant growth. However, most current methods rely on proprietary datasets and evaluation rubrics, making fair comparisons between different approaches challenging. Moreover, these methods often employ evaluation metrics that capture only a subset of the desired properties of designed proteins, lacking a comprehensive assessment framework. To address these, we introduce PDFBench, the first comprehensive benchmark for evaluating de novo protein design from function. PDFBench supports two tasks: description-guided design and keyword-guided design. To ensure fair and multifaceted evaluation, we compile 22 metrics covering sequence plausibility, structural fidelity, and language-protein alignment, along with measures of novelty and diversity. We evaluate five state-of-the-art baselines, revealing their respective strengths and weaknesses across tasks. Finally, we analyze inter-metric correlations, exploring the relationships between four categories of metrics, and offering guidelines for metric selection. PDFBench establishes a unified framework to drive future advances in function-driven de novo protein design. | [
{
"version": "v1",
"created": "Sun, 25 May 2025 05:40:15 GMT"
}
] | 2025-05-28T00:00:00 | [
[
"Kuang",
"Jiahao",
""
],
[
"Liu",
"Nuowei",
""
],
[
"Sun",
"Changzhi",
""
],
[
"Ji",
"Tao",
""
],
[
"Wu",
"Yuanbin",
""
]
] |
2505.20347 | Wenkai Fang | Wenkai Fang, Shunyu Liu, Yang Zhou, Kongcheng Zhang, Tongya Zheng, Kaixuan Chen, Mingli Song, Dacheng Tao | SeRL: Self-Play Reinforcement Learning for Large Language Models with Limited Data | null | null | null | null | cs.CL cs.AI | http://arxiv.org/licenses/nonexclusive-distrib/1.0/ | Recent advances have demonstrated the effectiveness of Reinforcement Learning (RL) in improving the reasoning capabilities of Large Language Models (LLMs). However, existing works inevitably rely on high-quality instructions and verifiable rewards for effective training, both of which are often difficult to obtain in specialized domains. In this paper, we propose Self-play Reinforcement Learning(SeRL) to bootstrap LLM training with limited initial data. Specifically, SeRL comprises two complementary modules: self-instruction and self-rewarding. The former module generates additional instructions based on the available data at each training step, employing robust online filtering strategies to ensure instruction quality, diversity, and difficulty. The latter module introduces a simple yet effective majority-voting mechanism to estimate response rewards for additional instructions, eliminating the need for external annotations. Finally, SeRL performs conventional RL based on the generated data, facilitating iterative self-play learning. Extensive experiments on various reasoning benchmarks and across different LLM backbones demonstrate that the proposed SeRL yields results superior to its counterparts and achieves performance on par with those obtained by high-quality data with verifiable rewards. Our code is available at https://github.com/wantbook-book/SeRL. | [
{
"version": "v1",
"created": "Sun, 25 May 2025 13:28:04 GMT"
}
] | 2025-05-28T00:00:00 | [
[
"Fang",
"Wenkai",
""
],
[
"Liu",
"Shunyu",
""
],
[
"Zhou",
"Yang",
""
],
[
"Zhang",
"Kongcheng",
""
],
[
"Zheng",
"Tongya",
""
],
[
"Chen",
"Kaixuan",
""
],
[
"Song",
"Mingli",
""
],
[
"Tao",
"Dacheng",
""
]
] |
2505.20348 | Steven Samuels | Steven Samuels, William Campbell, Michael E. Tobar, and Maxim Goryachev | Cryogenic Microwave Whispering Gallery Mode Spectroscopy of Paramagnetic Impurities in High-Purity Crystalline LiF | null | null | null | null | quant-ph cond-mat.mtrl-sci | http://creativecommons.org/licenses/by/4.0/ | A low-noise cryogenic microwave spectroscopy experiment was performed on a high-purity lithium fluoride (LiF) crystal. The spectroscopy data revealed avoided level crossing interactions in whispering gallery modes, indicative of electron spin resonance (ESR) coupling with paramagnetic impurities. Analysis of the interaction spectra identified distinct spin systems corresponding to ${S = 3/2, I = 7/2}$, ${S = 1, I = 7/2}$, and ${S = 3/2, I = 0}$. The number of hyperfine splittings observed, together with the natural abundance of ions possessing the appropriate nuclear spin values, suggest that V$^{2+}$ and V$^{3+}$ impurities, exhibiting orthorhombic distortion, are the most likely sources of the narrow interaction features. This interpretation is supported by earlier ESR studies and established manufacturing records for LiF crystal growth. Additionally, a separate set of broader interaction points is consistent with an orthorhombic model involving a ${S = 3/2, I = 0}$ spin system, although the specific impurity responsible for this interaction remains unidentified. | [
{
"version": "v1",
"created": "Sun, 25 May 2025 14:42:02 GMT"
}
] | 2025-05-28T00:00:00 | [
[
"Samuels",
"Steven",
""
],
[
"Campbell",
"William",
""
],
[
"Tobar",
"Michael E.",
""
],
[
"Goryachev",
"Maxim",
""
]
] |
2505.20349 | Haixin Wang | Haixin Wang, Ruoyan Li, Fred Xu, Fang Sun, Kaiqiao Han, Zijie Huang, Guancheng Wan, Ching Chang, Xiao Luo, Wei Wang, Yizhou Sun | FD-Bench: A Modular and Fair Benchmark for Data-driven Fluid Simulation | 31 pages, 18 figures, paper under review | null | null | null | physics.flu-dyn cs.LG | http://arxiv.org/licenses/nonexclusive-distrib/1.0/ | Data-driven modeling of fluid dynamics has advanced rapidly with neural PDE solvers, yet a fair and strong benchmark remains fragmented due to the absence of unified PDE datasets and standardized evaluation protocols. Although architectural innovations are abundant, fair assessment is further impeded by the lack of clear disentanglement between spatial, temporal and loss modules. In this paper, we introduce FD-Bench, the first fair, modular, comprehensive and reproducible benchmark for data-driven fluid simulation. FD-Bench systematically evaluates 85 baseline models across 10 representative flow scenarios under a unified experimental setup. It provides four key contributions: (1) a modular design enabling fair comparisons across spatial, temporal, and loss function modules; (2) the first systematic framework for direct comparison with traditional numerical solvers; (3) fine-grained generalization analysis across resolutions, initial conditions, and temporal windows; and (4) a user-friendly, extensible codebase to support future research. Through rigorous empirical studies, FD-Bench establishes the most comprehensive leaderboard to date, resolving long-standing issues in reproducibility and comparability, and laying a foundation for robust evaluation of future data-driven fluid models. The code is open-sourced at https://anonymous.4open.science/r/FD-Bench-15BC. | [
{
"version": "v1",
"created": "Sun, 25 May 2025 23:24:18 GMT"
}
] | 2025-05-28T00:00:00 | [
[
"Wang",
"Haixin",
""
],
[
"Li",
"Ruoyan",
""
],
[
"Xu",
"Fred",
""
],
[
"Sun",
"Fang",
""
],
[
"Han",
"Kaiqiao",
""
],
[
"Huang",
"Zijie",
""
],
[
"Wan",
"Guancheng",
""
],
[
"Chang",
"Ching",
""
],
[
"Luo",
"Xiao",
""
],
[
"Wang",
"Wei",
""
],
[
"Sun",
"Yizhou",
""
]
] |
2505.20350 | Jifeng Hu | Jifeng Hu, Sili Huang, Siyuan Guo, Zhaogeng Liu, Li Shen, Lichao Sun, Hechang Chen, Yi Chang, Dacheng Tao | Decision Flow Policy Optimization | null | null | null | null | cs.LG cs.AI | http://arxiv.org/licenses/nonexclusive-distrib/1.0/ | In recent years, generative models have shown remarkable capabilities across diverse fields, including images, videos, language, and decision-making. By applying powerful generative models such as flow-based models to reinforcement learning, we can effectively model complex multi-modal action distributions and achieve superior robotic control in continuous action spaces, surpassing the limitations of single-modal action distributions with traditional Gaussian-based policies. Previous methods usually adopt the generative models as behavior models to fit state-conditioned action distributions from datasets, with policy optimization conducted separately through additional policies using value-based sample weighting or gradient-based updates. However, this separation prevents the simultaneous optimization of multi-modal distribution fitting and policy improvement, ultimately hindering the training of models and degrading the performance. To address this issue, we propose Decision Flow, a unified framework that integrates multi-modal action distribution modeling and policy optimization. Specifically, our method formulates the action generation procedure of flow-based models as a flow decision-making process, where each action generation step corresponds to one flow decision. Consequently, our method seamlessly optimizes the flow policy while capturing multi-modal action distributions. We provide rigorous proofs of Decision Flow and validate the effectiveness through extensive experiments across dozens of offline RL environments. Compared with established offline RL baselines, the results demonstrate that our method achieves or matches the SOTA performance. | [
{
"version": "v1",
"created": "Mon, 26 May 2025 03:42:20 GMT"
}
] | 2025-05-28T00:00:00 | [
[
"Hu",
"Jifeng",
""
],
[
"Huang",
"Sili",
""
],
[
"Guo",
"Siyuan",
""
],
[
"Liu",
"Zhaogeng",
""
],
[
"Shen",
"Li",
""
],
[
"Sun",
"Lichao",
""
],
[
"Chen",
"Hechang",
""
],
[
"Chang",
"Yi",
""
],
[
"Tao",
"Dacheng",
""
]
] |
2505.20351 | Tomer Shoham | Tomer Shoham and Katrina Ligettt | Differentially private ratio statistics | 32 pages, 3 figures, under review | null | null | null | stat.ML cs.LG | http://creativecommons.org/licenses/by/4.0/ | Ratio statistics--such as relative risk and odds ratios--play a central role in hypothesis testing, model evaluation, and decision-making across many areas of machine learning, including causal inference and fairness analysis. However, despite privacy concerns surrounding many datasets and despite increasing adoption of differential privacy, differentially private ratio statistics have largely been neglected by the literature and have only recently received an initial treatment by Lin et al. [1]. This paper attempts to fill this lacuna, giving results that can guide practice in evaluating ratios when the results must be protected by differential privacy. In particular, we show that even a simple algorithm can provide excellent properties concerning privacy, sample accuracy, and bias, not just asymptotically but also at quite small sample sizes. Additionally, we analyze a differentially private estimator for relative risk, prove its consistency, and develop a method for constructing valid confidence intervals. Our approach bridges a gap in the differential privacy literature and provides a practical solution for ratio estimation in private machine learning pipelines. | [
{
"version": "v1",
"created": "Mon, 26 May 2025 04:28:27 GMT"
}
] | 2025-05-28T00:00:00 | [
[
"Shoham",
"Tomer",
""
],
[
"Ligettt",
"Katrina",
""
]
] |
2505.20352 | Jie Wu | Jie Wu | Some results on the k-strong parity property in a graph | 11 pages | null | null | null | math.CO | http://arxiv.org/licenses/nonexclusive-distrib/1.0/ | A graph $G$ has the $k$-strong parity property if for any $X\subseteq V(G)$ with $|X|$ even, $G$ contains a spanning subgraph $F$ with $d_F(u)\equiv1$ (mod 2) for each $u\in X$ and $d_F(v)\in\{k,k+2,k+4,\ldots\}$ for each $v\in V(G)\setminus X$, where $k\geq2$ is an even integer. Kano and Matsumura proposed a characterization for a graph with the $k$-strong parity property (M. Kano, H. Matsumura, Odd-even factors of graphs, Graphs Combin. 41 (2025) 55). In this paper, we first give a size condition for a graph to have the $k$-strong parity property. Then we establish a signless Laplacian spectral radius condition to guarantee that a graph has the $k$-strong parity property. | [
{
"version": "v1",
"created": "Mon, 26 May 2025 05:22:27 GMT"
}
] | 2025-05-28T00:00:00 | [
[
"Wu",
"Jie",
""
]
] |
2505.20353 | Dong Liu | Dong Liu, Jiayi Zhang, Yifan Li, Yanxuan Yu, Ben Lengerich, Ying Nian Wu | FastCache: Fast Caching for Diffusion Transformer Through Learnable Linear Approximation | null | null | null | null | cs.LG cs.AI cs.CV cs.MM cs.PF | http://creativecommons.org/licenses/by/4.0/ | Diffusion Transformers (DiT) are powerful generative models but remain computationally intensive due to their iterative structure and deep transformer stacks. To alleviate this inefficiency, we propose FastCache, a hidden-state-level caching and compression framework that accelerates DiT inference by exploiting redundancy within the model's internal representations. FastCache introduces a dual strategy: (1) a spatial-aware token selection mechanism that adaptively filters redundant tokens based on hidden state saliency, and (2) a transformer-level cache that reuses latent activations across timesteps when changes are statistically insignificant. These modules work jointly to reduce unnecessary computation while preserving generation fidelity through learnable linear approximation. Theoretical analysis shows that FastCache maintains bounded approximation error under a hypothesis-testing-based decision rule. Empirical evaluations across multiple DiT variants demonstrate substantial reductions in latency and memory usage, with best generation output quality compared to other cache methods, as measured by FID and t-FID. Code implementation of FastCache is available on GitHub at https://github.com/NoakLiu/FastCache-xDiT. | [
{
"version": "v1",
"created": "Mon, 26 May 2025 05:58:49 GMT"
}
] | 2025-05-28T00:00:00 | [
[
"Liu",
"Dong",
""
],
[
"Zhang",
"Jiayi",
""
],
[
"Li",
"Yifan",
""
],
[
"Yu",
"Yanxuan",
""
],
[
"Lengerich",
"Ben",
""
],
[
"Wu",
"Ying Nian",
""
]
] |
2505.20354 | Juntong Wu | Juntong Wu, Zijing Liu, He Cao, Hao Li, Bin Feng, Zishan Shu, Ke Yu, Li Yuan, Yu Li | Rethinking Text-based Protein Understanding: Retrieval or LLM? | null | null | null | null | cs.CL cs.AI | http://creativecommons.org/licenses/by-nc-nd/4.0/ | In recent years, protein-text models have gained significant attention for their potential in protein generation and understanding. Current approaches focus on integrating protein-related knowledge into large language models through continued pretraining and multi-modal alignment, enabling simultaneous comprehension of textual descriptions and protein sequences. Through a thorough analysis of existing model architectures and text-based protein understanding benchmarks, we identify significant data leakage issues present in current benchmarks. Moreover, conventional metrics derived from natural language processing fail to accurately assess the model's performance in this domain. To address these limitations, we reorganize existing datasets and introduce a novel evaluation framework based on biological entities. Motivated by our observation, we propose a retrieval-enhanced method, which significantly outperforms fine-tuned LLMs for protein-to-text generation and shows accuracy and efficiency in training-free scenarios. Our code and data can be seen at https://github.com/IDEA-XL/RAPM. | [
{
"version": "v1",
"created": "Mon, 26 May 2025 06:25:43 GMT"
}
] | 2025-05-28T00:00:00 | [
[
"Wu",
"Juntong",
""
],
[
"Liu",
"Zijing",
""
],
[
"Cao",
"He",
""
],
[
"Li",
"Hao",
""
],
[
"Feng",
"Bin",
""
],
[
"Shu",
"Zishan",
""
],
[
"Yu",
"Ke",
""
],
[
"Yuan",
"Li",
""
],
[
"Li",
"Yu",
""
]
] |
2505.20355 | Yeonjoon Jung | Yeonjoon Jung, Daehyun Ahn, Hyungjun Kim, Taesu Kim, and Eunhyeok Park | GraLoRA: Granular Low-Rank Adaptation for Parameter-Efficient Fine-Tuning | null | null | null | null | cs.LG cs.AI | http://creativecommons.org/licenses/by/4.0/ | Low-Rank Adaptation (LoRA) is a popular method for parameter-efficient fine-tuning (PEFT) of generative models, valued for its simplicity and effectiveness. Despite recent enhancements, LoRA still suffers from a fundamental limitation: overfitting when the bottleneck is widened. It performs best at ranks 32-64, yet its accuracy stagnates or declines at higher ranks, still falling short of full fine-tuning (FFT) performance. We identify the root cause as LoRA's structural bottleneck, which introduces gradient entanglement to the unrelated input channels and distorts gradient propagation. To address this, we introduce a novel structure, Granular Low-Rank Adaptation (GraLoRA) that partitions weight matrices into sub-blocks, each with its own low-rank adapter. With negligible computational or storage cost, GraLoRA overcomes LoRA's limitations, effectively increases the representational capacity, and more closely approximates FFT behavior. Experiments on code generation and commonsense reasoning benchmarks show that GraLoRA consistently outperforms LoRA and other baselines, achieving up to +8.5% absolute gain in Pass@1 on HumanEval+. These improvements hold across model sizes and rank settings, making GraLoRA a scalable and robust solution for PEFT. Code, data, and scripts are available at https://github.com/SqueezeBits/GraLoRA.git | [
{
"version": "v1",
"created": "Mon, 26 May 2025 06:48:20 GMT"
}
] | 2025-05-28T00:00:00 | [
[
"Jung",
"Yeonjoon",
""
],
[
"Ahn",
"Daehyun",
""
],
[
"Kim",
"Hyungjun",
""
],
[
"Kim",
"Taesu",
""
],
[
"Park",
"Eunhyeok",
""
]
] |
2505.20356 | Shuoming Zhang | Shuoming Zhang, Jiacheng Zhao, Chunwei Xia, Zheng Wang, Yunji Chen, Xiaobing Feng, Huimin Cui | LEGO-Compiler: Enhancing Neural Compilation Through Translation Composability | 30 pages, 8 figures, 4 tables. Preprint. Under review | null | null | null | cs.PL cs.AI cs.SE | http://creativecommons.org/licenses/by/4.0/ | Large language models (LLMs) have the potential to revolutionize how we design and implement compilers and code translation tools. However, existing LLMs struggle to handle long and complex programs. We introduce LEGO-Compiler, a novel neural compilation system that leverages LLMs to translate high-level languages into assembly code. Our approach centers on three key innovations: LEGO translation, which decomposes the input program into manageable blocks; breaking down the complex compilation process into smaller, simpler verifiable steps by organizing it as a verifiable LLM workflow by external tests; and a feedback mechanism for self-correction. Supported by formal proofs of translation composability, LEGO-Compiler demonstrates high accuracy on multiple datasets, including over 99% on ExeBench and 97.9% on industrial-grade AnsiBench. Additionally, LEGO-Compiler has also acheived near one order-of-magnitude improvement on compilable code size scalability. This work opens new avenues for applying LLMs to system-level tasks, complementing traditional compiler technologies. | [
{
"version": "v1",
"created": "Mon, 26 May 2025 07:07:54 GMT"
}
] | 2025-05-28T00:00:00 | [
[
"Zhang",
"Shuoming",
""
],
[
"Zhao",
"Jiacheng",
""
],
[
"Xia",
"Chunwei",
""
],
[
"Wang",
"Zheng",
""
],
[
"Chen",
"Yunji",
""
],
[
"Feng",
"Xiaobing",
""
],
[
"Cui",
"Huimin",
""
]
] |
2505.20357 | Jun Tian | Jun Tian, He Wang, Jibo He, Yu Pan, Shuo Cao, and Qingquan Jiang | Learning and Interpreting Gravitational-Wave Features from CNNs with a Random Forest Approach | null | null | null | null | cs.LG gr-qc physics.data-an | http://creativecommons.org/licenses/by/4.0/ | Convolutional neural networks (CNNs) have become widely adopted in gravitational wave (GW) detection pipelines due to their ability to automatically learn hierarchical features from raw strain data. However, the physical meaning of these learned features remains underexplored, limiting the interpretability of such models. In this work, we propose a hybrid architecture that combines a CNN-based feature extractor with a random forest (RF) classifier to improve both detection performance and interpretability. Unlike prior approaches that directly connect classifiers to CNN outputs, our method introduces four physically interpretable metrics - variance, signal-to-noise ratio (SNR), waveform overlap, and peak amplitude - computed from the final convolutional layer. These are jointly used with the CNN output in the RF classifier to enable more informed decision boundaries. Tested on long-duration strain datasets, our hybrid model outperforms a baseline CNN model, achieving a relative improvement of 21\% in sensitivity at a fixed false alarm rate of 10 events per month. Notably, it also shows improved detection of low-SNR signals (SNR $\le$ 10), which are especially vulnerable to misclassification in noisy environments. Feature attribution via the RF model reveals that both CNN-extracted and handcrafted features contribute significantly to classification decisions, with learned variance and CNN outputs ranked among the most informative. These findings suggest that physically motivated post-processing of CNN feature maps can serve as a valuable tool for interpretable and efficient GW detection, bridging the gap between deep learning and domain knowledge. | [
{
"version": "v1",
"created": "Mon, 26 May 2025 07:33:27 GMT"
}
] | 2025-05-28T00:00:00 | [
[
"Tian",
"Jun",
""
],
[
"Wang",
"He",
""
],
[
"He",
"Jibo",
""
],
[
"Pan",
"Yu",
""
],
[
"Cao",
"Shuo",
""
],
[
"Jiang",
"Qingquan",
""
]
] |
2505.20360 | Samuele Marco Silveravalle | Alfio Bonanno, Samuele Silveravalle | Spontaneous ghostification: how a dying black hole comes back as a naked singularity | 2 pages. Contribution to the 2025 Gravitation session of the 59th Rencontres de Moriond | null | null | null | gr-qc hep-th | http://arxiv.org/licenses/nonexclusive-distrib/1.0/ | A quantum ghost that destabilizes the Schwarzschild solution, transforming it into a naked singularity, may seem like a physicist's worst nightmare. However, we argue that this scenario represents the natural evolution of a black hole within a conservative high-energy gravity framework and may, in fact, be a desirable outcome. Quadratic curvature terms typically appear as corrections to the Einstein-Hilbert action at high energies; nonetheless, such theories are generally considered incomplete due to the presence of ghost particles at the quantum level, which can spoil vacuum stability. We argue that this instability can only be triggered at the final stages of black hole evaporation, starting a phase transition-like process that alters the nature of the spacetime, similarly to spontaneous scalarization. We propose that the endpoint is a stable, exotic naked singularity, possible only in modified gravity theories, and avoids some of the pathological features associated with standard naked singularities. | [
{
"version": "v1",
"created": "Mon, 26 May 2025 08:31:40 GMT"
}
] | 2025-05-28T00:00:00 | [
[
"Bonanno",
"Alfio",
""
],
[
"Silveravalle",
"Samuele",
""
]
] |
2505.20361 | Chuanxing Wang | Chuanxing Wang, Hui Luo, Kai Wang, Guohuai Zhu, Mingxing Luo | Solving Euler equations with Multiple Discontinuities via Separation-Transfer Physics-Informed Neural Networks | null | null | null | null | physics.flu-dyn cs.LG | http://arxiv.org/licenses/nonexclusive-distrib/1.0/ | Despite the remarkable progress of physics-informed neural networks (PINNs) in scientific computing, they continue to face challenges when solving hydrodynamic problems with multiple discontinuities. In this work, we propose Separation-Transfer Physics Informed Neural Networks (ST-PINNs) to address such problems. By sequentially resolving discontinuities from strong to weak and leveraging transfer learning during training, ST-PINNs significantly reduce the problem complexity and enhance solution accuracy. To the best of our knowledge, this is the first study to apply a PINNs-based approach to the two-dimensional unsteady planar shock refraction problem, offering new insights into the application of PINNs to complex shock-interface interactions. Numerical experiments demonstrate that ST-PINNs more accurately capture sharp discontinuities and substantially reduce solution errors in hydrodynamic problems involving multiple discontinuities. | [
{
"version": "v1",
"created": "Mon, 26 May 2025 08:55:04 GMT"
}
] | 2025-05-28T00:00:00 | [
[
"Wang",
"Chuanxing",
""
],
[
"Luo",
"Hui",
""
],
[
"Wang",
"Kai",
""
],
[
"Zhu",
"Guohuai",
""
],
[
"Luo",
"Mingxing",
""
]
] |
2505.20362 | Jiahui Geng | Jiahui Geng, Qing Li, Zongxiong Chen, Yuxia Wang, Derui Zhu, Zhuohan Xie, Chenyang Lyu, Xiuying Chen, Preslav Nakov, Fakhri Karray | VSCBench: Bridging the Gap in Vision-Language Model Safety Calibration | null | null | null | null | cs.IR cs.AI | http://creativecommons.org/licenses/by/4.0/ | The rapid advancement of vision-language models (VLMs) has brought a lot of attention to their safety alignment. However, existing methods have primarily focused on model undersafety, where the model responds to hazardous queries, while neglecting oversafety, where the model refuses to answer safe queries. In this paper, we introduce the concept of $\textit{safety calibration}$, which systematically addresses both undersafety and oversafety. Specifically, we present $\textbf{VSCBench}$, a novel dataset of 3,600 image-text pairs that are visually or textually similar but differ in terms of safety, which is designed to evaluate safety calibration across image-centric and text-centric scenarios. Based on our benchmark, we evaluate safety calibration across eleven widely used VLMs. Our extensive experiments revealed major issues with both undersafety and oversafety. We further investigated four approaches to improve the model's safety calibration. We found that even though some methods effectively calibrated the models' safety problems, these methods also lead to the degradation of models' utility. This trade-off underscores the urgent need for advanced calibration methods, and our benchmark provides a valuable tool for evaluating future approaches. Our code and data are available at https://github.com/jiahuigeng/VSCBench.git. | [
{
"version": "v1",
"created": "Mon, 26 May 2025 09:01:46 GMT"
}
] | 2025-05-28T00:00:00 | [
[
"Geng",
"Jiahui",
""
],
[
"Li",
"Qing",
""
],
[
"Chen",
"Zongxiong",
""
],
[
"Wang",
"Yuxia",
""
],
[
"Zhu",
"Derui",
""
],
[
"Xie",
"Zhuohan",
""
],
[
"Lyu",
"Chenyang",
""
],
[
"Chen",
"Xiuying",
""
],
[
"Nakov",
"Preslav",
""
],
[
"Karray",
"Fakhri",
""
]
] |
2505.20363 | Jean-Marie MALHERBE | Jean-Marie Malherbe (LIRA, PSL) | The Paris Meudon ground based support to the NASA Solar Maximum Mission in the eighties | null | null | null | null | astro-ph.IM astro-ph.SR | http://arxiv.org/licenses/nonexclusive-distrib/1.0/ | The Solar Maximum Mission of NASA was one of the first satellites with on board digitization of observations. It was launched for the solar maximum of cycle 21 (1980) in order to study the solar activity. It carried many instruments, such as coronagraphs, X and $\gamma$ ray detectors, an Ultra Violet spectrometer and a radiometer. Ground based support was offered by many institutes, such as Paris Meudon observatory under the form of systematic observations or coordinated campaigns with specific instruments. We present here the Meudon Solar Tower (MST) and magnetograph which offered in the eighties a major contribution with observations of velocity and magnetic fields of the photosphere and chromosphere, while SMM was observing the transition region and corona above. | [
{
"version": "v1",
"created": "Mon, 26 May 2025 09:46:10 GMT"
}
] | 2025-05-28T00:00:00 | [
[
"Malherbe",
"Jean-Marie",
"",
"LIRA, PSL"
]
] |
2505.20364 | Nikolaos Mavromatos | Nick E. Mavromatos, Andreas Mershin, Dimitri V. Nanopoulos | On the Potential of Microtubules for Scalable Quantum Computation | 24 pages revtex, 8 pdf figures incorporated | KCL-PH-TH/2025-18 | null | null | physics.bio-ph quant-ph | http://arxiv.org/licenses/nonexclusive-distrib/1.0/ | We examine the quantum coherence properties of tubulin heterodimers in the Microtubule (MT) lattice. In the cavity-MT model proposed by the authors, according to which the MT interiors are modeled as high-Q quantum-electrodynamics cavities, decoherence-resistant entangled states have been argued to emerge under physiological conditions, with decoherence times of order $\mathcal{O}(10^{-6})$ s. The latter is the result of strong electric-dipole interactions of tubulin dimers with ordered-water dipole quanta in the MT interior. We re-interpret the classical nonlinear (pseudospin) $\sigma$-models, describing the emergent dynamics of solitonic excitations in such systems, as representing quantum coherent (or possibly pointer) states, arising from the incomplete collapse of quantum-coherent dipole states. These solitons mediate dissipation-free energy transfer across the MT networks. We underpin logic-gate-like behavior through MT-associated proteins and detail how these structures may support scalable, ambient-temperature quantum computation, with the fundamental unit of information storage being a quDit associated with the basic unit of the MT honeycomb lattice. We describe in detail the decision-making process, after the action of an external stimulus, during which optimal path selection for energy-loss-free signal and information transport across the MT network emerges. Finally, we propose experimental pathways, including Rabi-splitting spectroscopy and entangled surface plasmon probes, to experimentally validate our predictions for MT-based, scalable quantum computation. | [
{
"version": "v1",
"created": "Mon, 26 May 2025 09:47:33 GMT"
}
] | 2025-05-28T00:00:00 | [
[
"Mavromatos",
"Nick E.",
""
],
[
"Mershin",
"Andreas",
""
],
[
"Nanopoulos",
"Dimitri V.",
""
]
] |
2505.20365 | Syed Rehan Shah | Areesha Rehman, Zikria Saleem, Jarrar Amjad, Syed Rehan Shah, Kamran Siddique | A Comparison of Bacterial Colonies Count from Petri Dishes Utilizing Hough Transform and Traditional Manual Counting | 13 pages | null | null | null | q-bio.QM | http://creativecommons.org/licenses/by/4.0/ | Bacterial colony enumeration is an essential stage in microbiological research, allowing susceptibility to antibiotics assessment, monitoring of the environment, and clinical diagnostics. Traditional manual counting methods are costly and susceptible to human mistakes, prompting the creation of automated detection systems. This research compares the efficacy of the Hough Circle Transform method for automated colony detection to hand counting of E. coli, S. aureus, and P. aeruginosa colonies on 200 petri plates. These bacteria are among the most clinically relevant pathogens, with E. coli frequently causing urinary tract infections, S. aureus connected with skin and bloodstream infections, and P. aeruginosa a significant issue in hospital-acquired infections. When colonies were counted automatically without visual correction, the mean difference from manual counts was 59.7%, with overestimation and underestimation occurring in 29% and 45% of cases, whereas S. aureus and P. aeruginosa had higher error rates. The proposed methodology achieved an overall accuracy of 95% for E. coli, 90% for S. aureus, and 84% for P. aeruginosa, with associated recall values of 95%, 91%, and 86%. The F-measure remained continuously high, ranging between 0.85 and 0.95. Regarding efficiency, manual counting required an average of 70 seconds per plate, while automated counting without and with visual correction took 30 seconds. Despite issues with segmentation in high-density plates, automated approaches offer a potential approach to high-throughput bacterial enumeration by decreasing labor-intensive operations while retaining adequate accuracy. Future research should enhance colony algorithmic segmentation and picture preprocessing approaches to improve detection performance, especially on crowded petri plates. | [
{
"version": "v1",
"created": "Mon, 26 May 2025 10:13:11 GMT"
}
] | 2025-05-28T00:00:00 | [
[
"Rehman",
"Areesha",
""
],
[
"Saleem",
"Zikria",
""
],
[
"Amjad",
"Jarrar",
""
],
[
"Shah",
"Syed Rehan",
""
],
[
"Siddique",
"Kamran",
""
]
] |
2505.20366 | Jan Janssen | Jan Janssen, Janine George, Julian Geiger, Marnik Bercx, Xing Wang, Christina Ertural, Joerg Schaarschmidt, Alex M. Ganose, Giovanni Pizzi, Tilmann Hickel, Joerg Neugebauer | A Python workflow definition for computational materials design | null | null | null | null | cs.SE cond-mat.mtrl-sci | http://creativecommons.org/licenses/by/4.0/ | Numerous Workflow Management Systems (WfMS) have been developed in the field of computational materials science with different workflow formats, hindering interoperability and reproducibility of workflows in the field. To address this challenge, we introduce here the Python Workflow Definition (PWD) as a workflow exchange format to share workflows between Python-based WfMS, currently AiiDA, jobflow, and pyiron. This development is motivated by the similarity of these three Python-based WfMS, that represent the different workflow steps and data transferred between them as nodes and edges in a graph. With the PWD, we aim at fostering the interoperability and reproducibility between the different WfMS in the context of Findable, Accessible, Interoperable, Reusable (FAIR) workflows. To separate the scientific from the technical complexity, the PWD consists of three components: (1) a conda environment that specifies the software dependencies, (2) a Python module that contains the Python functions represented as nodes in the workflow graph, and (3) a workflow graph stored in the JavaScript Object Notation (JSON). The first version of the PWD supports directed acyclic graph (DAG)-based workflows. Thus, any DAG-based workflow defined in one of the three WfMS can be exported to the PWD and afterwards imported from the PWD to one of the other WfMS. After the import, the input parameters of the workflow can be adjusted and computing resources can be assigned to the workflow, before it is executed with the selected WfMS. This import from and export to the PWD is enabled by the PWD Python library that implements the PWD in AiiDA, jobflow, and pyiron. | [
{
"version": "v1",
"created": "Mon, 26 May 2025 11:05:50 GMT"
}
] | 2025-05-28T00:00:00 | [
[
"Janssen",
"Jan",
""
],
[
"George",
"Janine",
""
],
[
"Geiger",
"Julian",
""
],
[
"Bercx",
"Marnik",
""
],
[
"Wang",
"Xing",
""
],
[
"Ertural",
"Christina",
""
],
[
"Schaarschmidt",
"Joerg",
""
],
[
"Ganose",
"Alex M.",
""
],
[
"Pizzi",
"Giovanni",
""
],
[
"Hickel",
"Tilmann",
""
],
[
"Neugebauer",
"Joerg",
""
]
] |
2505.20367 | Sen Yan | Sen Yan, Fabrizio Gabellieri, Etienne Goffinet, Filippo Castiglione, Thomas Launey | DiffNMR: Advancing Inpainting of Randomly Sampled Nuclear Magnetic Resonance Signals | null | null | null | null | eess.IV cs.LG | http://creativecommons.org/licenses/by/4.0/ | Nuclear Magnetic Resonance (NMR) spectroscopy leverages nuclear magnetization to probe molecules' chemical environment, structure, and dynamics, with applications spanning from pharmaceuticals to the petroleum industry. Despite its utility, the high cost of NMR instrumentation, operation and the lengthy duration of experiments necessitate the development of computational techniques to optimize acquisition times. Non-Uniform sampling (NUS) is widely employed as a sub-sampling method to address these challenges, but it often introduces artifacts and degrades spectral quality, offsetting the benefits of reduced acquisition times. In this work, we propose the use of deep learning techniques to enhance the reconstruction quality of NUS spectra. Specifically, we explore the application of diffusion models, a relatively untapped approach in this domain. Our methodology involves applying diffusion models to both time-time and time-frequency NUS data, yielding satisfactory reconstructions of challenging spectra from the benchmark Artina dataset. This approach demonstrates the potential of diffusion models to improve the efficiency and accuracy of NMR spectroscopy as well as the superiority of using a time-frequency domain data over the time-time one, opening new landscapes for future studies. | [
{
"version": "v1",
"created": "Mon, 26 May 2025 11:06:11 GMT"
}
] | 2025-05-28T00:00:00 | [
[
"Yan",
"Sen",
""
],
[
"Gabellieri",
"Fabrizio",
""
],
[
"Goffinet",
"Etienne",
""
],
[
"Castiglione",
"Filippo",
""
],
[
"Launey",
"Thomas",
""
]
] |
2505.20368 | Jaeyoung Choe | Jaeyoung Choe, Jihoon Kim, Woohwan Jung | Hierarchical Retrieval with Evidence Curation for Open-Domain Financial Question Answering on Standardized Documents | ACL 2025 (Findings) | null | null | null | cs.IR cs.AI cs.CL | http://arxiv.org/licenses/nonexclusive-distrib/1.0/ | Retrieval-augmented generation (RAG) based large language models (LLMs) are widely used in finance for their excellent performance on knowledge-intensive tasks. However, standardized documents (e.g., SEC filing) share similar formats such as repetitive boilerplate texts, and similar table structures. This similarity forces traditional RAG methods to misidentify near-duplicate text, leading to duplicate retrieval that undermines accuracy and completeness. To address these issues, we propose the Hierarchical Retrieval with Evidence Curation (HiREC) framework. Our approach first performs hierarchical retrieval to reduce confusion among similar texts. It first retrieve related documents and then selects the most relevant passages from the documents. The evidence curation process removes irrelevant passages. When necessary, it automatically generates complementary queries to collect missing information. To evaluate our approach, we construct and release a Large-scale Open-domain Financial (LOFin) question answering benchmark that includes 145,897 SEC documents and 1,595 question-answer pairs. Our code and data are available at https://github.com/deep-over/LOFin-bench-HiREC. | [
{
"version": "v1",
"created": "Mon, 26 May 2025 11:08:23 GMT"
}
] | 2025-05-28T00:00:00 | [
[
"Choe",
"Jaeyoung",
""
],
[
"Kim",
"Jihoon",
""
],
[
"Jung",
"Woohwan",
""
]
] |
2505.20369 | Michel Bakni | Michel Bakni (ESTIA), Abbad Diraneyya, Wael Tellat | WikiTermBase: An AI-Augmented Term Base to Standardize Arabic Translation on Wikipedia | null | Wiki Workshop 2025, Wikimedia Foundation, May 2025, Onlince Conference, United States. pp.1-3 | null | null | cs.IR | http://arxiv.org/licenses/nonexclusive-distrib/1.0/ | Term bases are recognized as one of the most effective components of translation software in time saving and consistency. In spite of the many recent advances in natural language processing (NLP) and large language models (LLMs), major translation platforms have yet to take advantage of these tools to improve their term bases and support scalable content for underrepresented languages, which often struggle with localizing technical terminology. Language academies in the Arab World, for example, have struggled since the 1940s to unify the way new scientific terms enter the Arabic language at scale. This abstract introduces an open source tool, WikiTermBase, with a systematic approach for building a lexicographical database with over 900K terms, which were collected and mapped from a multitude of sources on a semantic and morphological basis. The tool was successfully implemented on Arabic Wikipedia to standardize translated English and French terms. | [
{
"version": "v1",
"created": "Mon, 26 May 2025 11:27:01 GMT"
}
] | 2025-05-28T00:00:00 | [
[
"Bakni",
"Michel",
"",
"ESTIA"
],
[
"Diraneyya",
"Abbad",
""
],
[
"Tellat",
"Wael",
""
]
] |
2505.20370 | Martine Dyring Hansen | Martine Dyring Hansen, Elena Celledoni, Benjamin Kwanen Tapley | Learning mechanical systems from real-world data using discrete forced Lagrangian dynamics | null | null | null | null | eess.SY cs.LG cs.SY | http://creativecommons.org/licenses/by/4.0/ | We introduce a data-driven method for learning the equations of motion of mechanical systems directly from position measurements, without requiring access to velocity data. This is particularly relevant in system identification tasks where only positional information is available, such as motion capture, pixel data or low-resolution tracking. Our approach takes advantage of the discrete Lagrange-d'Alembert principle and the forced discrete Euler-Lagrange equations to construct a physically grounded model of the system's dynamics. We decompose the dynamics into conservative and non-conservative components, which are learned separately using feed-forward neural networks. In the absence of external forces, our method reduces to a variational discretization of the action principle naturally preserving the symplectic structure of the underlying Hamiltonian system. We validate our approach on a variety of synthetic and real-world datasets, demonstrating its effectiveness compared to baseline methods. In particular, we apply our model to (1) measured human motion data and (2) latent embeddings obtained via an autoencoder trained on image sequences. We demonstrate that we can faithfully reconstruct and separate both the conservative and forced dynamics, yielding interpretable and physically consistent predictions. | [
{
"version": "v1",
"created": "Mon, 26 May 2025 12:13:00 GMT"
}
] | 2025-05-28T00:00:00 | [
[
"Hansen",
"Martine Dyring",
""
],
[
"Celledoni",
"Elena",
""
],
[
"Tapley",
"Benjamin Kwanen",
""
]
] |
2505.20371 | Jing-Fu Guo | Zhipeng Tong (a), Xiaoqi Sun (a, c), Guiya Qin (b, d), Jinpu Bai (a), Qi Zhao (a), Aimin Ren (b), and Jingfu Guo (a) ((a) College of physics, Northeast Normal Univb Institute of theoretical chemistry, College of Chemistry, Jilin University, Changchun, (b) Institute of theoretical chemistry, College of Chemistry, Jilin University, Changchun, (c) Department of Physics, College of Science, Yanbian University, Yanji, (d) Chemistry Teaching and Research Section, Department of Basic Sciences, Jilin Jianzhu University, Changchun) | Theoretical study on charge transfer properties of triphenylamino-ethynyl Polycyclic Aromatic Hydrocarbon derivatives | in Chinese language | null | null | null | physics.chem-ph physics.comp-ph | http://creativecommons.org/licenses/by-nc-nd/4.0/ | This study systematically investigates the regulation mechanisms of backbone topology (tri-/tetracyclic arenes), substitution positions, and functional groups on charge transport properties through molecular design of triphenylamine-ethynylene fused acene derivatives. By integrating Marcus charge transfer theory with kinetic Monte Carlo simulations, we demonstrate that sulfur-doped tricyclic arene backbones (benzodithiophene and anthracene) effectively suppress high-frequency vibrational modes reducing reorganization energy to 146.1 meV. Concurrent optimization of intermolecular $\pi$-$\pi$ slippage enhances 2D hole mobility. Notably, asymmetric charge transport pathways in 2,7-disubstituted pyrene(27DTEP) decrease transfer integrals by 34%, while 1,6-substitution (16DTEP)reconstructs HOMO orbital distribution and induces rotational stacking, boosting transfer integrals by 28% and improving mobility isotropy. We further propose a "backbone-functional group synergy" strategy, revealing that concentrated orbital localization on the backbone amplifies transfer integral gains, outweighing the 38% increase in reorganization energy and significantly enhancing mobility. These findings establish a theoretical framework and quantitative model for the rational design of high-mobility organic ultraviolet photodetectors. | [
{
"version": "v1",
"created": "Mon, 26 May 2025 13:15:18 GMT"
}
] | 2025-05-28T00:00:00 | [
[
"Tong",
"Zhipeng",
"",
"a"
],
[
"Sun",
"Xiaoqi",
"",
"a, c"
],
[
"Qin",
"Guiya",
"",
"b, d"
],
[
"Bai",
"Jinpu",
"",
"a"
],
[
"Zhao",
"Qi",
"",
"a"
],
[
"Ren",
"Aimin",
"",
"b"
],
[
"Guo",
"Jingfu",
"",
"a"
]
] |
2505.20372 | Anton Ponomarev | Anton Ponomarev, Lutz Gr\"oll | Stability of two-dimensional SISO LTI system with bounded feedback gain that has bounded derivative | Submitted to the European Control Conference (ECC) 2025 | null | null | null | math.OC cs.SY eess.SY | http://creativecommons.org/licenses/by-nc-nd/4.0/ | We consider a two-dimensional SISO LTI system closed by uncertain linear feedback. The feedback gain is time-varying, bounded, and has a bounded derivative (both bounds are known). We investigate the asymptotic stability of this system under all admissible behaviors of the gain. Note that the situation is similar to the classical absolute stability problem of Lurie--Aizerman with two differences: linearity and derivative constraint. Our method of analysis is therefore inspired by the variational ideas of Pyatnitskii, Barabanov, Margaliot, and others developed for the absolute stability problem. We derive the Hamilton--Jacobi--Bellman equation for a function describing the "most unstable" of the possible portraits of the closed-loop system. A numerical method is proposed for solving the equation. Based on the solution, sufficient conditions are formulated for the asymptotic stability and instability. The method is applied to an equation arising from the analysis of a power electronics synchronization circuit. | [
{
"version": "v1",
"created": "Mon, 26 May 2025 13:19:35 GMT"
}
] | 2025-05-28T00:00:00 | [
[
"Ponomarev",
"Anton",
""
],
[
"Gröll",
"Lutz",
""
]
] |
2505.20373 | Andre Pointner | A. Pointner, D. Thalheim, S. Belasi, L. Heinen, C. Bonato, T. Luehmann, J. Meijer, R. Tietze, C. Alexiou, R. Schneider-Stock, R. Nagy | Optimizing SPION Labeling for Single-Cell Magnetic Microscopy | null | null | null | null | physics.bio-ph quant-ph | http://creativecommons.org/licenses/by/4.0/ | This study explores the correlation between iron mass on cell surfaces and the resultant magnetic field. Human colorectal cancer cells (HT29 line) were labeled with varying concentrations of SPIONs and imaged via a NV center widefield magnetic microscope. To assess the labeling efficacy, a convolutional neural network trained on simulated magnetic dipole data was utilized to reconstruct key labeling parameters on a cell-by-cell basis, including cell diameter, sensor proximity, and the iron mass associated with surface-bound SPIONs.
Our analysis provided quantitative metrics for these parameters across a range of labeling concentrations. The findings indicated that increasing SPION concentration enhances both the cell-surface iron mass and magnetic field strength, demonstrating a saturation effect. This methodology offers a coherent framework for the quantitative, high-throughput characterization of magnetically labeled cells, presenting significant implications for the fields of cell biology and magnetic sensing applications. | [
{
"version": "v1",
"created": "Mon, 26 May 2025 13:47:40 GMT"
}
] | 2025-05-28T00:00:00 | [
[
"Pointner",
"A.",
""
],
[
"Thalheim",
"D.",
""
],
[
"Belasi",
"S.",
""
],
[
"Heinen",
"L.",
""
],
[
"Bonato",
"C.",
""
],
[
"Luehmann",
"T.",
""
],
[
"Meijer",
"J.",
""
],
[
"Tietze",
"R.",
""
],
[
"Alexiou",
"C.",
""
],
[
"Schneider-Stock",
"R.",
""
],
[
"Nagy",
"R.",
""
]
] |
2505.20374 | Anton Ponomarev | Anton Ponomarev, Lutz Gr\"oll, Veit Hagenmeyer | Numerical estimation of the lock-in domain of a DC/AC inverter | Submitted to the IFAC Symposium on Nonlinear Control Systems (NOLCOS) 2025 | null | null | null | math.OC cs.SY eess.SY | http://creativecommons.org/licenses/by-nc-nd/4.0/ | We estimate the lock-in domain of the origin of a current control system which is used in common DC/AC inverter designs. The system is a cascade connection of a 4-dimensional linear system (current controller, CC) followed by a two-dimensional nonlinear system (phase-locked loop, PLL). For the PLL, we construct a Lyapunov function via numerical approximation of its level curves. In combination with the quadratic Lyapunov function of the CC, it forms a vector Lyapunov function (VLF) for the overall system. A forward-invariant set of the VLF is found via numerical application of the comparison principle. By LaSalle's invariance principle, convergence to the origin is established. | [
{
"version": "v1",
"created": "Mon, 26 May 2025 13:51:01 GMT"
}
] | 2025-05-28T00:00:00 | [
[
"Ponomarev",
"Anton",
""
],
[
"Gröll",
"Lutz",
""
],
[
"Hagenmeyer",
"Veit",
""
]
] |
2505.20375 | Xiaosong Li | Agam Shayit (1), Can Liao (2), Shiv Upadhyay (2), Hang Hu (2), Tianyuan Zhang (2), Eugene DePrince III (3), Chao Yang (4), Xiaosong Li (2) ((1) Department of Physics, University of Washington, USA, (2) Department of Chemistry, University of Washington, USA, (3) Department of Chemistry and Biochemistry, Florida State University, USA, (4) Applied Mathematics and Computational Research Division, Lawrence Berkeley National Laboratory, USA) | Breaking the Quadrillion Determinant Barrier in Numerically Exact Configuration Interaction | null | null | null | null | physics.chem-ph | http://arxiv.org/licenses/nonexclusive-distrib/1.0/ | The combinatorial scaling of configuration interaction (CI) has long restricted its applicability to only the simplest molecular systems. Here, we report the first numerically exact CI calculation exceeding one quadrillion ($10^{15}$) determinants, enabled by lossless categorical compression within the small-tensor-product distributed active space (STP-DAS) framework. As a demonstration, we converged the relativistic full CI (FCI) ground state of a magnesium atom involving over $10^{15}$ complex-valued 2-spinor determinants in under 8.6 hours (time-to-completion) using 1500 nodes, representing the largest FCI calculation reported to date. Additionally, we achieved $\boldsymbol{\sigma}$-build times of just 5 minutes for systems with approximately 150 billion complex-valued 2-spinor determinants using only a few compute nodes. Extensive benchmarks confirm that the method retains numerical exactness with drastically reduced resource demands. Compared to previous state-of-the-art FCI calculations, this work represents a 3-orders-of-magnitude increase in CI space, a 6-orders-of-magnitude increase in FLOP count, and a 6-orders-of-magnitude improvement in computational speed. By introducing a lossless, categorically compressed representation of the CI expansion vectors and reformulating the $\boldsymbol{\sigma}$-build accordingly, we eliminate memory bottlenecks associated with storing excitation lists and CI vectors while significantly reducing computational cost. A compression-compatible preconditioner further enhances performance by generating compressed CI expansion vectors throughout Davidson iterations. This work establishes a new computational frontier for numerically exact CI methods, enabling chemically and physically accurate simulations of strongly correlated, spin-orbit coupled systems previously thought to be beyond reach. | [
{
"version": "v1",
"created": "Mon, 26 May 2025 14:20:04 GMT"
}
] | 2025-05-28T00:00:00 | [
[
"Shayit",
"Agam",
"",
"Department of Physics, University of Washington, USA"
],
[
"Liao",
"Can",
"",
"Department of Chemistry, University of Washington, USA"
],
[
"Upadhyay",
"Shiv",
"",
"Department of Chemistry, University of Washington, USA"
],
[
"Hu",
"Hang",
"",
"Department of Chemistry, University of Washington, USA"
],
[
"Zhang",
"Tianyuan",
"",
"Department of Chemistry, University of Washington, USA"
],
[
"DePrince",
"Eugene",
"III",
"Department of Chemistry and Biochemistry, Florida State University, USA"
],
[
"Yang",
"Chao",
"",
"Applied Mathematics and Computational Research Division, Lawrence Berkeley National Laboratory, USA"
],
[
"Li",
"Xiaosong",
"",
"Department of Chemistry, University of Washington, USA"
]
] |
2505.20376 | Shengjie Yu | Shengjie Yu, Zhaoxuan Zhu, and Laurent Sanchez-Palencia | Exact solution and Luttinger liquid behavior of the quantum 1D hard rod model | null | null | null | null | cond-mat.str-el cond-mat.quant-gas quant-ph | http://arxiv.org/licenses/nonexclusive-distrib/1.0/ | The quantum hard rod model, a one-dimensional extension of the Lieb-Liniger model, is exactly solved using an adapted Bethe ansatz. Our solution, benchmarked against path-integral quantum Monte Carlo calculations, reveals significant corrections to the excitation spectrum and thermodynamic properties, previously overlooked by the standard excluded-volume approach. We also show that the hard rod model exhibits Luttinger liquid behavior across a wide range of parameters, at zero and finite temperature, as unveiled by correlation functions. This work provides a comprehensive framework for understanding strongly correlated regimes in dilute 1D systems, with applications to quantum wires, spin chains, and ultracold atoms. | [
{
"version": "v1",
"created": "Mon, 26 May 2025 15:09:17 GMT"
}
] | 2025-05-28T00:00:00 | [
[
"Yu",
"Shengjie",
""
],
[
"Zhu",
"Zhaoxuan",
""
],
[
"Sanchez-Palencia",
"Laurent",
""
]
] |
2505.20377 | Alona Zharova | Lennart Ullner, Alona Zharova, Felix Creutzig | Algorithmic Control Improves Residential Building Energy and EV Management when PV Capacity is High but Battery Capacity is Low | null | null | null | null | eess.SY cs.AI cs.LG cs.SY | http://creativecommons.org/licenses/by-sa/4.0/ | Efficient energy management in prosumer households is key to alleviating grid stress in an energy transition marked by electric vehicles (EV), renewable energies and battery storage. However, it is unclear how households optimize prosumer EV charging. Here we study real-world data from 90 households on fixed-rate electricity tariffs in German-speaking countries to investigate the potential of Deep Reinforcement Learning (DRL) and other control approaches (Rule-Based, Model Predictive Control) to manage the dynamic and uncertain environment of Home Energy Management (HEM) and optimize household charging patterns. The DRL agent efficiently aligns charging of EV and battery storage with photovoltaic (PV) surplus. We find that frequent EV charging transactions, early EV connections and PV surplus increase optimization potential. A detailed analysis of nine households (1 hour resolution, 1 year) demonstrates that high battery capacity facilitates self optimization; in this case further algorithmic control shows little value. In cases with relatively low battery capacity, algorithmic control with DRL improves energy management and cost savings by a relevant margin. This result is further corroborated by our simulation of a synthetic household. We conclude that prosumer households with optimization potential would profit from DRL, thus benefiting also the full electricity system and its decarbonization. | [
{
"version": "v1",
"created": "Mon, 26 May 2025 15:19:01 GMT"
}
] | 2025-05-28T00:00:00 | [
[
"Ullner",
"Lennart",
""
],
[
"Zharova",
"Alona",
""
],
[
"Creutzig",
"Felix",
""
]
] |
2505.20378 | Martin Guillemaud | Martin Guillemaud, Vera Dinkelacker, Mario Chavez | Hyperbolic embedding of multilayer networks | 9 pages, 4 figures | null | null | null | cs.SI physics.data-an q-bio.NC | http://creativecommons.org/licenses/by/4.0/ | Multilayer networks offer a powerful framework for modeling complex systems across diverse domains, effectively capturing multiple types of connections and interdependent subsystems commonly found in real world scenarios. To analyze these networks, embedding techniques that project nodes into a lower-dimensional geometric space are essential. This paper introduces a novel hyperbolic embedding framework that advances the state of the art in multilayer network analysis. Our method, which supports heterogeneous node sets across networks and inter-layer connections, generates layer-specific hyperbolic embeddings, enabling detailed intra-layer analysis and inter-layer comparisons, while simultaneously preserving the global multilayer structure within hyperbolic space, a capability that sets it apart from existing approaches, which typically rely on independent embedding of layers. Through experiments on synthetic multilayer stochastic block models, we demonstrate that our approach effectively preserves community structure, even when layers consist of different node sets. When applied to real brain networks, the method successfully clusters disease-related brain regions from different patients, outperforming layer-independent approaches and highlighting its relevance for comparative analysis. Overall, this work provides a robust tool for multilayer network analysis, enhancing interpretability and offering new insights into the structure and function of complex systems. | [
{
"version": "v1",
"created": "Mon, 26 May 2025 15:53:38 GMT"
}
] | 2025-05-28T00:00:00 | [
[
"Guillemaud",
"Martin",
""
],
[
"Dinkelacker",
"Vera",
""
],
[
"Chavez",
"Mario",
""
]
] |
2505.20379 | Eliran Sherzer Dr. | Eliran Sherzer, Yehezkel Resheff, and Miklos Telek | An Unconstrained Optimization Approach to Moment Fitting with Phase Type Distributions | null | null | null | null | math.OC math.PR | http://creativecommons.org/licenses/by/4.0/ | Phase type (PH) distributions are widely used in modeling and simulation due to their generality and analytical properties. In such settings, it is often necessary to construct a PH distribution that aligns with real-world data by matching a set of prescribed moments. Existing approaches provide either exact closed-form solutions or iterative procedures that may yield exact or approximate results. However, these methods are limited to matching a small number of moments using PH distributions with a small number of phases, or are restricted to narrow subclasses within the PH family. We address the problem of approximately fitting a larger set of given moments using potentially large PH distributions. We introduce an optimization methodology that relies on a re-parametrization of the Markovian representation, formulated in a space that enables unconstrained optimization of the moment-matching objective. This reformulation allows us to scale to significantly larger PH distributions and capture higher moments. Results on a large and diverse set of moment targets show that the proposed method is, in the vast majority of cases, capable of fitting as many as 20 moments to PH distributions with as many as 100 phases, with small relative errors on the order of under 0.5% from each target. We further demonstrate an application of the optimization framework where we search for a PH distribution that conforms not only to a given set of moments but also to a given shape. Finally, we illustrate the practical utility of this approach through a queueing application, presenting a case study that examines the influence of the i^{th} moment of the inter-arrival and service time distributions on the steady-state probabilities of the GI/GI/1 queue length. | [
{
"version": "v1",
"created": "Mon, 26 May 2025 17:14:50 GMT"
}
] | 2025-05-28T00:00:00 | [
[
"Sherzer",
"Eliran",
""
],
[
"Resheff",
"Yehezkel",
""
],
[
"Telek",
"Miklos",
""
]
] |
2505.20380 | Simin Fan | Simin Fan, Maria Ios Glarou, Martin Jaggi | GRAPE: Optimize Data Mixture for Group Robust Multi-target Adaptive Pretraining | null | null | null | null | cs.LG | http://arxiv.org/licenses/nonexclusive-distrib/1.0/ | The performance of large language models (LLMs) across diverse downstream applications is fundamentally governed by the quality and composition of their pretraining corpora. Existing domain reweighting algorithms primarily optimize data mixtures for a single target task, thereby resulting in models that overfit to specialized objectives while exhibiting substantial performance degradation on other benchmarks. This paper introduces Group Robust Multi-target Adaptive PrEtraining (GRAPE), a novel multi-source-multi-target domain reweighting framework designed to calibrate pretraining data mixtures for robust performance across multiple target tasks simultaneously. GRAPE dynamically adjusts sampling weights across source domains (domain weights) while concurrently modulating task weights that quantify the relative importance of each individual target task. This adaptive process prioritizes tasks based on their learning difficulty throughout training. We formulate this interleaved reweighting mechanism as a minimax optimization problem: The inner maximization adjusts task weights leveraging group distributed-robust-optimization (DRO), where those tasks demonstrating the least improvement under the current data mixture are prioritized with higher weights; The outer minimization then optimizes domain weights to maximize loss reduction on the prioritized tasks. Experiments on ClimbLab and SlimPajama datasets demonstrate that GRAPE consistently outperforms baseline methods in terms of reasoning performance across 6 benchmarks. Furthermore, when applied to multilingual targets, GRAPE effectively identifies optimal training mixtures from mainstream languages, achieving superior language modeling capabilities across 8 low-resource target languages. | [
{
"version": "v1",
"created": "Mon, 26 May 2025 17:32:14 GMT"
}
] | 2025-05-28T00:00:00 | [
[
"Fan",
"Simin",
""
],
[
"Glarou",
"Maria Ios",
""
],
[
"Jaggi",
"Martin",
""
]
] |
2505.20381 | Sijia Chen | Sijia Chen, Yanqiu Yu, En Yu, Wenbing Tao | ReaMOT: A Benchmark and Framework for Reasoning-based Multi-Object Tracking | 19 pages, 11 figures, 6 tables | null | null | null | cs.CV | http://creativecommons.org/licenses/by/4.0/ | Referring Multi-object tracking (RMOT) is an important research field in computer vision. Its task form is to guide the models to track the objects that conform to the language instruction. However, the RMOT task commonly requires clear language instructions, such methods often fail to work when complex language instructions with reasoning characteristics appear. In this work, we propose a new task, called Reasoning-based Multi-Object Tracking (ReaMOT). ReaMOT is a more challenging task that requires accurate reasoning about objects that match the language instruction with reasoning characteristic and tracking the objects' trajectories. To advance the ReaMOT task and evaluate the reasoning capabilities of tracking models, we construct ReaMOT Challenge, a reasoning-based multi-object tracking benchmark built upon 12 datasets. Specifically, it comprises 1,156 language instructions with reasoning characteristic, 423,359 image-language pairs, and 869 diverse scenes, which is divided into three levels of reasoning difficulty. In addition, we propose a set of evaluation metrics tailored for the ReaMOT task. Furthermore, we propose ReaTrack, a training-free framework for reasoning-based multi-object tracking based on large vision-language models (LVLM) and SAM2, as a baseline for the ReaMOT task. Extensive experiments on the ReaMOT Challenge benchmark demonstrate the effectiveness of our ReaTrack framework. | [
{
"version": "v1",
"created": "Mon, 26 May 2025 17:55:19 GMT"
}
] | 2025-05-28T00:00:00 | [
[
"Chen",
"Sijia",
""
],
[
"Yu",
"Yanqiu",
""
],
[
"Yu",
"En",
""
],
[
"Tao",
"Wenbing",
""
]
] |
2505.20382 | Pierre Guilmin | Pierre Guilmin, Pierre Rouchon and Antoine Tilloy | Time-averaged continuous quantum measurement | 5+2 pages, 3 figures | null | null | null | quant-ph | http://arxiv.org/licenses/nonexclusive-distrib/1.0/ | The theory of continuous quantum measurement allows to reconstruct the state $\rho_t$ of a system from a continuous stochastic measurement record $I_t$. However, this truly continuous-time signal $I_t$ is never available in practice. In experiments, one generally has access to its digitization, i.e., to a series of time averages $I_k$ over finite intervals of duration $\Delta t$. In this letter, we take this digitization seriously and define $\bar{\rho}_n$ as the best Bayesian estimate of the quantum state given (only) a digitized record $(I_1,\dots,I_n)$. We show that $\bar{\rho}_{n+1}$ can be computed recursively from $I_{n+1}$ and $\bar{\rho}_n$ using an exact formula. The latter can be evaluated numerically exactly, or used as the basis for a perturbative expansion into successive powers of $\sqrt{\Delta t}$. This allows reconstructing quantum trajectories in regimes of coarse $\Delta t$ where existing methods fail, estimating parameters at fixed $\Delta t$ without bias, and directly sampling digitized quantum trajectories with schemes of arbitrarily high order. | [
{
"version": "v1",
"created": "Mon, 26 May 2025 18:00:00 GMT"
}
] | 2025-05-28T00:00:00 | [
[
"Guilmin",
"Pierre",
""
],
[
"Rouchon",
"Pierre",
""
],
[
"Tilloy",
"Antoine",
""
]
] |
2505.20383 | Aurelio Amerio | Elena Pinetti, Veronika Vodeb, Aurelio Amerio, Alessandro Cuoco, Stefano Camera, Nicolao Fornengo, and Gabrijela Zaharijas | Across the Universe: Dark Matter and Galaxy Cross-Correlations with the Cherenkov Telescope Array Observatory | 9 pages + Appendix, 28 Figures | null | null | null | astro-ph.HE astro-ph.CO hep-ph | http://creativecommons.org/licenses/by/4.0/ | The Cherenkov Telescope Array Observatory (CTAO) will be a ground-based Cherenkov telescope performing wide-sky surveys, ideal for anisotropy studies such as cross-correlations with tracers of the cosmic large-scale structure. Cross-correlations can shed light on high-energy $\gamma$-ray sources and potentially reveal exotic signals from particle dark matter. In this work, we investigate CTAO sensitivity to cross-correlation signals between $\gamma$-ray emission and galaxy distributions. We find that by using dense, low-redshift catalogs like 2MASS, and for integration times around 50 hours, this technique achieves sensitivities to both annihilating and decaying dark matter signals that are competitive with those from dwarf galaxy and cluster analyses. | [
{
"version": "v1",
"created": "Mon, 26 May 2025 18:00:00 GMT"
}
] | 2025-05-28T00:00:00 | [
[
"Pinetti",
"Elena",
""
],
[
"Vodeb",
"Veronika",
""
],
[
"Amerio",
"Aurelio",
""
],
[
"Cuoco",
"Alessandro",
""
],
[
"Camera",
"Stefano",
""
],
[
"Fornengo",
"Nicolao",
""
],
[
"Zaharijas",
"Gabrijela",
""
]
] |
2505.20384 | Ryohei Kobayashi | Po-Shen Hsin, Ryohei Kobayashi | Generalized Hall Conductivities in Local Commuting Projector Models: Generalized Symmetries and Protected Surface Modes | 21 pages, 7 figures | null | null | null | cond-mat.str-el hep-th quant-ph | http://creativecommons.org/licenses/by/4.0/ | Hall conductivities are important characterizations of phases of matter. It is known that nonzero Hall conductivities are difficult to realize in local commuting projector lattice models due to no-go theorems in (2+1)D. In this work we construct local commuting projector models in (2+1)D and (3+1)D with nonzero generalized Hall conductivities for ordinary and higher-form continuous symmetries on tensor product Hilbert space of finite local dimension. The model is given by a standard $\mathbb{Z}_N$ toric code, but the symmetries do not admit expression in terms of onsite charge operators. The symmetry do not have local charges or currents on the lattice in the absence of boundaries, but there is still notion of Hall conductivities that coincide with the continuum field theories. We construct protected gapless boundaries of the lattice models using modified Villain formalism. The generalized Hall conductivities are computed by surface currents as well as bulk flux insertion and many body Chern number. | [
{
"version": "v1",
"created": "Mon, 26 May 2025 18:00:00 GMT"
}
] | 2025-05-28T00:00:00 | [
[
"Hsin",
"Po-Shen",
""
],
[
"Kobayashi",
"Ryohei",
""
]
] |
2505.20385 | Hao Geng | Hao Geng, Ling-Yan Hung, Yikun Jiang | It from ETH: Multi-interval Entanglement and Replica Wormholes from Large-$c$ BCFT Ensemble | 70 pages, 42 figures | null | null | null | hep-th gr-qc | http://creativecommons.org/licenses/by/4.0/ | We provide a derivation of the Ryu-Takayanagi (RT) formula in 3D gravity for generic boundary subsystems--including RT surface phase transitions--directly from the dual two-dimensional conformal field theory (CFT). Our approach relies on the universal statistics of the algebraic conformal data and the large-$c$ behavior of conformal blocks with Cardy boundaries involved. We observe the emergence of 3D multi-boundary black holes with Karch-Randall branes from entangled states of any number of CFT's with and without Cardy boundaries. We obtain the RT formula from the CFT, in the high-temperature regime. Two direct applications are: $\textbf{1)}$ A simple derivation of the multi-interval entanglement entropy for the vacuum state of a single CFT; $\textbf{2)}$ A CFT-based detection of the emergence of replica wormholes in the context of entanglement islands and black hole microstate counting. Our framework yields the first holographic random tensor network that faithfully captures the entanglement structure of holographic CFTs. These results imply that bulk spacetime geometries indeed emerge from the eigenstate thermalization hypothesis (ETH) in the dual field theory in the large-$c$ limit--a paradigm we refer to as $\textit{It from ETH}$. | [
{
"version": "v1",
"created": "Mon, 26 May 2025 18:00:00 GMT"
}
] | 2025-05-28T00:00:00 | [
[
"Geng",
"Hao",
""
],
[
"Hung",
"Ling-Yan",
""
],
[
"Jiang",
"Yikun",
""
]
] |
2505.20386 | Thomas Steingasser | Thomas Steingasser, Mark P. Hertzberg and David I. Kaiser | Precision Unitarity Calculations in Inflationary Models | 39 pages, 3 figures | null | null | MIT-CTP/5860 | hep-ph astro-ph.CO hep-th | http://creativecommons.org/licenses/by/4.0/ | We revisit perturbative unitarity in scalar field inflation with a nonminimal coupling, with Higgs inflation serving as the most prominent example. Although such models are phenomenologically successful, it is critical to examine whether or not unitarity violations spoil their theoretical self-consistency. The analysis of these issues has so far typically relied on order-of-magnitude estimates of scattering amplitudes, which are appropriate for generic parameters. It is not evident that these methods apply to scenarios relying on a near-critical inflationary potential, for which an interplay of both small scalar self-couplings and nonminimal couplings could partially alleviate the unitarity issues. To allow for an exploration of this possibility, we consider the full $S$-matrix for the relevant scattering processes, taking into account important phase space volume factors, leading to a precise evaluation of the cut-off scale. In the single-field case, we demonstrate that near-criticality raises the cut-off scale considerably, compared to previous estimates. In the multifield case, momentum-dependent self-interactions in the kinetic sector lower the cut-off compared to the single-field case to a value comparable to but slightly larger than previous estimates. We carefully study both the single-field and multifield cases in metric and metric-affine (Palatini) formulations of gravity, as well as introduce a new phenomenologically viable model with a canonical kinetic term and a significantly raised cut-off, and discuss the importance of background field effects. | [
{
"version": "v1",
"created": "Mon, 26 May 2025 18:00:00 GMT"
}
] | 2025-05-28T00:00:00 | [
[
"Steingasser",
"Thomas",
""
],
[
"Hertzberg",
"Mark P.",
""
],
[
"Kaiser",
"David I.",
""
]
] |
2505.20387 | Jad C. Halimeh | Julian Schuhmacher, Guo-Xian Su, Jesse J. Osborne, Anthony Gandon, Jad C. Halimeh, Ivano Tavernelli | Observation of hadron scattering in a lattice gauge theory on a quantum computer | $11+13$ pages, $4+13$ figures, $0+1$ table | null | null | null | quant-ph cond-mat.quant-gas cond-mat.str-el hep-lat nucl-th | http://creativecommons.org/licenses/by/4.0/ | Scattering experiments are at the heart of high-energy physics (HEP), breaking matter down to its fundamental constituents, probing its formation, and providing deep insight into the inner workings of nature. In the current huge drive to forge quantum computers into complementary venues that are ideally suited to capture snapshots of far-from-equilibrium HEP dynamics, a major goal is to utilize these devices for scattering experiments. A major obstacle in this endeavor has been the hardware overhead required to access the late-time post-collision dynamics while implementing the underlying gauge symmetry. Here, we report on the first quantum simulation of scattering in a lattice gauge theory (LGT), performed on \texttt{IBM}'s \texttt{ibm\_marrakesh} quantum computer. Specifically, we quantum-simulate the collision dynamics of electrons and positrons as well as mesons in a $\mathrm{U}(1)$ LGT representing $1+1$D quantum electrodynamics (QED), uncovering rich post-collision dynamics that we can precisely tune with a topological $\Theta$-term and the fermionic mass. By monitoring the time evolution of the scattering processes, we are able to distinguish between two main regimes in the wake of the collision. The first is characterized by the delocalization of particles when the topological $\Theta$-term is weak, while the second regime shows localized particles with a clear signature when the $\Theta$-term is nontrivial. Furthermore, we show that by quenching to a small mass at the collision point, inelastic scattering occurs with a large production of matter reminiscent of quantum many-body scarring. Our work provides a major step forward in the utility of quantum computers for investigating the real-time quantum dynamics of HEP collisions. | [
{
"version": "v1",
"created": "Mon, 26 May 2025 18:00:00 GMT"
}
] | 2025-05-28T00:00:00 | [
[
"Schuhmacher",
"Julian",
""
],
[
"Su",
"Guo-Xian",
""
],
[
"Osborne",
"Jesse J.",
""
],
[
"Gandon",
"Anthony",
""
],
[
"Halimeh",
"Jad C.",
""
],
[
"Tavernelli",
"Ivano",
""
]
] |
2505.20388 | Carlos Nunez | Carlos Nunez and Dibakar Roychowdhury | Time-like Entanglement Entropy: a top-down approach | 9 pages, four figures | null | null | null | hep-th | http://creativecommons.org/licenses/by/4.0/ | We investigate the concept of time-like entanglement entropy (tEE) within the framework of holography. We introduce a robust top-down prescription for computing tEE in higher-dimensional QFTs, both conformal and confining, eliminating the ambiguities typically associated with analytic continuation from Euclidean to Lorentzian signatures. We present accurate analytic approximations for tEE and time-like separations in slab geometries. We establish a clear stability criterion for bulk embeddings and demonstrate that tEE serves as a powerful tool for computing CFT central charges, extending and strengthening previous results. Finally, we apply our framework to holographic confining backgrounds, revealing distinctive behaviours like phase transitions. | [
{
"version": "v1",
"created": "Mon, 26 May 2025 18:00:00 GMT"
}
] | 2025-05-28T00:00:00 | [
[
"Nunez",
"Carlos",
""
],
[
"Roychowdhury",
"Dibakar",
""
]
] |
2505.20389 | Damiano Francesco Giuseppe Fiorillo | Damiano F. G. Fiorillo, Georg G. Raffelt | Dispersion relation of the neutrino plasma: Unifying fast, slow, and collisional instabilities | 43 pages, 5 figures | null | null | null | hep-ph astro-ph.CO astro-ph.HE | http://arxiv.org/licenses/nonexclusive-distrib/1.0/ | In neutrino-dense astrophysical environments, these particles exchange flavor through a coherent weak field, forming a collisionless neutrino plasma with collective flavor dynamics. Instabilities, which grow and affect the environment, may arise from neutrino-neutrino refraction alone (fast limit), vacuum energy splittings caused by masses (slow limit), or neutrino-matter scattering (collisional limit). We present a comprehensive analytical description of the dispersion relation governing these unstable modes. Treating vacuum energy splittings and collision rates as small perturbations, we construct a unified framework for fast, slow, and collisional instabilities. We classify modes into gapped, where collective excitations are already present in the fast limit but rendered unstable by slow or collisional effects, and gapless, which are purely generated by these effects. For each class, we derive approximate dispersion relations for generic energy and angle distributions, which reveal the order of magnitude of the growth rates and the nature of the instabilities without solving directly the dispersion relation. This approach confirms that slow and collisionally unstable waves generally grow much more slowly than they oscillate. Consequently, the common fast-mode approximation of local evolution within small boxes is unjustified. Even for fast modes, neglecting large-distance propagation of growing waves, as usually done, may be a poor approximation. Our unified framework provides an intuitive understanding of the linear phase of flavor evolution across all regimes and paves the way for a quasi-linear treatment of the instability's nonlinear development. | [
{
"version": "v1",
"created": "Mon, 26 May 2025 18:00:00 GMT"
}
] | 2025-05-28T00:00:00 | [
[
"Fiorillo",
"Damiano F. G.",
""
],
[
"Raffelt",
"Georg G.",
""
]
] |
2505.20390 | Yasunori Nomura | Yasunori Nomura, Tomonori Ugajin | Nonperturbative Quantum Gravity in a Closed Lorentzian Universe | 20 pages, 8 figures | null | null | RIKEN-iTHEMS-Report-25 | hep-th gr-qc | http://creativecommons.org/licenses/by/4.0/ | We study how meaningful physical predictions can arise in nonperturbative quantum gravity in a closed Lorentzian universe. In such settings, recent developments suggest that the quantum gravitational Hilbert space is one-dimensional and real for each $\alpha$-sector, as induced by spacetime wormholes. This appears to obstruct the conventional quantum-mechanical prescription of assigning probabilities via projection onto a basis of states. While previous approaches have introduced external observers or augmented the theory to resolve this issue, we argue that quantum gravity itself contains all the necessary ingredients to make physical predictions. We demonstrate that the emergence of classical observables and probabilistic outcomes can be understood as a consequence of partial observability: physical observers access only a subsystem of the universe. Tracing out the inaccessible degrees of freedom yields reduced density matrices that encode classical information, with uncertainties exponentially suppressed by the environment's entropy. We develop this perspective using both the Lorentzian path integral and operator formalisms and support it with a simple microscopic model. Our results show that quantum gravity in a closed universe naturally gives rise to meaningful, robust predictions without recourse to external constructs. | [
{
"version": "v1",
"created": "Mon, 26 May 2025 18:00:00 GMT"
}
] | 2025-05-28T00:00:00 | [
[
"Nomura",
"Yasunori",
""
],
[
"Ugajin",
"Tomonori",
""
]
] |
2505.20391 | Aris Tritsis | Aris Tritsis | Ambipolar diffusion and the mass-to-flux ratio in a turbulent collapsing cloud | 11 pages, 13 figures. Submitted to A&A. Comments are welcome | null | null | null | astro-ph.GA | http://creativecommons.org/licenses/by/4.0/ | The formation of stars is governed by the intricate interplay of nonideal magnetohydrodynamic (MHD) effects, gravity, and turbulence. Computational challenges have hindered a comprehensive 3D exploration of this interplay, posing a longstanding challenge in our understanding of clouds and cores. Our objective was to study the spatial features and time evolution of the neutral-ion drift velocity and the mass-to-flux ratio in a 3D nonideal MHD chemo-dynamical simulation of a supercritical turbulent collapsing molecular cloud. The resistivities of the cloud were computed self-consistently from a vast non-equilibrium chemical network containing 115 species. To compute the resistivities we used different mean collisional rates for each charged species in our network. We additionally developed a new generalized method for measuring the true mass-to-flux ratio in 3D simulations. Despite the cloud's turbulent nature, at early times, the neutral-ion drift velocity follows the expected structure from 2D axisymmetric non-ideal MHD simulations with an hourglass magnetic field. At later times, however, the neutral-ion drift velocity becomes increasingly complex, with many vectors pointing outward from the cloud's center. Specifically, we find that the drift velocity above and below the cloud's ``midplane'' is in ``antiphase''. We explain these features on the basis of magnetic helical loops and the correlation of the drift velocity with the magnetic tension force per unit volume. Despite the complex structure of the neutral-ion drift velocity, we demonstrate that, when averaged over a region, the true mass-to-flux ratio monotonically increases as a function of time and decreases as a function of the radius from the center of the cloud. In contrast, the ``observed'' mass-to-flux ratio shows poor correlation with the true mass-to-flux ratio and the density structure of the cloud. | [
{
"version": "v1",
"created": "Mon, 26 May 2025 18:00:01 GMT"
}
] | 2025-05-28T00:00:00 | [
[
"Tritsis",
"Aris",
""
]
] |
2505.20392 | Francesco Salvestrini | Francesco Salvestrini, Simone Bianchi, Edvige Corbelli | Exploring the interplay of dust and gas phases in DustPedia star-forming galaxies | 12 pages, 7 figures, 1 table, 6 appendices Accepted and in press at A&A | null | null | null | astro-ph.GA | http://creativecommons.org/licenses/by-nc-nd/4.0/ | Molecular gas is the key ingredient of the star formation cycle, and tracing its dependencies on other galaxy properties is essential for understanding galaxy evolution. In this work, we explore the relation between the different phases of the interstellar medium (ISM), namely molecular gas, atomic gas, and dust, and galaxy properties using a sample of nearby late-type galaxies. To this goal, we collect CO maps for 121 galaxies from the DustPedia project, ensuring an accurate determination of $M_{H2}$, the global molecular gas mass. We investigate which scaling relations provide the best description of $M_{H2}$, based on the strength of the correlation and its intrinsic dispersion. Commonly used correlations between $M_{H2}$ and star formation rate (SFR) and stellar mass ($M_{\star}$) are affected by large scatter, which accounts for galaxies that are experiencing quenching of their star formation activity. This issue can be partially mitigated by considering a "fundamental plane" of star formation, fitting together $M_{H2}$, $M_{\star}$, and SFR. We confirm previous results from the DustPedia collaboration that the total gas mass has the tightest connection with the dust mass and that the molecular component also establishes a good correlation with dust. Although dust grains are necessary for the formation of hydrogen molecules, the strength of gravitational potential driven by the stellar component plays a key role in driving density enhancements and the atomic-to-molecular phase transition. Eventually, we investigated the correlations between ISM components and monochromatic luminosities at different wavelengths: we proposed mid and far-IR luminosities as reliable proxies of $L^{\prime}_{CO}$ for sources lacking dedicated millimeter observations. Luminosities in mid-IR photometric bands collecting PAH emission can be used to trace molecular gas and dust masses. | [
{
"version": "v1",
"created": "Mon, 26 May 2025 18:00:01 GMT"
}
] | 2025-05-28T00:00:00 | [
[
"Salvestrini",
"Francesco",
""
],
[
"Bianchi",
"Simone",
""
],
[
"Corbelli",
"Edvige",
""
]
] |
2505.20393 | Mingyang Zhuang | Ming-Yang Zhuang, Junyao Li, Yue Shen, Xiaojing Lin, Alice E. Shapley, Feige Wang, Qiaoya Wu, Qian Yang | NEXUS: A Spectroscopic Census of Broad-line AGNs and Little Red Dots at $3\lesssim z\lesssim 6$ | 26 pages, 17 figures, 4 tables. Submitted to ApJ | null | null | null | astro-ph.GA | http://creativecommons.org/licenses/by/4.0/ | We present a spectroscopic sample of 23 broad-line AGNs (BLAGNs) at $3\lesssim z\lesssim 6$ selected using F322W2+F444W NIRCam/WFSS grism spectroscopy of the central 100 ${\rm arcmin^2}$ area of the NEXUS survey. Among these BLAGNs, 15 are classified as Little Red Dots (LRDs) based on their rest-frame UV-optical spectral slopes and compact morphology. The number density of LRDs is $\sim 10^{-5}\,{\rm cMpc^{-3}}$, with a hint of declining towards the lower end of the probed redshift range. These BLAGNs and LRDs span broad H$\alpha$ luminosities of $\sim 10^{42.2}-10^{43.7}\,{\rm erg\,s^{-1}}$, black hole masses of $\sim 10^{6.3}-10^{8.4}\,M_\odot$, and Eddington ratios of $\sim 0.1-1$ (median value 0.4), though the black hole mass and Eddington ratio estimates carry large systematic uncertainties. Half of the LRDs show strong Balmer absorption, suggesting high-density gas surrounding the line-emitting region. We detect extended (hundreds of parsec) rest-frame UV-optical emission from the host galaxy in the majority of these LRDs, which contributes significantly or even dominantly to their total UV emission. This host emission largely accounts for the peculiar UV upturn of the LRD spectral energy distribution. We also measure the small-scale ($\lesssim 1\,{\rm cMpc}$) clustering of these BLAGNs and LRDs by cross-correlating with a photometric galaxy sample. Extrapolating the power-law two-point correlation function model to large linear scales, we infer a linear bias of $3.30_{-2.04}^{+2.88}$ and typical halo masses of a few $\times 10^{11}\,h^{-1}M_\odot$ for BLAGNs at the sample median redshift of $z\sim 4.5$. However, the inferred linear bias and halo masses of LRDs, while formally consistent with those for BLAGNs at $\sim 1.5\sigma$, appear too large to be compatible with their space density, suggesting LRDs may have strong excess clustering on small scales. | [
{
"version": "v1",
"created": "Mon, 26 May 2025 18:00:01 GMT"
}
] | 2025-05-28T00:00:00 | [
[
"Zhuang",
"Ming-Yang",
""
],
[
"Li",
"Junyao",
""
],
[
"Shen",
"Yue",
""
],
[
"Lin",
"Xiaojing",
""
],
[
"Shapley",
"Alice E.",
""
],
[
"Wang",
"Feige",
""
],
[
"Wu",
"Qiaoya",
""
],
[
"Yang",
"Qian",
""
]
] |
2505.20394 | Christopher Boettner | Christopher Boettner | gallifrey: JAX-based Gaussian process structure learning for astronomical time series | 16 pages, 10 figures, 3 tables, accepted in A&A, code available at https://github.com/ChrisBoettner/gallifrey, documentation available at https://chrisboettner.github.io/gallifrey/ | null | 10.1051/0004-6361/202554518 | null | astro-ph.IM astro-ph.EP | http://creativecommons.org/licenses/by-sa/4.0/ | Gaussian processes (GPs) have become a common tool in astronomy for analysing time series data, particularly in exoplanet science and stellar astrophysics. However, choosing the appropriate covariance structure for a GP model remains a challenge in many situations, limiting model flexibility and performance. This work provides an introduction to recent advances in GP structure learning methods, which enable the automated discovery of optimal GP kernels directly from the data, with the aim of making these methods more accessible to the astronomical community. We present gallifrey, a JAX-based Python package that implements a sequential Monte Carlo algorithm for Bayesian kernel structure learning. This approach defines a prior distribution over kernel structures and hyperparameters, and efficiently samples the GP posterior distribution using a novel involutive Markov chain Monte Carlo procedure. We applied gallifrey to common astronomical time series tasks, including stellar variability modelling, exoplanet transit modelling, and transmission spectroscopy. We show that this methodology can accurately interpolate and extrapolate stellar variability, recover transit parameters with robust uncertainties, and derive transmission spectra by effectively separating the background from the transit signal. When compared with traditional fixed-kernel approaches, we show that structure learning has advantages in terms of accuracy and uncertainty estimation. Structure learning can enhance the performance of GP regression for astronomical time series modelling. We discuss a road map for algorithmic improvements in terms of scalability to larger datasets, so that the methods presented here can be applied to future stellar and exoplanet missions such as PLATO. | [
{
"version": "v1",
"created": "Mon, 26 May 2025 18:00:01 GMT"
}
] | 2025-05-28T00:00:00 | [
[
"Boettner",
"Christopher",
""
]
] |
2505.20395 | Alessandro Strumia | Ferruccio Feruglio, Antonio Marrone, Alessandro Strumia, Arsenii Titov | Solving the strong CP problem in string-inspired theories with modular invariance | 19 pages, 3 figures. Webinar presentation: https://youtu.be/gJPWXWFYnl4 | null | null | null | hep-ph hep-th | http://arxiv.org/licenses/nonexclusive-distrib/1.0/ | We show that solutions to the strong CP problem based on modular invariance can be extended to incorporate features that appear in string compactifications: quarks with mostly positive modular weights and non-trivial gauge kinetic functions. This requires assuming that singularities and zeroes only appear at special points, such as decompactification limits. We discuss the impact of these assumptions on string gauge unification. | [
{
"version": "v1",
"created": "Mon, 26 May 2025 18:00:01 GMT"
}
] | 2025-05-28T00:00:00 | [
[
"Feruglio",
"Ferruccio",
""
],
[
"Marrone",
"Antonio",
""
],
[
"Strumia",
"Alessandro",
""
],
[
"Titov",
"Arsenii",
""
]
] |
2505.20396 | Sebastian Trojanowski | Sebastian Trojanowski, Lei Zu | Cosmological impact of $\nu$DM interactions enhanced in narrow redshift ranges | 10 pages, 4 figures | null | null | null | astro-ph.CO hep-ph | http://arxiv.org/licenses/nonexclusive-distrib/1.0/ | The impact of dark matter-neutrino ($\nu$DM) interactions on cosmological perturbations has regained attention, spurred by indications of non-zero couplings from high-multipole cosmic microwave background data, weak lensing, and Lyman-$\alpha$ observations. We demonstrate that a similar observational preference is obtained if $\nu$DM interactions are primarily enhanced during a specific epoch, $z\sim (10^4-10^5)$, leading to $>3\sigma$ preference for a non-zero interaction in the combined Atacama Cosmology Telescope and cosmic shear data. This redshift-limited enhancement circumvents other cosmological and astrophysical bounds and can be achieved within a neutrino portal dark matter framework incorporating resonantly enhanced scattering rates. | [
{
"version": "v1",
"created": "Mon, 26 May 2025 18:00:01 GMT"
}
] | 2025-05-28T00:00:00 | [
[
"Trojanowski",
"Sebastian",
""
],
[
"Zu",
"Lei",
""
]
] |
2505.20397 | Emre Can Sert\"oz | Emre Can Sert\"oz, Jo\"el Ouaknine, James Worrell | Computing transcendence and linear relations of 1-periods | 68 pages, comments welcome | null | null | null | math.AG cs.SC math.NT | http://creativecommons.org/licenses/by/4.0/ | A 1-period is a complex number given by the integral of a univariate algebraic function, where all data involved -- the integrand and the domain of integration -- are defined over algebraic numbers. We give an algorithm that, given a finite collection of 1-periods, computes the space of all linear relations among them with algebraic coefficients. In particular, the algorithm decides whether a given 1-period is transcendental, and whether two 1-periods are equal. This resolves, in the case of 1-periods, a problem posed by Kontsevich and Zagier, asking for an algorithm to decide equality of periods. The algorithm builds on the work of Huber and W\"ustholz, who showed that all linear relations among 1-periods arise from 1-motives; we make this perspective effective by reducing the problem to divisor arithmetic on curves and providing the theoretical foundations for a practical and fully explicit algorithm. To illustrate the broader applicability of our methods, we also give an algorithmic classification of autonomous first-order (non-linear) differential equations. | [
{
"version": "v1",
"created": "Mon, 26 May 2025 18:00:01 GMT"
}
] | 2025-05-28T00:00:00 | [
[
"Sertöz",
"Emre Can",
""
],
[
"Ouaknine",
"Joël",
""
],
[
"Worrell",
"James",
""
]
] |
2505.20398 | Pok Man Tam | Pok Man Tam, Hao Chen, Biao Lian | Quantized Transport of Disordered Superconducting $\nu=2/3$ Fractional Quantum Hall Edges | Main: 4.5 pages, 3 figures; Supplemental Material: 5 sections | null | null | null | cond-mat.mes-hall cond-mat.dis-nn cond-mat.str-el cond-mat.supr-con | http://arxiv.org/licenses/nonexclusive-distrib/1.0/ | The $\nu=2/3$ fractional quantum Hall (FQH) edge states, which have counter-propagating modes, are known to flow under relevant neutral disorders into a stable Kane-Fisher-Polchinski (KFP) renormalization group (RG) fixed point, which yields a quantized two-terminal conductance $\frac{2}{3}\frac{e^2}{h}$ consistent with the experiments. Motivated by growing interests in superconducting (SC) quantum Hall systems, we study the $\nu=2/3$ FQH edge states with disordered SC proximity by generalizing the KFP analysis, which may also apply to the recently realized $\nu=2/3$ fractional Chern insulator. We show that the $\nu=2/3$ FQH edge theory has an infinite number of stable RG fixed points SC$_N$ labeled by an integer $N\in\mathbb{Z}$, each of which is driven by a relevant disordered charge-$2q_N$ SC tunneling with $q_N\in\mathbb{Z}$ depending on $N$. The $N<0$ ($N\ge0$) phases are favored by an attractive (repulsive) inter-mode interaction on the edge. We further predict that the edge states in a SC$_N$ phase with $q_N\neq0$ yields a quantized downstream resistance $R_d=\frac{h}{2q_N^2e^2}$ measurable in a FQH-SC junction. For edge states staying in the KFP phase ($N=0$ and $q_N=0$) under SC proximity, we arrive at a nonlinear $R_d\propto V^{-\alpha}$ or $T^{-\alpha}$ with voltage bias $V$ or temperature $T$, where $\alpha=4$ ($\alpha=1$) if pairing (vortex tunneling) dominates. | [
{
"version": "v1",
"created": "Mon, 26 May 2025 18:00:02 GMT"
}
] | 2025-05-28T00:00:00 | [
[
"Tam",
"Pok Man",
""
],
[
"Chen",
"Hao",
""
],
[
"Lian",
"Biao",
""
]
] |
2505.20399 | Benedikt Diemer | Benedikt Diemer | ULULA: An ultra-lightweight 2D hydrodynamics code for teaching and experimentation | 11 pages, 8 figures. Comments welcome. Documentation at https://bdiemer.bitbucket.io/ulula/index.html | null | null | null | astro-ph.IM | http://creativecommons.org/licenses/by/4.0/ | Hydrodynamics is a difficult subject to teach in the classroom because most relevant problems must be solved numerically rather than analytically. While there are numerous public hydrodynamics codes, the complexity of production-level software obscures the underlying physics and can be overwhelming to first-time users. Here we present ULULA, an ultra-lightweight python code to solve hydrodynamics and gravity in 2D. The main goal is for the code to be easy to understand, extend, and experiment with. The simulation framework consists of fewer than 800 active lines of pure python code, but it includes a robust MUSCL-Hancock scheme with exchangeable components such as Riemann solvers, reconstruction schemes, boundary conditions, and equations of state. Numerous well-known hydrodynamics problems are provided and can be run in a few minutes on a laptop. The code is open-source, generously commented, and extensively documented. | [
{
"version": "v1",
"created": "Mon, 26 May 2025 18:00:02 GMT"
}
] | 2025-05-28T00:00:00 | [
[
"Diemer",
"Benedikt",
""
]
] |
2505.20400 | Hynek Paul | Nikolay Bobev and Hynek Paul | Holographic Correlators Beyond Maximal Supersymmetry | 52 pages, 3 figures | null | null | null | hep-th | http://arxiv.org/licenses/nonexclusive-distrib/1.0/ | We use the AdS/CFT correspondence to explicitly calculate some of the three-point functions in the planar limit of the 4d $\mathcal{N}=1$ Leigh-Strassler SCFT. This strongly interacting CFT can be obtained as a mass deformation of the 4d $\mathcal{N}=4$ SYM theory and admits a dual description in terms of an AdS$_5$ background of type IIB supergravity. Our analysis is based on the existence of a consistent truncation of the 10d supergravity to a tractable 5d gravitational theory with 10 scalar fields dual to some of the low-lying operators in the spectrum of the LS SCFT. We apply standard holographic techniques to this 10-scalar model to analytically calculate the correlators of interest and thus provide a rare example of explicit three-point functions of scalar non-BPS operators in strongly coupled 4d CFTs. Using superconformal Ward identities we perform several consistency checks of these holographic correlators. As a byproduct of our analysis we discuss some subtleties related to the calculation of extremal correlators in AdS/CFT and the contribution of scalar derivative couplings to the evaluation of Witten diagrams. Our work provides a blueprint for the holographic calculation of other correlators in the LS SCFT and similar top-down holographic models. | [
{
"version": "v1",
"created": "Mon, 26 May 2025 18:00:02 GMT"
}
] | 2025-05-28T00:00:00 | [
[
"Bobev",
"Nikolay",
""
],
[
"Paul",
"Hynek",
""
]
] |
2505.20401 | Ricardo Castillo | Brandon Carhuas, Ricardo Castillo, Ricardo Freire, Alex Lira, Miguel Loayza | Global and nonglobal solutions for a mixed local-nonlocal heat equation | 7 pages | null | null | null | math.AP | http://creativecommons.org/licenses/by/4.0/ | In this work, we establish optimal conditions concerning the global and nonglobal existence of solutions of a semilinear parabolic equations governed by a mixed local-nonlocal operator. Furthermore, our findings recover the Fujita exponent recently derived by Biagi, Punzo and Vecchi, as well as by Del Pezzo and Ferreira. | [
{
"version": "v1",
"created": "Mon, 26 May 2025 18:00:03 GMT"
}
] | 2025-05-28T00:00:00 | [
[
"Carhuas",
"Brandon",
""
],
[
"Castillo",
"Ricardo",
""
],
[
"Freire",
"Ricardo",
""
],
[
"Lira",
"Alex",
""
],
[
"Loayza",
"Miguel",
""
]
] |
2505.20402 | Clemente Smarra | Clemente Smarra, Lodovico Capuano, Adrien Kuntz | Probing supermassive black hole scalarization with Pulsar Timing Arrays | 23 pages, 9 figures | null | null | null | gr-qc astro-ph.CO astro-ph.HE hep-th | http://creativecommons.org/licenses/by/4.0/ | Scalar-tensor theories with a scalar field coupled to the Gauss-Bonnet invariant can evade no-hair theorems and allow for non-trivial scalar profiles around black holes. This coupling is characterized by a length scale $\lambda$, which, in an effective field theory perspective, sets the threshold below which deviations from General Relativity become significant. LIGO/VIRGO constraints indicate $\lambda$ is small, implying supermassive black holes should not scalarize. However, recent work suggests that scalarization can occur within a narrow window of masses, allowing supermassive black holes to scalarize, while leaving LIGO/VIRGO sources unaffected. We explore the impact of this scenario on the stochastic gravitational wave background recently observed by Pulsar Timing Arrays. We find that scalarization can alter the characteristic strain produced by circularly inspiralling SMBH binaries and that current data shows a marginal preference for a non-zero $\lambda$. However, similar signatures could arise from astrophysical effects such as orbital eccentricity or environmental interactions, emphasizing the need for improved modeling and longer observations to discriminate among the different scenarios. | [
{
"version": "v1",
"created": "Mon, 26 May 2025 18:00:03 GMT"
}
] | 2025-05-28T00:00:00 | [
[
"Smarra",
"Clemente",
""
],
[
"Capuano",
"Lodovico",
""
],
[
"Kuntz",
"Adrien",
""
]
] |
2505.20403 | Andrea Bulgarelli | Andrea Bulgarelli, Michele Caselle, Alessandro Nada and Marco Panero | Casimir effect in critical $\mathrm{O}(N)$ models from non-equilibrium Monte Carlo simulations | 11 pages, 7 figures | null | null | HIP-2025-15/TH | cond-mat.stat-mech hep-lat hep-th | http://creativecommons.org/licenses/by-nc-nd/4.0/ | $\mathrm{O}(N)$ vector models in three dimensions, when defined in a geometry with a compact direction and tuned to criticality, exhibit long-range fluctuations which induce a Casimir effect. The strength of the resulting interaction is encoded in the excess free-energy density, which depends on a universal coefficient: the Casimir amplitude. We present a high-precision numerical calculation of the latter, by means of a novel non-equilibrium Monte Carlo algorithm, and compare our findings with results obtained from large-$N$ expansions and from the conformal bootstrap. | [
{
"version": "v1",
"created": "Mon, 26 May 2025 18:00:05 GMT"
}
] | 2025-05-28T00:00:00 | [
[
"Bulgarelli",
"Andrea",
""
],
[
"Caselle",
"Michele",
""
],
[
"Nada",
"Alessandro",
""
],
[
"Panero",
"Marco",
""
]
] |
2505.20404 | Sha Yi | Sha Yi, Xueqian Bai, Adabhav Singh, Jianglong Ye, Michael T Tolley, Xiaolong Wang | Co-Design of Soft Gripper with Neural Physics | null | null | null | null | cs.RO | http://arxiv.org/licenses/nonexclusive-distrib/1.0/ | For robot manipulation, both the controller and end-effector design are crucial. Soft grippers are generalizable by deforming to different geometries, but designing such a gripper and finding its grasp pose remains challenging. In this paper, we propose a co-design framework that generates an optimized soft gripper's block-wise stiffness distribution and its grasping pose, using a neural physics model trained in simulation. We derived a uniform-pressure tendon model for a flexure-based soft finger, then generated a diverse dataset by randomizing both gripper pose and design parameters. A neural network is trained to approximate this forward simulation, yielding a fast, differentiable surrogate. We embed that surrogate in an end-to-end optimization loop to optimize the ideal stiffness configuration and best grasp pose. Finally, we 3D-print the optimized grippers of various stiffness by changing the structural parameters. We demonstrate that our co-designed grippers significantly outperform baseline designs in both simulation and hardware experiments. | [
{
"version": "v1",
"created": "Mon, 26 May 2025 18:00:09 GMT"
}
] | 2025-05-28T00:00:00 | [
[
"Yi",
"Sha",
""
],
[
"Bai",
"Xueqian",
""
],
[
"Singh",
"Adabhav",
""
],
[
"Ye",
"Jianglong",
""
],
[
"Tolley",
"Michael T",
""
],
[
"Wang",
"Xiaolong",
""
]
] |
2505.20405 | Davide Bucciarelli | Lorenzo Baraldi, Davide Bucciarelli, Federico Betti, Marcella Cornia, Lorenzo Baraldi, Nicu Sebe, Rita Cucchiara | What Changed? Detecting and Evaluating Instruction-Guided Image Edits with Multimodal Large Language Models | null | null | null | null | cs.CV cs.AI cs.CL cs.MM | http://arxiv.org/licenses/nonexclusive-distrib/1.0/ | Instruction-based image editing models offer increased personalization opportunities in generative tasks. However, properly evaluating their results is challenging, and most of the existing metrics lag in terms of alignment with human judgment and explainability. To tackle these issues, we introduce DICE (DIfference Coherence Estimator), a model designed to detect localized differences between the original and the edited image and to assess their relevance to the given modification request. DICE consists of two key components: a difference detector and a coherence estimator, both built on an autoregressive Multimodal Large Language Model (MLLM) and trained using a strategy that leverages self-supervision, distillation from inpainting networks, and full supervision. Through extensive experiments, we evaluate each stage of our pipeline, comparing different MLLMs within the proposed framework. We demonstrate that DICE effectively identifies coherent edits, effectively evaluating images generated by different editing models with a strong correlation with human judgment. We publicly release our source code, models, and data. | [
{
"version": "v1",
"created": "Mon, 26 May 2025 18:00:10 GMT"
}
] | 2025-05-28T00:00:00 | [
[
"Baraldi",
"Lorenzo",
""
],
[
"Bucciarelli",
"Davide",
""
],
[
"Betti",
"Federico",
""
],
[
"Cornia",
"Marcella",
""
],
[
"Baraldi",
"Lorenzo",
""
],
[
"Sebe",
"Nicu",
""
],
[
"Cucchiara",
"Rita",
""
]
] |
2505.20406 | M. Schuyler Moss | M. Schuyler Moss, Roeland Wiersema, Mohamed Hibat-Allah, Juan Carrasquilla, Roger G. Melko | Leveraging recurrence in neural network wavefunctions for large-scale simulations of Heisenberg antiferromagnets: the triangular lattice | 20 pages, 15 figures, 5 tables | null | null | null | cond-mat.str-el cond-mat.dis-nn quant-ph | http://arxiv.org/licenses/nonexclusive-distrib/1.0/ | Variational Monte Carlo simulations have been crucial for understanding quantum many-body systems, especially when the Hamiltonian is frustrated and the ground-state wavefunction has a non-trivial sign structure. In this paper, we use recurrent neural network (RNN) wavefunction ans\"{a}tze to study the triangular-lattice antiferromagnetic Heisenberg model (TLAHM) for lattice sizes up to $30\times30$. In a recent study [M. S. Moss et al. arXiv:2502.17144], the authors have demonstrated how RNN wavefunctions can be iteratively retrained in order to obtain variational results for multiple lattice sizes with a reasonable amount of compute. That study, which looked at the sign-free, square-lattice antiferromagnetic Heisenberg model, showed favorable scaling properties, allowing accurate finite-size extrapolations to the thermodynamic limit. In contrast, our present results illustrate in detail the relative difficulty in simulating the sign-problematic TLAHM. We find that the accuracy of our simulations can be significantly improved by transforming the Hamiltonian with a judicious choice of basis rotation. We also show that a similar benefit can be achieved by using variational neural annealing, an alternative optimization technique that minimizes a pseudo free energy. Ultimately, we are able to obtain estimates of the ground-state properties of the TLAHM in the thermodynamic limit that are in close agreement with values in the literature, showing that RNN wavefunctions provide a powerful toolbox for performing finite-size scaling studies for frustrated quantum many-body systems. | [
{
"version": "v1",
"created": "Mon, 26 May 2025 18:00:20 GMT"
}
] | 2025-05-28T00:00:00 | [
[
"Moss",
"M. Schuyler",
""
],
[
"Wiersema",
"Roeland",
""
],
[
"Hibat-Allah",
"Mohamed",
""
],
[
"Carrasquilla",
"Juan",
""
],
[
"Melko",
"Roger G.",
""
]
] |
Subsets and Splits
No community queries yet
The top public SQL queries from the community will appear here once available.