ArXiv Domain 2025-07-17
数据来源：ArXiv Domain
LLM Domain Papers
1. AirLLM: Diffusion Policy-based Adaptive LoRA for Remote Fine-Tuning of LLM over the Air
Operating Large Language Models (LLMs) on edge devices is increasingly challenged by limited communication bandwidth and strained computational and memory costs. Thus, cloud-assisted remote fine-tuning becomes indispensable. Nevertheless, existing Low-Rank Adaptation (LoRA) approaches typically employ fixed or heuristic rank configurations, and the subsequent over-the-air transmission of all LoRA parameters could be rather inefficient. To address this limitation, we develop AirLLM, a hierarchical diffusion policy framework for communication-aware LoRA adaptation. Specifically, AirLLM models the rank configuration as a structured action vector that spans all LoRA-inserted projections. To solve the underlying high-dimensional sequential decision-making problem, a Proximal Policy Optimization (PPO) agent generates coarse-grained decisions by jointly observing wireless states and linguistic complexity, which are then refined via Denoising Diffusion Implicit Models (DDIM) to produce high-resolution, task- and channel-adaptive rank vectors. The two modules are optimized alternatively, with the DDIM trained under the Classifier-Free Guidance (CFG) paradigm to maintain alignment with PPO rewards. Experiments under varying signal-to-noise ratios demonstrate that AirLLM consistently enhances fine-tuning performance while significantly reducing transmission costs, highlighting the effectiveness of reinforcement-driven, diffusion-refined rank adaptation for scalable and efficient remote fine-tuning over the air.
Authors: Shiyi Yang, Xiaoxue Yu, Rongpeng Li, Jianhang Zhu, Zhifeng Zhao, Honggang Zhang
Categories: cs.LG, cs.AI, cs.CL
PDF URL: https://arxiv.org/pdf/2507.11515v1.pdf
Published: 2025-07-15T17:36:37Z
2. Conversation Forests: The Key to Fine Tuning Large Language Models for Multi-Turn Medical Conversations is Branching
Fine-tuning methods such as Direct Preference Optimization (DPO) and Group Relative Policy Optimization (GRPO) have demonstrated success in training large language models (LLMs) for single-turn tasks. However, these methods fall short in multi-turn applications, such as diagnostic patient interviewing, where understanding how early conversational turns influence downstream completions and outcomes is essential. In medicine, a multi-turn perspective is critical for learning diagnostic schemas and better understanding conversation dynamics. To address this gap, I introduce Savage Conversation Forests (SCF), a reinforcement learning framework that leverages a branched conversation architecture to fine-tune LLMs for multi-turn dialogue. SCF generates multiple possible conversation continuations at each turn, enabling the model to learn how different early responses affect downstream interactions and diagnostic outcomes. In experiments simulating doctor-patient conversations, SCF with branching outperforms linear conversation architectures on diagnostic accuracy. I hypothesize that SCF’s improvements stem from its ability to provide richer, interdependent training signals across conversation turns. These results suggest that a branched training architecture is an important strategy for fine tuning LLMs in complex multi-turn conversational tasks.
Authors: Thomas Savage
Categories: cs.CL, cs.AI
PDF URL: https://arxiv.org/pdf/2507.04099v2.pdf
Published: 2025-07-05T16:49:34Z
3. A Generative Approach to LLM Harmfulness Detection with Special Red Flag Tokens
Most safety training methods for large language models (LLMs) are based on fine-tuning that forces models to shift from an unsafe answer to refusal when faced with harmful requests. Unfortunately, these drastic distribution shifts generally compromise model capabilities. To avoid that, we propose to expand the model’s vocabulary with a special token we call red flag token (
Authors: Sophie Xhonneux, David Dobre, Mehrnaz Mofakhami, Leo Schwinn, Gauthier Gidel
Categories: cs.CL, cs.AI, cs.CR, cs.LG
PDF URL: https://arxiv.org/pdf/2502.16366v3.pdf
Published: 2025-02-22T21:48:48Z
4. Reasoning Strategies in Large Language Models: Can They Follow, Prefer, and Optimize?
Human reasoning involves different strategies, each suited to specific problems. Prior work shows that large language model (LLMs) tend to favor a single reasoning strategy, potentially limiting their effectiveness in diverse reasoning challenges. In this work, we investigate whether prompting can control LLMs reasoning strategies and assess its impact on logical problem-solving. While our experiments show that no single strategy consistently improves accuracy, performance could be enhanced if models could adaptively choose the optimal strategy. We propose methods to guide LLMs in strategy selection, highlighting new ways to refine their reasoning abilities.
Authors: Yanjian Zhang, Guillaume Wisniewski, Nadi Tomeh, Thierry Charnois
Categories: cs.CL
PDF URL: https://arxiv.org/pdf/2507.11423v2.pdf
Published: 2025-07-15T15:47:47Z
5. Hallucination Stations: On Some Basic Limitations of Transformer-Based Language Models
In this paper we explore hallucinations and related capability limitations in LLMs and LLM-based agents from the perspective of computational complexity. We show that beyond a certain complexity, LLMs are incapable of carrying out computational and agentic tasks or verifying their accuracy.
Authors: Varin Sikka, Vishal Sikka
Categories: cs.CL, cs.AI
PDF URL: https://arxiv.org/pdf/2507.07505v3.pdf
Published: 2025-07-10T07:50:52Z
6. KisMATH: Do LLMs Have Knowledge of Implicit Structures in Mathematical Reasoning?
Chain-of-thought traces have been shown to improve performance of large language models in a plethora of reasoning tasks, yet there is no consensus on the mechanism through which this performance boost is achieved. To shed more light on this, we introduce Causal CoT Graphs (CCGs), which are directed acyclic graphs automatically extracted from reasoning traces that model fine-grained causal dependencies in the language model output. A collection of $1671$ mathematical reasoning problems from MATH500, GSM8K and AIME, and their associated CCGs are compiled into our dataset — \textbf{KisMATH}. Our detailed empirical analysis with 15 open-weight LLMs shows that (i) reasoning nodes in the CCG are mediators for the final answer, a condition necessary for reasoning; and (ii) LLMs emphasise reasoning paths given by the CCG, indicating that models internally realise structures akin to our graphs. KisMATH enables controlled, graph-aligned interventions and opens up avenues for further investigation into the role of chain-of-thought in LLM reasoning.
Authors: Soumadeep Saha, Akshay Chaturvedi, Saptarshi Saha, Utpal Garain, Nicholas Asher
Categories: cs.CL, cs.AI, I.2.7
PDF URL: https://arxiv.org/pdf/2507.11408v1.pdf
Published: 2025-07-15T15:28:37Z
7. EXAONE 4.0: Unified Large Language Models Integrating Non-reasoning and Reasoning Modes
This technical report introduces EXAONE 4.0, which integrates a Non-reasoning mode and a Reasoning mode to achieve both the excellent usability of EXAONE 3.5 and the advanced reasoning abilities of EXAONE Deep. To pave the way for the agentic AI era, EXAONE 4.0 incorporates essential features such as agentic tool use, and its multilingual capabilities are extended to support Spanish in addition to English and Korean. The EXAONE 4.0 model series consists of two sizes: a mid-size 32B model optimized for high performance, and a small-size 1.2B model designed for on-device applications. The EXAONE 4.0 demonstrates superior performance compared to open-weight models in its class and remains competitive even against frontier-class models. The models are publicly available for research purposes and can be easily downloaded via https://huggingface.co/LGAI-EXAONE.
Authors: LG AI Research, :, Kyunghoon Bae, Eunbi Choi, Kibong Choi, Stanley Jungkyu Choi, Yemuk Choi, Kyubeen Han, Seokhee Hong, Junwon Hwang, Taewan Hwang, Joonwon Jang, Hyojin Jeon, Kijeong Jeon, Gerrard Jeongwon Jo, Hyunjik Jo, Jiyeon Jung, Euisoon Kim, Hyosang Kim, Jihoon Kim, Joonkee Kim, Seonghwan Kim, Soyeon Kim, Sunkyoung Kim, Yireun Kim, Yongil Kim, Youchul Kim, Edward Hwayoung Lee, Gwangho Lee, Haeju Lee, Honglak Lee, Jinsik Lee, Kyungmin Lee, Sangha Park, Young Min Paik, Yongmin Park, Youngyong Park, Sanghyun Seo, Sihoon Yang, Heuiyeen Yeen, Sihyuk Yi, Hyeongu Yun
Categories: cs.CL, cs.AI
PDF URL: https://arxiv.org/pdf/2507.11407v1.pdf
Published: 2025-07-15T15:24:51Z
8. DCR: Quantifying Data Contamination in LLMs Evaluation
The rapid advancement of large language models (LLMs) has heightened concerns about benchmark data contamination (BDC), where models inadvertently memorize evaluation data, inflating performance metrics and undermining genuine generalization assessment. This paper introduces the Data Contamination Risk (DCR) framework, a lightweight, interpretable pipeline designed to detect and quantify BDC across four granular levels: semantic, informational, data, and label. By synthesizing contamination scores via a fuzzy inference system, DCR produces a unified DCR Factor that adjusts raw accuracy to reflect contamination-aware performance. Validated on 9 LLMs (0.5B-72B) across sentiment analysis, fake news detection, and arithmetic reasoning tasks, the DCR framework reliably diagnoses contamination severity and with accuracy adjusted using the DCR Factor to within 4% average error across the three benchmarks compared to the uncontaminated baseline. Emphasizing computational efficiency and transparency, DCR provides a practical tool for integrating contamination assessment into routine evaluations, fostering fairer comparisons and enhancing the credibility of LLM benchmarking practices.
Authors: Cheng Xu, Nan Yan, Shuhao Guan, Changhong Jin, Yuke Mei, Yibing Guo, M-Tahar Kechadi
Categories: cs.CL
PDF URL: https://arxiv.org/pdf/2507.11405v1.pdf
Published: 2025-07-15T15:23:53Z
9. Addressing Data Imbalance in Transformer-Based Multi-Label Emotion Detection with Weighted Loss
This paper explores the application of a simple weighted loss function to Transformer-based models for multi-label emotion detection in SemEval-2025 Shared Task 11. Our approach addresses data imbalance by dynamically adjusting class weights, thereby enhancing performance on minority emotion classes without the computational burden of traditional resampling methods. We evaluate BERT, RoBERTa, and BART on the BRIGHTER dataset, using evaluation metrics such as Micro F1, Macro F1, ROC-AUC, Accuracy, and Jaccard similarity coefficients. The results demonstrate that the weighted loss function improves performance on high-frequency emotion classes but shows limited impact on minority classes. These findings underscore both the effectiveness and the challenges of applying this approach to imbalanced multi-label emotion detection.
Authors: Xia Cui
Categories: cs.CL
PDF URL: https://arxiv.org/pdf/2507.11384v1.pdf
Published: 2025-07-15T14:53:33Z
10. What is the Best Process Model Representation? A Comparative Analysis for Process Modeling with Large Language Models
Large Language Models (LLMs) are increasingly applied for Process Modeling (PMo) tasks such as Process Model Generation (PMG). To support these tasks, researchers have introduced a variety of Process Model Representations (PMRs) that serve as model abstractions or generation targets. However, these PMRs differ widely in structure, complexity, and usability, and have never been systematically compared. Moreover, recent PMG approaches rely on distinct evaluation strategies and generation techniques, making comparison difficult. This paper presents the first empirical study that evaluates multiple PMRs in the context of PMo with LLMs. We introduce the PMo Dataset, a new dataset containing 55 process descriptions paired with models in nine different PMRs. We evaluate PMRs along two dimensions: suitability for LLM-based PMo and performance on PMG. \textit{Mermaid} achieves the highest overall score across six PMo criteria, whereas \textit{BPMN text} delivers the best PMG results in terms of process element similarity.
Authors: Alexis Brissard, Frédéric Cuppens, Amal Zouaq
Categories: cs.CL
PDF URL: https://arxiv.org/pdf/2507.11356v1.pdf
Published: 2025-07-15T14:26:50Z
Agent Domain Papers
1. Seven Security Challenges That Must be Solved in Cross-domain Multi-agent LLM Systems
Large language models (LLMs) are rapidly evolving into autonomous agents that cooperate across organizational boundaries, enabling joint disaster response, supply-chain optimization, and other tasks that demand decentralized expertise without surrendering data ownership. Yet, cross-domain collaboration shatters the unified trust assumptions behind current alignment and containment techniques. An agent benign in isolation may, when receiving messages from an untrusted peer, leak secrets or violate policy, producing risks driven by emergent multi-agent dynamics rather than classical software bugs. This position paper maps the security agenda for cross-domain multi-agent LLM systems. We introduce seven categories of novel security challenges, for each of which we also present plausible attacks, security evaluation metrics, and future research guidelines.
Authors: Ronny Ko, Jiseong Jeong, Shuyuan Zheng, Chuan Xiao, Tae-Wan Kim, Makoto Onizuka, Won-Yong Shin
Categories: cs.CR, cs.AI
PDF URL: https://arxiv.org/pdf/2505.23847v3.pdf
Published: 2025-05-28T18:19:03Z
2. Possible Principles for Aligned Structure Learning Agents
This paper offers a roadmap for the development of scalable aligned artificial intelligence (AI) from first principle descriptions of natural intelligence. In brief, a possible path toward scalable aligned AI rests upon enabling artificial agents to learn a good model of the world that includes a good model of our preferences. For this, the main objective is creating agents that learn to represent the world and other agents’ world models; a problem that falls under structure learning (a.k.a. causal representation learning or model discovery). We expose the structure learning and alignment problems with this goal in mind, as well as principles to guide us forward, synthesizing various ideas across mathematics, statistics, and cognitive science. 1) We discuss the essential role of core knowledge, information geometry and model reduction in structure learning, and suggest core structural modules to learn a wide range of naturalistic worlds. 2) We outline a way toward aligned agents through structure learning and theory of mind. As an illustrative example, we mathematically sketch Asimov’s Laws of Robotics, which prescribe agents to act cautiously to minimize the ill-being of other agents. We supplement this example by proposing refined approaches to alignment. These observations may guide the development of artificial intelligence in helping to scale existing — or design new — aligned structure learning systems.
Authors: Lancelot Da Costa, Tomáš Gavenčiak, David Hyland, Mandana Samiei, Cristian Dragos-Manta, Candice Pattisapu, Adeel Razi, Karl Friston
Categories: cs.AI, q-bio.NC
PDF URL: https://arxiv.org/pdf/2410.00258v2.pdf
Published: 2024-09-30T22:06:06Z
3. Taming Uncertainty via Automation: Observing, Analyzing, and Optimizing Agentic AI Systems
Large Language Models (LLMs) are increasingly deployed within agentic systems-collections of interacting, LLM-powered agents that execute complex, adaptive workflows using memory, tools, and dynamic planning. While enabling powerful new capabilities, these systems also introduce unique forms of uncertainty stemming from probabilistic reasoning, evolving memory states, and fluid execution paths. Traditional software observability and operations practices fall short in addressing these challenges. This paper introduces AgentOps: a comprehensive framework for observing, analyzing, optimizing, and automating operation of agentic AI systems. We identify distinct needs across four key roles-developers, testers, site reliability engineers (SREs), and business users-each of whom engages with the system at different points in its lifecycle. We present the AgentOps Automation Pipeline, a six-stage process encompassing behavior observation, metric collection, issue detection, root cause analysis, optimized recommendations, and runtime automation. Throughout, we emphasize the critical role of automation in managing uncertainty and enabling self-improving AI systems-not by eliminating uncertainty, but by taming it to ensure safe, adaptive, and effective operation.
Authors: Dany Moshkovich, Sergey Zeltyn
Categories: cs.AI, cs.MA
PDF URL: https://arxiv.org/pdf/2507.11277v1.pdf
Published: 2025-07-15T12:54:43Z
4. An Agentic Flow for Finite State Machine Extraction using Prompt Chaining
Finite-State Machines (FSMs) are critical for modeling the operational logic of network protocols, enabling verification, analysis, and vulnerability discovery. However, existing FSM extraction techniques face limitations such as scalability, incomplete coverage, and ambiguity in natural language specifications. In this paper, we propose FlowFSM, a novel agentic framework that leverages Large Language Models (LLMs) combined with prompt chaining and chain-of-thought reasoning to extract accurate FSMs from raw RFC documents. FlowFSM systematically processes protocol specifications, identifies state transitions, and constructs structured rule-books by chaining agent outputs. Experimental evaluation across FTP and RTSP protocols demonstrates that FlowFSM achieves high extraction precision while minimizing hallucinated transitions, showing promising results. Our findings highlight the potential of agent-based LLM systems in the advancement of protocol analysis and FSM inference for cybersecurity and reverse engineering applications.
Authors: Fares Wael, Youssef Maklad, Ali Hamdi, Wael Elsersy
Categories: cs.CL, cs.AI, cs.NI
PDF URL: https://arxiv.org/pdf/2507.11222v1.pdf
Published: 2025-07-15T11:50:25Z
5. Role-Playing LLM-Based Multi-Agent Support Framework for Detecting and Addressing Family Communication Bias
Well-being in family settings involves subtle psychological dynamics that conventional metrics often overlook. In particular, unconscious parental expectations, termed ideal parent bias, can suppress children’s emotional expression and autonomy. This suppression, referred to as suppressed emotion, often stems from well-meaning but value-driven communication, which is difficult to detect or address from outside the family. Focusing on these latent dynamics, this study explores Large Language Model (LLM)-based support for psychologically safe family communication. We constructed a Japanese parent-child dialogue corpus of 30 scenarios, each annotated with metadata on ideal parent bias and suppressed emotion. Based on this corpus, we developed a Role-Playing LLM-based multi-agent dialogue support framework that analyzes dialogue and generates feedback. Specialized agents detect suppressed emotion, describe implicit ideal parent bias in parental speech, and infer contextual attributes such as the child’s age and background. A meta-agent compiles these outputs into a structured report, which is then passed to five selected expert agents. These agents collaboratively generate empathetic and actionable feedback through a structured four-step discussion process. Experiments show that the system can detect categories of suppressed emotion with moderate accuracy and produce feedback rated highly in empathy and practicality. Moreover, simulated follow-up dialogues incorporating this feedback exhibited signs of improved emotional expression and mutual understanding, suggesting the framework’s potential in supporting positive transformation in family interactions.
Authors: Rushia Harada, Yuken Kimura, Keito Inoshita
Categories: cs.HC, cs.AI
PDF URL: https://arxiv.org/pdf/2507.11210v1.pdf
Published: 2025-07-15T11:27:32Z
6. Temperature and Persona Shape LLM Agent Consensus With Minimal Accuracy Gains in Qualitative Coding
Large Language Models (LLMs) enable new possibilities for qualitative research at scale, including coding and data annotation. While multi-agent systems (MAS) can emulate human coding workflows, their benefits over single-agent coding remain poorly understood. We conducted an experimental study of how agent persona and temperature shape consensus-building and coding accuracy of dialog segments based on a codebook with 8 codes. Our open-source MAS mirrors deductive human coding through structured agent discussion and consensus arbitration. Using six open-source LLMs (with 3 to 32 billion parameters) and 18 experimental configurations, we analyze over 77,000 coding decisions against a gold-standard dataset of human-annotated transcripts from online math tutoring sessions. Temperature significantly impacted whether and when consensus was reached across all six LLMs. MAS with multiple personas (including neutral, assertive, or empathetic), significantly delayed consensus in four out of six LLMs compared to uniform personas. In three of those LLMs, higher temperatures significantly diminished the effects of multiple personas on consensus. However, neither temperature nor persona pairing lead to robust improvements in coding accuracy. Single agents matched or outperformed MAS consensus in most conditions. Only one model (OpenHermesV2:7B) and code category showed above-chance gains from MAS deliberation when temperature was 0.5 or lower and especially when the agents included at least one assertive persona. Qualitative analysis of MAS collaboration for these configurations suggests that MAS may nonetheless aid in narrowing ambiguous code applications that could improve codebooks and human-AI coding. We contribute new insight into the limits of LLM-based qualitative methods, challenging the notion that diverse MAS personas lead to better outcomes. We open-source our MAS and experimentation code.
Authors: Conrad Borchers, Bahar Shahrokhian, Francesco Balzan, Elham Tajik, Sreecharan Sankaranarayanan, Sebastian Simon
Categories: cs.CL, cs.AI
PDF URL: https://arxiv.org/pdf/2507.11198v1.pdf
Published: 2025-07-15T11:06:32Z
7. An Agentic Framework for Autonomous Metamaterial Modeling and Inverse Design
Recent significant advances in integrating multiple Large Language Model (LLM) systems have enabled Agentic Frameworks capable of performing complex tasks autonomously, including novel scientific research. We develop and demonstrate such a framework specifically for the inverse design of photonic metamaterials. When queried with a desired optical spectrum, the Agent autonomously proposes and develops a forward deep learning model, accesses external tools via APIs for tasks like simulation and optimization, utilizes memory, and generates a final design via a deep inverse method. The framework’s effectiveness is demonstrated in its ability to automate, reason, plan, and adapt. Notably, the Agentic Framework possesses internal reflection and decision flexibility, permitting highly varied and potentially novel outputs.
Authors: Darui Lu, Jordan M. Malof, Willie J. Padilla
Categories: cs.AI, cond-mat.mtrl-sci
PDF URL: https://arxiv.org/pdf/2506.06935v2.pdf
Published: 2025-06-07T22:10:05Z
8. Collaborative Trustworthiness for Good Decision Making in Autonomous Systems
Autonomous systems are becoming an integral part of many application domains, like in the mobility sector. However, ensuring their safe and correct behaviour in dynamic and complex environments remains a significant challenge, where systems should autonomously make decisions e.g., about manoeuvring. We propose in this paper a general collaborative approach for increasing the level of trustworthiness in the environment of operation and improve reliability and good decision making in autonomous system. In the presence of conflicting information, aggregation becomes a major issue for trustworthy decision making based on collaborative data sharing. Unlike classical approaches in the literature that rely on consensus or majority as aggregation rule, we exploit the fact that autonomous systems have different quality attributes like perception quality. We use this criteria to determine which autonomous systems are trustworthy and borrow concepts from social epistemology to define aggregation and propagation rules, used for automated decision making. We use Binary Decision Diagrams (BDDs) as formal models for beliefs aggregation and propagation, and formulate reduction rules to reduce the size of the BDDs and allow efficient computation structures for collaborative automated reasoning.
Authors: Selma Saidi, Omar Laimona, Christoph Schmickler, Dirk Ziegenbein
Categories: cs.AI
PDF URL: https://arxiv.org/pdf/2507.11135v1.pdf
Published: 2025-07-15T09:37:28Z
9. Plancraft: an evaluation dataset for planning with LLM agents
We present Plancraft, a multi-modal evaluation dataset for LLM agents. Plancraft has both a text-only and multi-modal interface, based on the Minecraft crafting GUI. We include the Minecraft Wiki to evaluate tool use and Retrieval Augmented Generation (RAG), as well as a handcrafted planner and Oracle Retriever, to ablate the different components of a modern agent architecture. To evaluate decision-making, Plancraft also includes a subset of examples that are intentionally unsolvable, providing a realistic challenge that requires the agent not only to complete tasks but also to decide whether they are solvable at all. We benchmark both open-source and closed-source LLMs and compare their performance and efficiency to a handcrafted planner. Overall, we find that LLMs and VLMs struggle with the planning problems that Plancraft introduces, and offer suggestions on how to improve their capabilities.
Authors: Gautier Dagan, Frank Keller, Alex Lascarides
Categories: cs.CL, cs.AI
PDF URL: https://arxiv.org/pdf/2412.21033v2.pdf
Published: 2024-12-30T15:58:41Z
10. AI Agent Architecture for Decentralized Trading of Alternative Assets
Decentralized trading of real-world alternative assets (e.g., gold) requires bridging physical asset custody with blockchain systems while meeting strict requirements for compliance, liquidity, and risk management. We present GoldMine OS, a research oriented architecture that employs multiple specialized AI agents to automate and secure the tokenization and exchange of physical gold into a blockchain based stablecoin (“OZ”). Our approach combines on chain smart contracts for critical risk controls with off chain AI agents for decision making, blending the transparency and reliability of blockchains with the flexibility of AI driven automation. We describe four cooperative agents (Compliance, Token Issuance, Market Making, and Risk Control) and a coordinating core, and evaluate the system through simulation and a controlled pilot deployment. In experiments the prototype delivers on demand token issuance in under 1.2 s, more than 100 times faster than manual workflows. The Market Making agent maintains tight liquidity with spreads often below 0.5 percent even under volatile conditions. Fault injection tests show resilience: an oracle price spoofing attack is detected and mitigated within 10 s, and a simulated vault mis reporting halts issuance immediately with minimal user impact. The architecture scales to 5000 transactions per second with 10000 concurrent users in benchmarks. These results indicate that an AI agent based decentralized exchange for alternative assets can satisfy rigorous performance and safety requirements. We discuss broader implications for democratizing access to traditionally illiquid assets and explain how our governance model — multi signature agent updates and on chain community voting on risk parameters — provides ongoing transparency, adaptability, and formal assurance of system integrity.
Authors: Ailiya Borjigin, Cong He, Charles CC Lee, Wei Zhou
Categories: cs.AI
PDF URL: https://arxiv.org/pdf/2507.11117v1.pdf
Published: 2025-07-15T09:11:19Z
Evaluation Domain Papers
1. DrafterBench: Benchmarking Large Language Models for Tasks Automation in Civil Engineering
Large Language Model (LLM) agents have shown great potential for solving real-world problems and promise to be a solution for tasks automation in industry. However, more benchmarks are needed to systematically evaluate automation agents from an industrial perspective, for example, in Civil Engineering. Therefore, we propose DrafterBench for the comprehensive evaluation of LLM agents in the context of technical drawing revision, a representation task in civil engineering. DrafterBench contains twelve types of tasks summarized from real-world drawing files, with 46 customized functions/tools and 1920 tasks in total. DrafterBench is an open-source benchmark to rigorously test AI agents’ proficiency in interpreting intricate and long-context instructions, leveraging prior knowledge, and adapting to dynamic instruction quality via implicit policy awareness. The toolkit comprehensively assesses distinct capabilities in structured data comprehension, function execution, instruction following, and critical reasoning. DrafterBench offers detailed analysis of task accuracy and error statistics, aiming to provide deeper insight into agent capabilities and identify improvement targets for integrating LLMs in engineering applications. Our benchmark is available at https://github.com/Eason-Li-AIS/DrafterBench, with the test set hosted at https://huggingface.co/datasets/Eason666/DrafterBench.
Authors: Yinsheng Li, Zhen Dong, Yi Shao
Categories: cs.AI, cs.CE
PDF URL: https://arxiv.org/pdf/2507.11527v1.pdf
Published: 2025-07-15T17:56:04Z
2. LongDocURL: a Comprehensive Multimodal Long Document Benchmark Integrating Understanding, Reasoning, and Locating
Large vision language models (LVLMs) have improved the document understanding capabilities remarkably, enabling the handling of complex document elements, longer contexts, and a wider range of tasks. However, existing document understanding benchmarks have been limited to handling only a small number of pages and fail to provide a comprehensive analysis of layout elements locating. In this paper, we first define three primary task categories: Long Document Understanding, numerical Reasoning, and cross-element Locating, and then propose a comprehensive benchmark, LongDocURL, integrating above three primary tasks and comprising 20 sub-tasks categorized based on different primary tasks and answer evidences. Furthermore, we develop a semi-automated construction pipeline and collect 2,325 high-quality question-answering pairs, covering more than 33,000 pages of documents, significantly outperforming existing benchmarks. Subsequently, we conduct comprehensive evaluation experiments on both open-source and closed-source models across 26 different configurations, revealing critical performance gaps in this field.
Authors: Chao Deng, Jiale Yuan, Pi Bu, Peijie Wang, Zhong-Zhi Li, Jian Xu, Xiao-Hui Li, Yuan Gao, Jun Song, Bo Zheng, Cheng-Lin Liu
Categories: cs.AI, cs.CL
PDF URL: https://arxiv.org/pdf/2412.18424v3.pdf
Published: 2024-12-24T13:39:32Z
3. A Unified Framework for Evaluating the Effectiveness and Enhancing the Transparency of Explainable AI Methods in Real-World Applications
The fast growth of deep learning has brought great progress in AI-based applications. However, these models are often seen as “black boxes,” which makes them hard to understand, explain, or trust. Explainable Artificial Intelligence (XAI) tries to make AI decisions clearer so that people can understand how and why the model makes certain choices. Even though many studies have focused on XAI, there is still a lack of standard ways to measure how well these explanation methods work in real-world situations. This study introduces a single evaluation framework for XAI. It uses both numbers and user feedback to check if the explanations are correct, easy to understand, fair, complete, and reliable. The framework focuses on users’ needs and different application areas, which helps improve the trust and use of AI in important fields. To fix problems in current evaluation methods, we propose clear steps, including loading data, creating explanations, and fully testing them. We also suggest setting common benchmarks. We show the value of this framework through case studies in healthcare, finance, farming, and self-driving systems. These examples prove that our method can support fair and trustworthy evaluation of XAI methods. This work gives a clear and practical way to improve transparency and trust in AI systems used in the real world.
Authors: Md. Ariful Islam, Md Abrar Jahin, M. F. Mridha, Nilanjan Dey
Categories: cs.AI
PDF URL: https://arxiv.org/pdf/2412.03884v2.pdf
Published: 2024-12-05T05:30:10Z
4. ProtocolLLM: RTL Benchmark for SystemVerilog Generation of Communication Protocols
Recent advances in large language models (LLMs) have demonstrated strong performance in generating code for general-purpose programming languages. However, their potential for hardware description languages (HDLs), such as SystemVerilog, remains largely unexplored. HDL code generation poses unique challenges due to strict timing semantics, concurrency, and synthesizability constraints essential for correct hardware functionality. Further, HDL-based design flows encompass a broad set of tasks beyond structural code generation, including testbench development, assertion-based verification, timing closure, and protocol-level integration for on-chip communication. In this work, we evaluate the capabilities of both open-source and state-of-the-art LLMs in generating synthesizable and functionally accurate SystemVerilog implementations of widely used communication protocols that are critical components of embedded and System-on-Chip (SoC) systems. We introduce ProtocolLLM, the first benchmark suite specifically targeting these protocols with tasks spanning multiple design abstraction levels and varying prompt specificity. Our evaluation method also focuses on timing correctness in addition to synthesizability and syntactic correctness. We observe that most of the models fail to generate SystemVerilog code for communication protocols that follow timing constrains.
Authors: Arnav Sheth, Ivaxi Sheth, Mario Fritz
Categories: cs.AR, cs.AI, cs.CL
PDF URL: https://arxiv.org/pdf/2506.07945v2.pdf
Published: 2025-06-09T17:10:47Z
5. Assistance or Disruption? Exploring and Evaluating the Design and Trade-offs of Proactive AI Programming Support
AI programming tools enable powerful code generation, and recent prototypes attempt to reduce user effort with proactive AI agents, but their impact on programming workflows remains unexplored. We introduce and evaluate Codellaborator, a design probe LLM agent that initiates programming assistance based on editor activities and task context. We explored three interface variants to assess trade-offs between increasingly salient AI support: prompt-only, proactive agent, and proactive agent with presence and context (Codellaborator). In a within-subject study (N=18), we find that proactive agents increase efficiency compared to prompt-only paradigm, but also incur workflow disruptions. However, presence indicators and interaction context support alleviated disruptions and improved users’ awareness of AI processes. We underscore trade-offs of Codellaborator on user control, ownership, and code understanding, emphasizing the need to adapt proactivity to programming processes. Our research contributes to the design exploration and evaluation of proactive AI systems, presenting design implications on AI-integrated programming workflow.
Authors: Kevin Pu, Daniel Lazaro, Ian Arawjo, Haijun Xia, Ziang Xiao, Tovi Grossman, Yan Chen
Categories: cs.HC, cs.AI, cs.SE
PDF URL: https://arxiv.org/pdf/2502.18658v3.pdf
Published: 2025-02-25T21:37:25Z
6. Automated Novelty Evaluation of Academic Paper: A Collaborative Approach Integrating Human and Large Language Model Knowledge
Novelty is a crucial criterion in the peer review process for evaluating academic papers. Traditionally, it’s judged by experts or measure by unique reference combinations. Both methods have limitations: experts have limited knowledge, and the effectiveness of the combination method is uncertain. Moreover, it’s unclear if unique citations truly measure novelty. The large language model (LLM) possesses a wealth of knowledge, while human experts possess judgment abilities that the LLM does not possess. Therefore, our research integrates the knowledge and abilities of LLM and human experts to address the limitations of novelty assessment. One of the most common types of novelty in academic papers is the introduction of new methods. In this paper, we propose leveraging human knowledge and LLM to assist pretrained language models (PLMs, e.g. BERT etc.) in predicting the method novelty of papers. Specifically, we extract sentences related to the novelty of the academic paper from peer review reports and use LLM to summarize the methodology section of the academic paper, which are then used to fine-tune PLMs. In addition, we have designed a text-guided fusion module with novel Sparse-Attention to better integrate human and LLM knowledge. We compared the method we proposed with a large number of baselines. Extensive experiments demonstrate that our method achieves superior performance.
Authors: Wenqing Wu, Chengzhi Zhang, Yi Zhao
Categories: cs.CL, cs.AI, cs.DL, cs.HC
PDF URL: https://arxiv.org/pdf/2507.11330v2.pdf
Published: 2025-07-15T14:03:55Z
7. FeDa4Fair: Client-Level Federated Datasets for Fairness Evaluation
Federated Learning (FL) enables collaborative model training across multiple clients without sharing clients’ private data. However, fairness remains a key concern, as biases in local clients’ datasets can impact the entire federated system. Heterogeneous data distributions across clients may lead to models that are fairer for some clients than others. Although several fairness-enhancing solutions are present in the literature, most focus on mitigating bias for a single sensitive attribute, typically binary, overlooking the diverse and sometimes conflicting fairness needs of different clients. This limited perspective can limit the effectiveness of fairness interventions for the different clients. To support more robust and reproducible fairness research in FL, we aim to enable a consistent benchmarking of fairness-aware FL methods at both the global and client levels. In this paper, we contribute in three ways: (1) We introduce FeDa4Fair, a library to generate tabular datasets tailored to evaluating fair FL methods under heterogeneous client bias; (2) we release four bias-heterogeneous datasets and corresponding benchmarks to compare fairness mitigation methods in a controlled environment; (3) we provide ready-to-use functions for evaluating fairness outcomes for these datasets.
Authors: Xenia Heilmann, Luca Corbucci, Mattia Cerrato, Anna Monreale
Categories: cs.LG, cs.AI
PDF URL: https://arxiv.org/pdf/2506.21095v2.pdf
Published: 2025-06-26T08:43:12Z
8. ImpliRet: Benchmarking the Implicit Fact Retrieval Challenge
Retrieval systems are central to many NLP pipelines, but often rely on surface-level cues such as keyword overlap and lexical semantic similarity. To evaluate retrieval beyond these shallow signals, recent benchmarks introduce reasoning-heavy queries; however, they primarily shift the burden to query-side processing techniques — like prompting or multi-hop retrieval — that can help resolve complexity. In contrast, we present ImpliRet, a benchmark that shifts the reasoning challenge to document-side processing: The queries are simple, but relevance depends on facts stated implicitly in documents through temporal (e.g., resolving “two days ago”), arithmetic, and world knowledge relationships. We evaluate a range of sparse and dense retrievers, all of which struggle in this setting: the best nDCG@10 is only 14.91%. We also test whether long-context models can overcome this limitation. But even with a short context of only thirty documents, including the positive document, GPT-o4-mini scores only 55.54%, showing that document-side reasoning remains a challenge. Our codes are available at: github.com/ZeinabTaghavi/IMPLIRET
Authors: Zeinab Sadat Taghavi, Ali Modarressi, Yunpu Ma, Hinrich Schütze
Categories: cs.CL, cs.AI
PDF URL: https://arxiv.org/pdf/2506.14407v2.pdf
Published: 2025-06-17T11:08:29Z
9. An Explainable AI-Enhanced Machine Learning Approach for Cardiovascular Disease Detection and Risk Assessment
Heart disease remains a major global health concern, particularly in regions with limited access to medical resources and diagnostic facilities. Traditional diagnostic methods often fail to accurately identify and manage heart disease risks, leading to adverse outcomes. Machine learning has the potential to significantly enhance the accuracy, efficiency, and speed of heart disease diagnosis. In this study, we proposed a comprehensive framework that combines classification models for heart disease detection and regression models for risk prediction. We employed the Heart Disease dataset, which comprises 1,035 cases. To address the issue of class imbalance, the Synthetic Minority Oversampling Technique (SMOTE) was applied, resulting in the generation of an additional 100,000 synthetic data points. Performance metrics, including accuracy, precision, recall, F1-score, R2, MSE, RMSE, and MAE, were used to evaluate the model’s effectiveness. Among the classification models, Random Forest emerged as the standout performer, achieving an accuracy of 97.2% on real data and 97.6% on synthetic data. For regression tasks, Linear Regression demonstrated the highest R2 values of 0.992 and 0.984 on real and synthetic datasets, respectively, with the lowest error metrics. Additionally, Explainable AI techniques were employed to enhance the interpretability of the models. This study highlights the potential of machine learning to revolutionize heart disease diagnosis and risk prediction, thereby facilitating early intervention and enhancing clinical decision-making.
Authors: Md. Emon Akter Sourov, Md. Sabbir Hossen, Pabon Shaha, Mohammad Minoar Hossain, Md Sadiq Iqbal
Categories: cs.LG, cs.AI
PDF URL: https://arxiv.org/pdf/2507.11185v1.pdf
Published: 2025-07-15T10:38:38Z
10. Assessing Color Vision Test in Large Vision-language Models
With the widespread adoption of large vision-language models, the capacity for color vision in these models is crucial. However, the color vision abilities of large visual-language models have not yet been thoroughly explored. To address this gap, we define a color vision testing task for large vision-language models and construct a dataset \footnote{Anonymous Github Showing some of the data https://anonymous.4open.science/r/color-vision-test-dataset-3BCD} that covers multiple categories of test questions and tasks of varying difficulty levels. Furthermore, we analyze the types of errors made by large vision-language models and propose fine-tuning strategies to enhance their performance in color vision tests.
Authors: Hongfei Ye, Bin Chen, Wenxi Liu, Yu Zhang, Zhao Li, Dandan Ni, Hongyang Chen
Categories: cs.CV, cs.AI
PDF URL: https://arxiv.org/pdf/2507.11153v1.pdf
Published: 2025-07-15T10:03:06Z
AI Domain Papers
1. Streaming 4D Visual Geometry Transformer
Perceiving and reconstructing 4D spatial-temporal geometry from videos is a fundamental yet challenging computer vision task. To facilitate interactive and real-time applications, we propose a streaming 4D visual geometry transformer that shares a similar philosophy with autoregressive large language models. We explore a simple and efficient design and employ a causal transformer architecture to process the input sequence in an online manner. We use temporal causal attention and cache the historical keys and values as implicit memory to enable efficient streaming long-term 4D reconstruction. This design can handle real-time 4D reconstruction by incrementally integrating historical information while maintaining high-quality spatial consistency. For efficient training, we propose to distill knowledge from the dense bidirectional visual geometry grounded transformer (VGGT) to our causal model. For inference, our model supports the migration of optimized efficient attention operator (e.g., FlashAttention) from the field of large language models. Extensive experiments on various 4D geometry perception benchmarks demonstrate that our model increases the inference speed in online scenarios while maintaining competitive performance, paving the way for scalable and interactive 4D vision systems. Code is available at: https://github.com/wzzheng/StreamVGGT.
Authors: Dong Zhuo, Wenzhao Zheng, Jiahe Guo, Yuqi Wu, Jie Zhou, Jiwen Lu
Categories: cs.CV, cs.AI, cs.LG
PDF URL: https://arxiv.org/pdf/2507.11539v1.pdf
Published: 2025-07-15T17:59:57Z
2. How Many Instructions Can LLMs Follow at Once?
Production-grade LLM systems require robust adherence to dozens or even hundreds of instructions simultaneously. However, the instruction-following capabilities of LLMs at high instruction densities have not yet been characterized, as existing benchmarks only evaluate models on tasks with a single or few instructions. We introduce IFScale, a simple benchmark of 500 keyword-inclusion instructions for a business report writing task to measure how instruction-following performance degrades as instruction density increases. We evaluate 20 state-of-the-art models across seven major providers and find that even the best frontier models only achieve 68% accuracy at the max density of 500 instructions. Our analysis reveals model size and reasoning capability to correlate with 3 distinct performance degradation patterns, bias towards earlier instructions, and distinct categories of instruction-following errors. Our insights can help inform design of instruction-dense prompts in real-world applications and highlight important performance-latency tradeoffs. We open-source the benchmark and all results for further analysis at https://distylai.github.io/IFScale.
Authors: Daniel Jaroslawicz, Brendan Whiting, Parth Shah, Karime Maamari
Categories: cs.AI
PDF URL: https://arxiv.org/pdf/2507.11538v1.pdf
Published: 2025-07-15T17:59:42Z
3. Canonical Bayesian Linear System Identification
Standard Bayesian approaches for linear time-invariant (LTI) system identification are hindered by parameter non-identifiability; the resulting complex, multi-modal posteriors make inference inefficient and impractical. We solve this problem by embedding canonical forms of LTI systems within the Bayesian framework. We rigorously establish that inference in these minimal parameterizations fully captures all invariant system dynamics (e.g., transfer functions, eigenvalues, predictive distributions of system outputs) while resolving identifiability. This approach unlocks the use of meaningful, structure-aware priors (e.g., enforcing stability via eigenvalues) and ensures conditions for a Bernstein—von Mises theorem — a link between Bayesian and frequentist large-sample asymptotics that is broken in standard forms. Extensive simulations with modern MCMC methods highlight advantages over standard parameterizations: canonical forms achieve higher computational efficiency, generate interpretable and well-behaved posteriors, and provide robust uncertainty estimates, particularly from limited data.
Authors: Andrey Bryutkin, Matthew E. Levine, Iñigo Urteaga, Youssef Marzouk
Categories: stat.ML, cs.LG, cs.SY, eess.SY, stat.CO
PDF URL: https://arxiv.org/pdf/2507.11535v1.pdf
Published: 2025-07-15T17:58:55Z
4. Langevin Flows for Modeling Neural Latent Dynamics
Neural populations exhibit latent dynamical structures that drive time-evolving spiking activities, motivating the search for models that capture both intrinsic network dynamics and external unobserved influences. In this work, we introduce LangevinFlow, a sequential Variational Auto-Encoder where the time evolution of latent variables is governed by the underdamped Langevin equation. Our approach incorporates physical priors — such as inertia, damping, a learned potential function, and stochastic forces — to represent both autonomous and non-autonomous processes in neural systems. Crucially, the potential function is parameterized as a network of locally coupled oscillators, biasing the model toward oscillatory and flow-like behaviors observed in biological neural populations. Our model features a recurrent encoder, a one-layer Transformer decoder, and Langevin dynamics in the latent space. Empirically, our method outperforms state-of-the-art baselines on synthetic neural populations generated by a Lorenz attractor, closely matching ground-truth firing rates. On the Neural Latents Benchmark (NLB), the model achieves superior held-out neuron likelihoods (bits per spike) and forward prediction accuracy across four challenging datasets. It also matches or surpasses alternative methods in decoding behavioral metrics such as hand velocity. Overall, this work introduces a flexible, physics-inspired, high-performing framework for modeling complex neural population dynamics and their unobserved influences.
Authors: Yue Song, T. Anderson Keller, Yisong Yue, Pietro Perona, Max Welling
Categories: cs.LG, q-bio.NC
PDF URL: https://arxiv.org/pdf/2507.11531v1.pdf
Published: 2025-07-15T17:57:48Z
5. DrafterBench: Benchmarking Large Language Models for Tasks Automation in Civil Engineering
Large Language Model (LLM) agents have shown great potential for solving real-world problems and promise to be a solution for tasks automation in industry. However, more benchmarks are needed to systematically evaluate automation agents from an industrial perspective, for example, in Civil Engineering. Therefore, we propose DrafterBench for the comprehensive evaluation of LLM agents in the context of technical drawing revision, a representation task in civil engineering. DrafterBench contains twelve types of tasks summarized from real-world drawing files, with 46 customized functions/tools and 1920 tasks in total. DrafterBench is an open-source benchmark to rigorously test AI agents’ proficiency in interpreting intricate and long-context instructions, leveraging prior knowledge, and adapting to dynamic instruction quality via implicit policy awareness. The toolkit comprehensively assesses distinct capabilities in structured data comprehension, function execution, instruction following, and critical reasoning. DrafterBench offers detailed analysis of task accuracy and error statistics, aiming to provide deeper insight into agent capabilities and identify improvement targets for integrating LLMs in engineering applications. Our benchmark is available at https://github.com/Eason-Li-AIS/DrafterBench, with the test set hosted at https://huggingface.co/datasets/Eason666/DrafterBench.
Authors: Yinsheng Li, Zhen Dong, Yi Shao
Categories: cs.AI, cs.CE
PDF URL: https://arxiv.org/pdf/2507.11527v1.pdf
Published: 2025-07-15T17:56:04Z
6. LongDocURL: a Comprehensive Multimodal Long Document Benchmark Integrating Understanding, Reasoning, and Locating
Large vision language models (LVLMs) have improved the document understanding capabilities remarkably, enabling the handling of complex document elements, longer contexts, and a wider range of tasks. However, existing document understanding benchmarks have been limited to handling only a small number of pages and fail to provide a comprehensive analysis of layout elements locating. In this paper, we first define three primary task categories: Long Document Understanding, numerical Reasoning, and cross-element Locating, and then propose a comprehensive benchmark, LongDocURL, integrating above three primary tasks and comprising 20 sub-tasks categorized based on different primary tasks and answer evidences. Furthermore, we develop a semi-automated construction pipeline and collect 2,325 high-quality question-answering pairs, covering more than 33,000 pages of documents, significantly outperforming existing benchmarks. Subsequently, we conduct comprehensive evaluation experiments on both open-source and closed-source models across 26 different configurations, revealing critical performance gaps in this field.
Authors: Chao Deng, Jiale Yuan, Pi Bu, Peijie Wang, Zhong-Zhi Li, Jian Xu, Xiao-Hui Li, Yuan Gao, Jun Song, Bo Zheng, Cheng-Lin Liu
Categories: cs.AI, cs.CL
PDF URL: https://arxiv.org/pdf/2412.18424v3.pdf
Published: 2024-12-24T13:39:32Z
7. EXPO: Stable Reinforcement Learning with Expressive Policies
We study the problem of training and fine-tuning expressive policies with online reinforcement learning (RL) given an offline dataset. Training expressive policy classes with online RL present a unique challenge of stable value maximization. Unlike simpler Gaussian policies commonly used in online RL, expressive policies like diffusion and flow-matching policies are parameterized by a long denoising chain, which hinders stable gradient propagation from actions to policy parameters when optimizing against some value function. Our key insight is that we can address stable value maximization by avoiding direct optimization over value with the expressive policy and instead construct an on-the-fly RL policy to maximize Q-value. We propose Expressive Policy Optimization (EXPO), a sample-efficient online RL algorithm that utilizes an on-the-fly policy to maximize value with two parameterized policies — a larger expressive base policy trained with a stable imitation learning objective and a light-weight Gaussian edit policy that edits the actions sampled from the base policy toward a higher value distribution. The on-the-fly policy optimizes the actions from the base policy with the learned edit policy and chooses the value maximizing action from the base and edited actions for both sampling and temporal-difference (TD) backup. Our approach yields up to 2-3x improvement in sample efficiency on average over prior methods both in the setting of fine-tuning a pretrained policy given offline data and in leveraging offline data to train online.
Authors: Perry Dong, Qiyang Li, Dorsa Sadigh, Chelsea Finn
Categories: cs.LG, cs.AI
PDF URL: https://arxiv.org/pdf/2507.07986v2.pdf
Published: 2025-07-10T17:57:46Z
8. CATVis: Context-Aware Thought Visualization
EEG-based brain-computer interfaces (BCIs) have shown promise in various applications, such as motor imagery and cognitive state monitoring. However, decoding visual representations from EEG signals remains a significant challenge due to their complex and noisy nature. We thus propose a novel 5-stage framework for decoding visual representations from EEG signals: (1) an EEG encoder for concept classification, (2) cross-modal alignment of EEG and text embeddings in CLIP feature space, (3) caption refinement via re-ranking, (4) weighted interpolation of concept and caption embeddings for richer semantics, and (5) image generation using a pre-trained Stable Diffusion model. We enable context-aware EEG-to-image generation through cross-modal alignment and re-ranking. Experimental results demonstrate that our method generates high-quality images aligned with visual stimuli, outperforming SOTA approaches by 13.43% in Classification Accuracy, 15.21% in Generation Accuracy and reducing Fr\’echet Inception Distance by 36.61%, indicating superior semantic alignment and image quality.
Authors: Tariq Mehmood, Hamza Ahmad, Muhammad Haroon Shakeel, Murtaza Taj
Categories: cs.CV, cs.LG
PDF URL: https://arxiv.org/pdf/2507.11522v1.pdf
Published: 2025-07-15T17:47:01Z
9. Hi Robot: Open-Ended Instruction Following with Hierarchical Vision-Language-Action Models
Generalist robots that can perform a range of different tasks in open-world settings must be able to not only reason about the steps needed to accomplish their goals, but also process complex instructions, prompts, and even feedback during task execution. Intricate instructions (e.g., “Could you make me a vegetarian sandwich?” or “I don’t like that one”) require not just the ability to physically perform the individual steps, but the ability to situate complex commands and feedback in the physical world. In this work, we describe a system that uses vision-language models in a hierarchical structure, first reasoning over complex prompts and user feedback to deduce the most appropriate next step to fulfill the task, and then performing that step with low-level actions. In contrast to direct instruction following methods that can fulfill simple commands (“pick up the cup”), our system can reason through complex prompts and incorporate situated feedback during task execution (“that’s not trash”). We evaluate our system across three robotic platforms, including single-arm, dual-arm, and dual-arm mobile robots, demonstrating its ability to handle tasks such as cleaning messy tables, making sandwiches, and grocery shopping. Videos are available at https://www.pi.website/research/hirobot
Authors: Lucy Xiaoyang Shi, Brian Ichter, Michael Equi, Liyiming Ke, Karl Pertsch, Quan Vuong, James Tanner, Anna Walling, Haohuan Wang, Niccolo Fusai, Adrian Li-Bell, Danny Driess, Lachy Groom, Sergey Levine, Chelsea Finn
Categories: cs.RO, cs.AI, cs.LG
PDF URL: https://arxiv.org/pdf/2502.19417v2.pdf
Published: 2025-02-26T18:58:41Z
10. AirLLM: Diffusion Policy-based Adaptive LoRA for Remote Fine-Tuning of LLM over the Air
Operating Large Language Models (LLMs) on edge devices is increasingly challenged by limited communication bandwidth and strained computational and memory costs. Thus, cloud-assisted remote fine-tuning becomes indispensable. Nevertheless, existing Low-Rank Adaptation (LoRA) approaches typically employ fixed or heuristic rank configurations, and the subsequent over-the-air transmission of all LoRA parameters could be rather inefficient. To address this limitation, we develop AirLLM, a hierarchical diffusion policy framework for communication-aware LoRA adaptation. Specifically, AirLLM models the rank configuration as a structured action vector that spans all LoRA-inserted projections. To solve the underlying high-dimensional sequential decision-making problem, a Proximal Policy Optimization (PPO) agent generates coarse-grained decisions by jointly observing wireless states and linguistic complexity, which are then refined via Denoising Diffusion Implicit Models (DDIM) to produce high-resolution, task- and channel-adaptive rank vectors. The two modules are optimized alternatively, with the DDIM trained under the Classifier-Free Guidance (CFG) paradigm to maintain alignment with PPO rewards. Experiments under varying signal-to-noise ratios demonstrate that AirLLM consistently enhances fine-tuning performance while significantly reducing transmission costs, highlighting the effectiveness of reinforcement-driven, diffusion-refined rank adaptation for scalable and efficient remote fine-tuning over the air.
Authors: Shiyi Yang, Xiaoxue Yu, Rongpeng Li, Jianhang Zhu, Zhifeng Zhao, Honggang Zhang
Categories: cs.LG, cs.AI, cs.CL
PDF URL: https://arxiv.org/pdf/2507.11515v1.pdf
Published: 2025-07-15T17:36:37Z