Last Updated: 2026-08-11 03:09:51
Total Papers: 85
- Agent Applications
- Benchmarks and Datasets
- LLM Agents
- Multi-Agent Systems
- Other Agent Research
- Planning and Reasoning
- Reinforcement Learning
IB-RL: Isolated Bilateral Reinforcement Learning for Strategic Dialogue Agents - Senhao Wang, Chenghao Cai, Haitao Hu, Mingxing Huang, Xingguang Wang, Wenhao Li, Zecheng Lin - [[pdf]](https://arxiv.org/pdf/2608.06735)
Abstract: Reinforcement learning (RL) has achieved strong results in improving large language models (LLMs) on tasks with stationary, verifiable rewards, such as mathematical reasoning and code execution. In these settings, the environment follows fixed rules and does not adapt strategically to the agent. Strategic dialogue differs in this respect: the environment is another agent that adapts to the policy, and success depends on the interaction between the two sides. Despite this interactive nature, current RL approaches typically train a target agent against a fixed counterpart or simulator. We find that this training paradigm encourages the policy to exploit counterpart-specific regularities rather than learn strategies that generalize across counterparts. We call this problem the static-counterpart mismatch, which we quantify directly in our experiments. To address it, we propose Isolated Bilateral Reinforcement Learning (IB-RL), in which the two roles coevolve through joint rollouts while each role optimizes its own reward through fully independent advantages, action masks, and update paths. We evaluate frozen policies against fully independent held-out counterparts in both domains. On Vehicle TeleSales, IB-RL achieves 89.6% Success@1, compared to 84.6% for the best unilateral RL baseline. On Deal-or-NoDeal, it reaches 98.4% agreement against DeepSeek V4 Pro, compared to 86.4% for the best unilateral baseline. These results indicate that jointly training both roles with strict peragent isolation produces policies that generalize more effectively to unseen counterparts.
arXiv ID: 2608.06735
FinanceHarness: Autonomous Financial Deep Research Framework - Yijia Xiao, Rujun Han, Yanfei Chen, Zifeng Wang, Ke Jiang, Zhongying CuiZhu, Vishy Tirumalashetty, Wei Wang, Burak Gokturk, Tomas Pfister, Chen-Yu Lee - [[pdf]](https://arxiv.org/pdf/2607.27853)
Abstract: Powered by advances in LLMs and autonomous agents, deep research has become one of the most widely adopted agentic products. However, most deep research systems write general-purpose reports, which are inadequate for financial deep research. Financial research demands specialized knowledge to analyze historical patterns and forecast upcoming events. Automating financial deep research therefore requires both a layered harness to drive the research agent and a verifiable, point-in-time benchmark that prevents leakage of future information. We present FinanceHarness, a harness that runs finance-oriented tools and practitioner-guided workflows, automating financial deep research end to end: environment and data construction, the agent execution loop, and reward modeling. We further propose FinanceGym, comprising thesis-driven research questions and rubrics that combine pre-cutoff and post-cutoff criteria. Professional expert validation yields an 82% pass rate. With the same open-weight backbone, FinanceHarness improves the overall rubric score from 25.3% to 32.4%, demonstrating the effectiveness of our specialized harness design. However, even pairing FinanceHarness with the most cutting edge LLM (e.g. Opus-5), the FinanceGym score is below 45%, showing that it is a challenging benchmark for financial deep research. Leaderboard is available at: this https URL and FinanceHarness code is available at: this https URL.
arXiv ID: 2607.27853
Benchmarks and Datasets (15 papers)
Shape Your Feed: An LLM-based Agentic System for Conversational Recommendation - Ziyun Xu, Bosen Ding, Yue Zhang, Ji Qi, Qingyuan Song, Jizhou Huang, Liwei Wang, Jefferey Santelli, Yue Weng, Qichao Que, Zhenheng Yang, Junfeng Pan, Linhong Zhu - [[pdf]](https://arxiv.org/pdf/2608.06632)
Abstract: Industrial recommendation systems predominantly adopt a passive ranking paradigm that infers user preferences from implicit behavioral signals (e.g., clicks, dwell time) rather than explicit, natural language inputs. As a result, users experience a persistent discrepancy between their explicit interests and what passive behavioral algorithms deliver, limiting their ability to express nuanced preferences or steer their feed in real time. To address this growing gap between how recommendations are optimized and how users wish to articulate their interests, we present Shape Your Feed (SYF), an LLM-based agentic recommendation framework that enables real-time, multimodal co-curation of content. SYF employs a three-tier architecture: (i) a Perception Flow that captures fine-grained user intent from text prompts, voice commands, and UI interactions; (ii) a Serving Flow that performs real-time agentic re-ranking and pruning of candidate items, grounded in a persistent Semantic Profile encoding evolving user preferences; and (iii) a Self-Evolution Flow that aligns system behavior with human judgments via Direct Preference Optimization (DPO) and an LLM-as-a-Judge ensemble. Offline evaluations show that SYF's alignment scoring module achieves 98.85% accuracy, substantially improving over strong few-shot baselines. Large-scale online A/B experiments on production traffic further demonstrate that SYF improves feed relevance and user sentiment, indicating a practical and scalable path toward interactive, user-steerable recommendation in industrial settings.
arXiv ID: 2608.06632
AgentPatch: Coarse-to-Fine Weak-Task Repair for Merging Agentic Multimodal Large Language Models - Zibo Shao, Baochen Xiong, Chengdong Xu, Linhui Xiao, Kaichen Li, Haoran Gong, Yan Li, Yaguang Song, Xiaoshan Yang - [[pdf]](https://arxiv.org/pdf/2608.06699)
Abstract: Agentic multimodal large language models (MLLMs) extend multimodal perception and reasoning with planning, tool use, and interaction in dynamic environments. Yet current models are specialized for particular tools or environments, complicating consolidation into a single generalist. We formulate Agentic MLLM Merging and identify two challenges: asymmetric capability preservation, whereby capabilities with different interaction complexity are retained unevenly, producing weak tasks after merging, and behavior-critical forgetting, whereby losing decisive actions can derail long-horizon execution. We propose AgentPatch, a training-free coarse-to-fine repair framework. It selects a stable merged backbone, restores diluted weak-task-specific signals through Weak-Task Unique Residual Recovery, and applies an Agent-Guided Behavior-Critical Patch that recovers decisive behaviors under explicit capability protection. AgentPatch produces a single static checkpoint without routing or ensembles. Experiments across six agentic and multimodal benchmarks show that AgentPatch improves diverse merged backbones, alleviates weak-task degradation, and better balances weak-task recovery with the preservation of complementary search and agentic visual processing capabilities. Code is available at this https URL.
arXiv ID: 2608.06699
MemPrism: Task-Conditioned Relational Memory Views for Long-Horizon Agents - Zhisheng Chen, Bingfan Zeng, Bangde Cao, Zhengwei Xie, Yuxuan Li, Jinhan Li, Zheng Lu, Xiangchen Guan, Zikai Xiao, Rui Qian, Jingwei Song - [[pdf]](https://arxiv.org/pdf/2608.06745)
Abstract: Long-horizon agents rely on memory to reuse experiences, yet existing memory systems often assume that evidence can be directly consumed through a fixed representation. This leads to representation mismatch, where relevant information is available but not organized for the current decision. To this end, we propose MemPrism, a task-conditioned relational memory framework that separates persistent experience storage from decision-time working memory. MemPrism records interactions as the event stream and dynamically constructs relational views according to the current task context. A lightweight view policy selects the relation structure, evidence range, outcome condition, and granularity, while a deterministic composer and render transform historical facts into a temporary optical working-memory view for a frozen task policy. Experiments on long-horizon embodied and web-agent benchmarks show that MemPrism consistently improves the task performance, especially as trajectories become longer, while reducing memory token consumption. Furthermore, the learned view policy transfers across different VLMs without additional adaptation, demonstrating the effectiveness of task-conditioned relational views as a general memory interface for agents.
arXiv ID: 2608.06745
SkillEval: Decomposing Agent Skill Quality into Interpretable Signals - Jiahui Han, Qinuo Li, Ziheng Peng, Haotian Wu, Haoze Liu, Danfeng Shan, Guanchu Wang, Huiqi Deng, Ninghao Liu - [[pdf]](https://arxiv.org/pdf/2608.06891)
Abstract: Agent skills provide reusable procedural knowledge that helps agents solve specialized tasks. As their use expands, evaluating skill quality becomes increasingly important. Existing evaluations often measure skill quality by testing whether a skill improves performance on specific downstream tasks. However, a reusable skill may apply to multiple task scenarios. Downstream evaluation mainly reflects the compatibility between a skill and the evaluated task, provides only a partial view of skill quality, and does not identify which aspect of the skill should be improved. We find that general properties of the \texttt{this http URL} document play an important role in skill quality. To evaluate these properties, we propose \textbf{SkillEval}, an interpretable framework for document-level skill evaluation. SkillEval evaluates each property using a fixed and inspectable scoring direction, producing interpretable scores. It further measures and reduces the influence of unrelated document features, such as length and formatting, so that each score captures its intended semantic property more specifically. Specifically, SkillEval learns an interpretable direction for each quality property from controlled positive--negative skill pairs in the hidden representation space of the model, and scores a new skill by projecting its representation onto these fixed directions. We use SkillEval to evaluate skills in controlled quality tests and show that SkillEval reliably distinguishes skills of different quality. In addition, SkillEval scores closely reflect downstream task performance, providing an early indication of whether a skill is likely to help an agent complete a task. We further explore SkillEval for diagnosing weaknesses in skill documents and guiding targeted revisions. The revised skills improve the targeted properties and achieve higher pass rates on downstream tasks.
arXiv ID: 2608.06891
CAi Copilot: Reducing Operational Workload in Molecular Design through Intent-Driven Agentic Workflows - Zhu Wang, Jiangyu Chen, Yingjun Shang, Yuhui Yao, Laiao Lu, Tianfan Fu, Na Zou - [[pdf]](https://arxiv.org/pdf/2608.06961)
Abstract: Early-stage molecular design is an iterative process, not just a task of generating molecules. Researchers turn broad goals into design strategies, refine candidates, assess many properties, and gather evidence before synthesis and tests. AI methods can generate molecules, optimize several goals, predict properties, dock compounds, and account for synthesis. Yet these functions are spread across specialized tools. Experts must still coordinate each step, judge interim results, and integrate evidence. The central challenge is thus to turn research intent into adaptive, traceable runs grounded in scientific tools. We cast this challenge as intent-to-evidence molecular design workflow execution and present CAi Copilot, an expert-oriented agent with three linked layers. The Research Interface Layer turns intent into an executable plan. The Agent Reasoning Layer uses interim results to guide each run. The Execution Substrate supplies molecular tools, metrics, reusable utilities, and backend services. Across 45 tasks, CAi achieves the strongest overall performance, with an outcome score of 84.59, exceeding the next-best result by 18.07 points. Additional benchmarks test how CAi coordinates generation, screening, and multi-criteria evaluation, while exposing limits in long-horizon execution. These results show that CAi turns broad molecular-design intent into transparent, traceable workflows that connect interim decisions to candidate-level evidence.
arXiv ID: 2608.06961
An End-to-End Agent Auditing Engine - Haoning Wang, Mingxun Zhang, Chenyue Yu, Yingjun Shang, Xia Hu, Guanchu Wang, Na Zou - [[pdf]](https://arxiv.org/pdf/2608.07346)
Abstract: With the rapid advancement of large language models (LLMs), harnesses have become essential infrastructure for deploying agents across a wide range of domains. The fast-evolving harness ecosystem has also made rigorous capability evaluation increasingly important. However, efficiently building an end-to-end, systematic, and comprehensive evaluation pipeline remains a significant challenge. To address this challenge, we introduce
arXiv ID: 2608.07346
Agentic Planning for Symbolic Execution - Daniel Koh Ji Yang, Yannic Noller, Corina S. Pasareanu, Youcheng Sun - [[pdf]](https://arxiv.org/pdf/2608.06397)
Abstract: Symbolic execution seeks to explore feasible program paths, yet a practical run may exhaust its resources while much program behaviour remains unreached. We investigate a complementary way of extending its practical reach by reasoning about how the same tool is utilised from one bounded run to the next, while leaving ordinary state exploration to the underlying tool. We present Agolic, an agentic planning system that uses evidence from earlier runs to choose and configure later bounded symbolic execution (BSE) runs, which the underlying symbolic execution tool then carries out. The planning intelligence, available evidence and execution modes can be adapted to the symbolic execution tool and analysis objective. We evaluate one adaptation for branch-coverage exploration, in which an LLM-based agent reasons over source code, replayed coverage and earlier targeting attempts. We evaluate Agolic on several C and C++ programs. On every program, it extends the branch coverage obtained by continuous symbolic execution and covers more than
arXiv ID: 2608.06397
StepJack: Benchmarking Computer-Use Agent Safety Against Multi-Step Indirect Prompt Injection - Zhuoxin Zhan, Akbar Rafiey, Avery Ma, Leila Pishdad, Layla El Asri - [[pdf]](https://arxiv.org/pdf/2608.06477)
Abstract: Computer-use agents (CUAs) face a growing threat from indirect prompt injection, where adversarial instructions are planted in the environment such as web pages. In this paper, we introduce multi-step indirect prompt injection, a new attack class against CUAs in which the adversarial goal is decomposed into multiple innocuous-looking sub-steps and distributed across a chain of pages referenced along the agent's navigation path. We develop a pipeline to automatically decompose an adversarial goal under the constraint that the execution of the decomposed sub-steps must achieve the original goal while optimizing the innocuousness of each decomposed sub-step. With this pipeline, we build StepJack, a CUA safety benchmark with 480 test examples. On this benchmark, we evaluate six state-of-the-art CUAs and find that at a fixed decomposition depth, multi-step attacks raise attack success rate (ASR) on three of six CUAs, by up to 31.2 points (e.g., GPT-5.4-mini: 41.7% at single-step to 72.9% at three-step); averaged over the five CUAs that can reliably follow the reference chain (all but EvoCUA-32B), ASR rises from 31.3% at single-step to 36.9% at three-step. Dataset and code are available at this https URL.
arXiv ID: 2608.06477
HLSmith: An Expert-Guided Agentic Framework for C/C++-to-HLS Translation - Yuebo Luo, Ahmad Sedigh Baroughi, Philip Stachura, Le Chen, Venkatram Vishwanath, Zhenman Fang, Caiwen Ding - [[pdf]](https://arxiv.org/pdf/2608.06791)
Abstract: Application-specific FPGA accelerators offer substantial performance and energy-efficiency gains across many application domains, but developing them is costly, often requiring months of specialized effort. Even with high-level synthesis (HLS), designers still need extensive hardware expertise to build high-performance accelerators. Although large language models (LLMs) have demonstrated strong software-generation capabilities, even frontier models lack the hardware intuition and procedural knowledge needed to reliably translate baseline C/C++ programs into high-performance HLS designs: they struggle to identify effective architectures, follow the optimization processes used by HLS experts, and apply hardware transformations consistently across diverse kernels. We present HLSmith, an expert-guided framework for translating C/C++ programs into optimized HLS accelerators. HLSmith combines three components: an HLS optimization expertise library that encodes guarded transformation recipes, their applicability and prerequisite conditions, and unsafe cases to avoid; a staged, feedback-driven orchestration flow modeled on expert HLS development practice that guides agents through synthesis, bottleneck analysis, and optimization; and a tool-grounded model-adaptation pipeline that converts optimization trajectories from commercial frontier models into training data for fine-tuning open-weight LLMs. We evaluate HLSmith on PolyBench against ChatHLS, a leading prior agent-orchestration framework for HLS accelerator development. HLSmith achieves a geometric mean speedup of 4.24x over ChatHLS while producing functionally correct designs, in both software and RTL simulation, for every benchmark, compared with ChatHLS's 57% valid-design rate. It further reaches speedups of up to 252x and 138x with commercial frontier models and open-weight models, respectively.
arXiv ID: 2608.06791
HarnessSafe: Evaluating Safety Across Persistent Carriers in Agent Harnesses - Xiao Zhang, Yusheng Wang, Yuhao Fei, Dongyuan Li, Zian Liang, Liuyu Xiang, Hongxun Gu, Zhaofeng He - [[pdf]](https://arxiv.org/pdf/2608.06984)
Abstract: Modern agent harnesses persist state across tasks and sessions through persistent carriers like memory, skills, tools, and shared artifacts. However, this capability creates delayed safety risks: attacker-influenced content can cross system boundaries and later affect the execution of a benign request. Existing benchmarks typically focus on a few carriers or harnesses, while end-to-end attack-success rates reveal little about how risks propagate. To this end, we present HarnessSafe, a benchmark comprising 328 executable cases across seven persistent-carrier families and evaluated on most mainstream agent harnesses. Each case is specified as a Persistent-Risk Lifecycle that traces attacker influence from its initial entry, through persistence across carriers and system boundaries, to a later benign trigger and an observable violation. We further introduce a multi-stage, trace-based evaluation that uses observable execution evidence to determine how far each attack chain progresses and where it is stopped. Experiments show that containment is carrier-specific and strongly depends on the harness-model configuration. Both the harness and model backend substantially shape containment outcomes, while attack success rates cannot reflect distinct lifecycle progression patterns.
arXiv ID: 2608.06984
OpenForgeRL: Train Harness-native Agents in Any Environment - Xiao Yu, Baolin Peng, Ruize Xu, Hao Zou, Qianhui Wu, Hao Cheng, Wenlin Yao, Nikhil Singh, Zhou Yu, Jianfeng Gao - [[pdf]](https://arxiv.org/pdf/2607.21557)
Abstract: Modern AI agents rely on elaborate inference harnesses such as Claude Code, Codex, and OpenClaw to drive multi-turn reasoning, tool use, and access to external systems. While powerful, these complex harnesses also make agents hard to train end-to-end with open infrastructure, whose SFT/RL stacks cannot natively express stateful, multi-process harness inference. To address this, we present OpenForgeRL, an open-source framework for training harness-based agents end-to-end in diverse environments. OpenForgeRL achieves this with a lightweight proxy that serves the harness's model calls while recording them as training data for a standard RL codebase (e.g., veRL), and a Kubernetes orchestrator that runs each rollout in its own remote container, together enabling training on any harness in any environment at scale. By decoupling training and inference, OpenForgeRL allows researchers to easily train, study, and improve agents directly in the real harnesses and environments they are deployed with. We validate our framework across diverse, complex harnesses and environments, spanning tool/claw-based agents and multimodal GUI browser- and computer-use agents. Using only hundreds to a few thousand tasks, OpenForgeClaw reaches 31.7 pass^3 and 55.9 pass@3 on ClawEval and 33.7 on QwenClawBench. OpenForgeGUI reaches 37.7 on OSWorld-Verified, 63.0 on Online-Mind2Web, and 72.3 on WebVoyager. Both outperform open baselines of similar size on nearly all benchmarks, and in the GUI setting match or surpass models several times larger. Beyond benchmarks, we analyze how harness choice (e.g., ZeroClaw, OpenClaw, Codex) and RL shape agent behavior. We find that some harnesses are substantially harder to learn than others, and that RL improves agentic reliability, such as self-verification, tool coverage, and completing multi-step plans, though critical abilities such as error recovery remain weak.
arXiv ID: 2607.21557
Can AI agents conduct open-ended AI research? Early evidence from two case studies - Peter Kirgis, Sayash Kapoor, Andrew Schwartz, Stephan Rabanser, David Africa, Konstantinos Voudouris, Viet Nguyen, Toby Pilditch, Magda Dubois, Harry Coppock, Cozmin Ududec, Nitya Nadgir, Matilda Orona, Tilman Bayer, Derrick Chan-Sew, Yue Ling, Abhishek Shetty, Helen Toner, Gillian Hadfield, Seth Lazar, Steve Newman, Shoshannah Tekofsky, Rishi Bommasani, Arvind Narayanan - [[pdf]](https://arxiv.org/pdf/2607.27191)
Abstract: Forecasts of explosive AI progress hinge on AI agents automating AI research. But evidence on whether agents can carry out open-ended AI research is thin. Current evaluations either test agents on narrow, verifiable tasks, which excludes open-ended research, or submit AI-generated papers to blind peer review, which is overstretched, stochastic, and suffers from poor review quality. We introduce a third way to measure progress towards AI R&D automation. An agent takes on the central, open-ended research question of a high-quality unpublished paper, and the paper's original authors grade its output. We call these shadow evaluations. We ran shadow evaluations on two unpublished NeurIPS 2026 submissions, giving frontier agents six days and thousands of dollars of compute. The agents completed all of the engineering without human help, yet could not make substantial progress towards answering the research questions. As a result, both papers were unambiguously rejected by the authors. We identify five recurring failure modes: poor judgment about the bar for publishable research, uncreative responses to shortcomings in the research design, ineffective backtracking from dead ends, poor resource awareness, and instruction drift. A robustness check with a second model and scaffold reproduced these failures. We release the expert reviews, survey responses, agent repositories, and logs. Our results provide early evidence that today's agents can do the engineering of AI research, but struggle with critical parts of the research lifecycle.
arXiv ID: 2607.27191
From Plan to Action: How Well Do Agents Follow the Plan? - Shuyang Liu, Saman Dehghan, Jatin Ganhotra, Martin Hirzel, Reyhaneh Jabbarvand - [[pdf]](https://arxiv.org/pdf/2604.12147)
Abstract: Agents are commonly instructed to follow a task-specific plan for guidance. However, it is unknown to what extent agents actually follow instructed plans. Without such an analysis, determining the extent agents comply with a given plan, it is impossible to assess whether a solution was reached through correct strategic reasoning or through other means, e.g., data contamination or overfitting to a benchmark. This paper presents the first extensive, systematic analysis of plan compliance in programming agents, examining 21,120 trajectories from SWE-agent across four LLMs on SWE-bench Verified and SWE-bench Pro under eight plan variations. Without an explicit plan, agents fall back on internalized workflows during training, which are often incomplete, overfit, or inconsistently applied. Providing the standard plan improves issue resolution, and we observe that periodic plan reminders can mitigate plan violations and improve task success. A subpar plan hurts performance even more than no plan at all. Surprisingly, inserting additional task-relevant phases in the early stage can degrade performance, particularly when these phases do not align with the model's internal problem-solving strategy. These findings call for fine-tuning paradigms that teach models to follow instructed plans, rather than encoding task-specific plans in them, so that they reason and act adaptively, rather than memorizing workflows.
arXiv ID: 2604.12147
Toward Reliable Context Compression for Long-Horizon Agents: An Empirical Study of Execution Instability - Guanghui Min, Liang Wu, Mayank Darbari, Chen Chen, Liangjie Hong - [[pdf]](https://arxiv.org/pdf/2608.06503)
Abstract: Recurrent context compression controls context growth in long-horizon agents, but its behavioral effects remain poorly understood. In this preliminary empirical study, we show that compression can weaken the influence of recent interactions, increasing blocked actions, repeated exploration, and instability across runs. Motivated by these observations, we introduce TRACE, a verifier-guided framework that evaluates individual compaction events through paired closed-loop continuations from the same environment state and uses summary preferences to optimize a natural-language compression prompt while keeping all models frozen. Initial results on AppWorld show improvements over existing compression baselines in task performance, multi-run reliability, and context--execution efficiency. These findings provide early evidence for boundary-local evaluation as a promising direction for reliable agent context compression.
arXiv ID: 2608.06503
WAM-Diff2: Hierarchical AR-to-Diffusion Distillation for Highly Efficient Autonomous Driving VLA - Zhihao Zhu, Hanlin Shang, Mingwang Xu, Feipeng Cai, Zhuolin He, Yaoyi Li, Jianhua Han, Hang Xu, Siyu Zhu - [[pdf]](https://arxiv.org/pdf/2608.01035)
Abstract: Vision-Language-Action (VLA) models have emerged as a prominent paradigm for end-to-end autonomous driving; however, their efficient deployment is severely constrained by high computational latency and exposure bias arising from sequential autoregressive decoding. Conversely, while specialized diffusion policies enable low-latency, parallel execution, training them from scratch typically yields narrow, single-task architectures that lack holistic visual-linguistic reasoning. Successfully transforming pre-trained autoregressive generalists into parallel diffusion models could combine multi-task cognitive intelligence with execution efficiency, yet this transition presents a formidable architectural challenge due to mismatched attention patterns (causal versus bidirectional) and divergent optimization objectives. To bridge this divide, we introduce WAM-Diff2, a multi-task discrete diffusion VLA framework powered by a three-stage hierarchical distillation strategy. By structuring the architectural shift through progressive block-wise adaptation, block-wise distillation, and model-wise cross-scale distillation, WAM-Diff2 preserves the underlying semantic foundations of the base model while accelerating inference. Extensive evaluations across driving understanding, perception, and planning benchmarks demonstrate that WAM-Diff2 effectively mitigates exposure bias and achieves performance parity with autoregressive baselines. Crucially, the autoregressive-to-diffusion transition yields a 2.8x decoding speedup, which scales to an ultimate 15.1x acceleration when combined with system-level optimizations including FlashInfer and CUDA Graphs.
arXiv ID: 2608.01035
LLM Agents (13 papers)
Gated-BEPO: Confidence-Gated Bellman Credit Assignment for Large Language Model Agents - Hongxi Yan, Ziyue Huang, Shichao Fan, Qingjie Liu - [[pdf]](https://arxiv.org/pdf/2608.06861)
Abstract: Training large language model agents in long-horizon environments requires assigning credit from sparse terminal outcomes to individual actions. Existing critic-free methods propagate trajectory-level rewards uniformly across steps, while recent approaches construct step-level groups by matching repeated states and compare actions within each group. The former cannot distinguish useful actions in failed trajectories from ineffective actions in successful ones. The latter rely on step credit derived directly from individual trajectory outcomes and fixed-weight fusion with episode-level credit. We propose Gated-BEPO, which derives step-level credit from empirical rollout graphs. For each rollout group, Gated-BEPO constructs an empirical graph and estimates node values through a mean-backup Bellman fixed point that reflects the empirical action distribution of the current policy. We then accumulate these temporal-difference residuals along each sampled trajectory using generalized advantage estimation, yielding step-level Bellman advantages that capture both immediate and downstream effects. To adaptively fuse episode- and step-level credit, a confidence gate incorporates Bellman credit only at states with multiple observed successors and otherwise uses episode-level credit. Experiments on WebShop, ALFWorld, and visual Sokoban show consistent improvements across language and vision-language models, while diagnostic ablations support the effectiveness of Bellman fixed-point value estimation and show that step-level credit should be incorporated selectively rather than uniformly into the final advantage.
arXiv ID: 2608.06861
NiyamAI - An Intent-Bound AI Agent with Cryptographically Verifiable Guardrails using Zero-Knowledge Proofs - Aditya Katkar, Om Karkele, Kartik Mandhane, Manisha More, Yash Kashid - [[pdf]](https://arxiv.org/pdf/2608.07167)
Abstract: Giving an AI agent the ability to send emails, query databases, or execute commands is useful--until the agent is tricked into doing something it shouldn't. Prompt injection, hallucinated reasoning, and unsafe tool calls form the primary attack surface for autonomous LLM agents. Existing defenses rely on software checks like system prompts or policy filters running on the same machine the attacker targets, offering no verifiable proof of execution. We introduce Niyam-AI, a framework that makes safety enforcement provable. At session start, permitted tools and constraints are locked into an Intent Contract committed via SHA-256. Every tool call is intercepted and validated by an isolated Judge model; upon passing, a zk-SNARK proof is generated via EZKL. The tool executes only after proof verification, allowing third parties to confirm enforcement without accessing Judge model weights. Evaluating Niyam-AI on 2,000 real-world scenarios from Agent-SafetyBench against NeMo Guardrails, Meta's Llama Prompt Guard 2, and OpenAI's GPT-OSS-Safeguard using 5-fold stratified cross-validation yields an F1 score of 88.5% with a 1.1% false-positive rate (bootstrap 95% CI: [85.19%, 91.88%], N=1000). McNemar's exact paired test confirms significant improvement: Niyam-AI wins 390 discordant scenarios against NeMo (vs 20 losses), 115 against Prompt Guard 2 (vs 13), and 384 against GPT-OSS-Safeguard (vs 19) with p < 0.0001 in all cases. Proof generation adds 2260.6 +/- 218.4 ms per approved action, while verification takes 53.1 +/- 11.8 ms. Niyam-AI provides a guardrail that is both highly accurate and mathematically verifiable--though this reflects a classifier adapted to Agent-SafetyBench evaluated against zero-shot baselines, a distinction discussed in Section IV.C.
arXiv ID: 2608.07167
Agent Memory Distillation: Empowering Small LLM Agents with Hierarchical Teacher Memory - Taeil Kim, Kangsan Kim, Sung Ju Hwang - [[pdf]](https://arxiv.org/pdf/2608.07169)
Abstract: Memory systems have shown promise for improving agent performance, but their potential remains largely unexplored for small language models, which struggle to generate sufficient successful trajectories on their own. We propose Agent Memory Distillation (AMD), a training-free framework that transfers structured knowledge from a large teacher agent to a small student agent through hierarchical memory. AMD constructs three complementary memory types from successful teacher trajectories: Workflow memory encodes task-level strategies, Subtask memory provides concrete behavioral examples at an intermediate granularity, and Function memory captures per-function calling conventions and common pitfalls. Workflow and Subtask memories are injected proactively at the start of each task, while Function memory is retrieved reactively upon tool-calling errors. We evaluate AMD on three tool-use benchmarks using four student models (4B-8B parameters) with GPT-5-mini as the teacher, achieving average accuracy gains of 27.2%p, 11.2%p, and 3.4%p on AppWorld, BFCL V3, and ToolSandbox, while consistently outperforming existing memory-based baselines. Further analysis shows that Subtask memory contributes the largest gains, teacher effectiveness depends on both teacher capability and student compatibility, and 4B-sized students benefit most from AMD.
arXiv ID: 2608.07169
Fisher-R1: Training LLM Agents for Reliable Hypothesis Testing - Jiacheng Miao, Jin Mu, Guanhua Chen, James Zou - [[pdf]](https://arxiv.org/pdf/2608.07437)
Abstract: Reliable hypothesis testing is the foundation of many empirical scientific claims. Large language model (LLM) agents are increasingly used to automate this process, as they can inspect datasets, generate code, and produce analyses end-to-end. However, we show that they frequently make subtle inferential errors that lead to incorrect conclusions despite correctly executed analyses. Existing benchmarks fail to capture this failure mode, as they rarely assess whether a reported p-value is statistically valid given the assumptions underlying the data. We address this gap by building P-Bench, a benchmark comprising 425 open-ended, realistic hypothesis-testing tasks spanning economics, biology, and medicine. Each task requires an agent to select a statistical method, compute a p-value, and draw a conclusion given only a scientific hypothesis and a dataset. We further introduce Fisher-R1, an open-weight LLM agent trained for rigorous hypothesis testing using synthetic tasks and reinforcement learning. On P-Bench, Fisher-R1-14B substantially improves over its backbone and outperforms strong proprietary and open-source baselines, including GPT-5.4 and DeepSeekV4-Pro, achieving a 21% average relative improvement in single-trial success over DeepSeek-V4-Pro, with gains up to 26% on the most challenging tasks. Our results demonstrate that current LLM agents lack reliable statistical reasoning for hypothesis testing and that reinforcement learning on tasks with verified statistical reward substantially improves reliability.
arXiv ID: 2608.07437
PsychoAgent: An Affect-Sensitive Cognitive Architecture for Conflict-Aware Memory in LLM Agents - Mohammad Amanlou, Parham Abed Azad, Farbod Davoodi, Mostafa Masumi, Behnam Bahrak, Abdol-Hossein Vahabie - [[pdf]](https://arxiv.org/pdf/2608.07438)
Abstract: Human-like cognition does not select past experience by topical similarity alone: affective significance and unresolved conflict also shape what becomes accessible. We present PsychoAgent, a cognitive architecture for LLM agents that separates factual and affective memory and integrates both through a conflict-aware executive controller. Affective memories are first filtered by semantic relevance and then re-ranked by salience, preserving topical fit while allowing emotionally important traces to enter the prompt. Across three controlled conflict scenarios, the full architecture retrieved more conflict-critical memories than semantic-affective and single-memory RAG baselines (0.933 vs. 0.500 and 0.667), with a small semantic-similarity cost. Five blinded raters evaluated 27 outputs. After within-rater standardization, the full architecture had the highest overall mean (+0.22 SD), but corrected pairwise differences were not significant. A three-day illustrative trace further shows persistent affect, offline memory recombination, and selective memory reweighting. The findings support affect-sensitive retrieval as an inspectable mechanism for modeling human-like conflict effects in LLM agents.
arXiv ID: 2608.07438
SkillProx: Self-Evolving Agent Skills via Proximal Textual Gradient Descent - Mingxuan Zheng, Yujin Zhou, Chuxue Cao, Boqin Yin, Yuyao Zhang, Jiapeng Sun, Shuaishuai Gong, Sirui Han, Yike Guo - [[pdf]](https://arxiv.org/pdf/2608.07449)
Abstract: LLM agents increasingly adapt to recurring tasks by accumulating procedural knowledge in skills. These skills are lightweight, reusable textual artifacts that are loaded into the agent's context without weight updates. Recent methods refine skills through iterative task execution, failure diagnosis, and trajectory-guided text-space updates. However, existing frameworks lack explicit diagnosis--outcome feedback and treat deletion as a generic edit operation rather than a dedicated mechanism for consolidating accumulated knowledge. We introduce SkillProx, a proximal-gradient-inspired forward--backward framework that couples closed-loop diagnostic evolution with utility-aware proximal refinement. Motivated by a composite objective balancing task loss and skill complexity, the forward stage re-executes diagnosis-driven edits on the same task batch, rolls back regressions, and feeds measured outcomes into subsequent diagnoses. The backward stage decomposes the resulting skill into auditable knowledge units, estimates their contributions using a frozen leave-one-out utility audit, and applies validation-gated consolidation, demotion, or removal. Experiments on in-distribution and out-of-distribution benchmarks across multiple backbone LLMs show that SkillProx improves average accuracy by 3.0 percentage points over the strongest gradient-based baseline. Component ablations demonstrate the complementary effects of closed-loop diagnosis and proximal refinement.
arXiv ID: 2608.07449
Do AI Personas Grow? Analyzing and Benchmarking Personality Evolution in LLM Agents After Life Events - Ming Wang, Peidong Wang, Xiaocui Yang, Daling Wang, Shi Feng, Fiona Fui-Hoon Nah, Ee-Peng Lim - [[pdf]](https://arxiv.org/pdf/2608.06485)
Abstract: Personality-conditioned LLM agents (PC-Agents) are increasingly used in emotional support, social simulation, and role-playing, motivating the development of lifelong agents that remain coherent over extended interactions. A key component of such coherence is personality evolution: agents should undergo plausible, psychology-grounded changes as they experience life events in different contexts. Although prior work shows that LLM personalities can shift under contextual perturbations, how these shifts vary across traits, events, personas, and models remains poorly understood. We study event-induced personality change after 11 major life events, using the Big Five traits as a psychometric anchor and interpreting the resulting trajectories against longitudinal evidence from human personality psychology. Across four diagnostic axes, PC-Agents exhibit measurable trait shifts at similar rates for event-trait pairs with and without documented human change directions. Even when shifts follow the expected direction, their magnitudes usually fall below human effect-size ranges. Gender and cultural-region prompts show little moderating effect, while persona-level dispersion is compressed three- to four-fold relative to human samples. To enable systematic comparison, we introduce BFI-Adapt, a reusable benchmark for scoring the directional fidelity of event-induced personality change, and use it to rank 14 models. A validation suite shows that the measured shifts exceed no-event retest noise, remain stable under independently paraphrased prompts, exhibit limited and model-dependent convergence with scenario-based behavioral choices, and persist across intervening unrelated dialogue. Together, these checks establish the measured trajectories as robust event-conditioned response patterns. Our results suggest that current PC-Agents simulate the mean of human personality dynamics, but not its shape.
arXiv ID: 2608.06485
Coupling Planning with Episodic Memory in LLM Agents for Software Issue Resolution - Jiahao Zhang, Yifan Zhang, Yu Huang - [[pdf]](https://arxiv.org/pdf/2608.06811)
Abstract: Resolving a real software issue with a large language model (LLM) agent is a long repair episode, often tens to hundreds of steps spanning exploration, hypothesis, implementation, and verification. Success depends on both the base model's local reasoning and the agent's ability to maintain an evolving plan and remember observations across phases. Existing repository-level agents typically strengthen planning or memory in isolation, leaving long trajectories vulnerable to stale evidence, repeated failed edits, and verification inferred from the agent's own claims instead of execution evidence. We present PMCoder, an issue-resolution agent that couples a hierarchical phase planner with episodic memory. The coupling is bidirectional: the current plan phase conditions memory retrieval, while memory-derived trajectory statistics inform stuck detection and replanning. When available, issue-reproduction verdicts ground verification progress in execution evidence rather than self-reported completion. On SWE-bench Verified, PMCoder resolves an average of
arXiv ID: 2608.06811
"LLM Agent Performance" Is Not a Single Evaluation Target - Pengyu Zhu, Li Sun, Philip S. Yu, Sen Su - [[pdf]](https://arxiv.org/pdf/2602.03238)
Abstract: LLM agent benchmark scores are shaped not only by the model but also by the agent harness, environment, evaluator, and inference budget. Unified execution controls these non-model factors by evaluating candidate models under the same configuration, making observed differences more attributable to the models themselves. However, model comparison is only one use of agent benchmarks. Other evaluations compare complete agent systems or test whether a fixed model or system remains stable across predeclared changes in its operating conditions. These results can all be reported under the common label of "LLM agent performance." Our position is that "LLM agent performance" does not denote a single evaluation target. Model comparisons under a reference stack and comparisons of complete agent systems answer different questions, while robustness asks whether either conclusion persists across predeclared conditions. The claim supported by a score therefore depends on the declared candidate boundary and condition policy. We derive implications for leaderboards, result reporting, and benchmark versioning, showing how distinguishing these classes preserves fair comparison while accommodating system innovation and robustness analysis.
arXiv ID: 2602.03238
ForesightSafety-SAGE:A Fully Automated Scenario Generation and Safety Evaluation Framework for LLM Agents - Lu Jia, Haibo Tong, Feifei Zhao, Jindong Li, Dongqi Liang, Ping Wu, Qian Zhang, Yi Zeng - [[pdf]](https://arxiv.org/pdf/2606.08531)
Abstract: Large language models (LLMs) are increasingly evolving from simple text-based interaction systems into LLM agents that can maintain memory, use tools, access external environments, and execute tasks. As their capabilities and autonomy expand, the safety risks they face also become more diverse. Existing evaluations often rely on manually written scenarios, static prompts, or final-output judgments, making it difficult to capture the diverse risks that agents may face during task execution. We introduce ForesightSafety-SAGE, a fully automated scenario generation and safety evaluation framework for LLM agents. Based on five risk dimensions,we instantiae abstract and diverse safety risks in real-world task execution into 1,072 measurable evaluation scenarios. Using the automated evaluation pipeline, 12 LLM agents are evaluated under two authority contexts. The results show that current agents still face substantial behavioral safety risks during task execution, with an average ASR of 47.1% and several models exceeding 70%. These findings demonstrate the importance of executable, process-level evaluation for understanding and improving LLM agent safety.
arXiv ID: 2606.08531
Harnessing the Synergy between LLM Agents and Knowledge Graphs for Urban Socioeconomic Prediction - Zhilun Zhou, Jingyang Fan, Yu Liu, Fengli Xu, Depeng Jin, Yong Li - [[pdf]](https://arxiv.org/pdf/2411.00028)
Abstract: Socioeconomic prediction aims to leverage various urban data to predict the socioeconomic indicators of regions such as population and commercial activity level, which plays an important role in understanding urban regions and supporting decision-making. Existing studies leverage knowledge graphs (KG) to model heterogeneous urban data, and further apply graph representation learning methods for socioeconomic prediction. However, these approaches heavily rely on heuristic ideas and expertise to extract task-relevant knowledge from diverse data, which may not be optimal for specific tasks. Additionally, they tend to overlook the inherent relationships between different indicators, limiting the prediction accuracy. Motivated by the remarkable abilities of large language models (LLMs), in this work, we propose a synergistic framework of LLM agents and KG, which integrates the reasoning and representation learning on KG with LLM agents. We first construct an urban knowledge graph (UrbanKG) to model multi-sourced urban data and finetune an embedding language model to generate embeddings for KG entities with semantic information. Then we leverage the reasoning power of LLM to identify relevant meta-paths in the UrbanKG for each type of socioeconomic prediction task, and design a semantic-guided attention module for knowledge fusion with meta-paths. Moreover, we introduce a cross-task communication mechanism to further enhance performance by enabling knowledge sharing across tasks at both LLM agent and KG levels. On the one hand, the LLM agents for different tasks collaborate to generate more diverse and comprehensive meta-paths. On the other hand, the embeddings from different tasks are adaptively merged. Experiments on two datasets demonstrate the effectiveness of the synergistic design between LLM and KG, providing insights for information sharing across socioeconomic prediction tasks.
arXiv ID: 2411.00028
The Horizon Gap: Planning, Memory, Execution, Training, and Evaluation for Long-Horizon LLM Agents - Mingguang Chen, Licheng Wang, Bo Qu - [[pdf]](https://arxiv.org/pdf/2608.06663)
Abstract: Frontier language models solve reasoning problems in a single forward pass that would have been research contributions years ago, yet fail at multi-hour tasks: losing track of earlier decisions, declaring half-finished work done, or drifting from goals. We call this the horizon gap and survey 1,547 arXiv papers (2024-2026) collected via systematic seed harvest with a disclosed 26.8% bleed filter, extended by targeted supplementation. We disambiguate three routinely conflated properties: long-horizon (task property: required steps), long-context (model property: token capacity), and long-term memory (system property: persistence across steps/sessions). We organize the corpus into six categories tracking a long-horizon task's lifecycle -- planning, memory, execution, training, evaluation, and foundations/safety -- crossed with an axis capturing where horizons are carried (within-context, within-task-beyond-context, or cross-task-persistent). Across all categories, we find the same pattern: outcome-only signals grow uninformative as horizons lengthen, and the field's response -- whether process reward models, credit assignment, or trajectory-level diagnostics -- manufactures denser step-level signals. We treat critical and diagnostic literature as first-class threads throughout, arguing that segregating critique from method would routinely split single papers across chapters. We close by naming open measurement problems: decomposing model versus harness capability, managing correlated bias in process-level signals used for both training and evaluation, and whether long-horizon reliability admits general predictive theory.
arXiv ID: 2608.06663
PULSE: Agentic Investigation with Passive Sensing for Proactive Affective Intervention in Cancer Survivorship - Zhiyuan Wang, Subigya Nepal, Ariful Islam, Indrajeet Ghosh, Xinyu Chen, Katharine E. Daniel, Laura E. Barnes, Philip Chow - [[pdf]](https://arxiv.org/pdf/2605.17679)
Abstract: Cancer survivors face elevated rates of depression, anxiety, and emotional distress, yet self-report may be unavailable at some moments when support is relevant, a challenge we term the diary paradox. We present PULSE, a system for agentic sensing investigation: LLM agents equipped with eight purpose-built tools query smartphone sensing data, compare current behavior with personal baselines, and retrieve outcome-labeled historical cases. Rather than receiving only a fixed feature summary, agents choose which modalities and time windows to inspect. We evaluate PULSE through a two-by-two evaluation design crossing system architecture (structured single-pass vs. multi-turn agentic) with concurrent input modality (no current diary vs. sensing plus current diary) on 50 cancer survivors. The agentic multimodal condition achieves balanced accuracy of 0.743 for emotion-regulation desire; the agentic no-current-diary condition achieves 0.713 for self-reported intervention availability. This is a system-level comparison because the architecture conditions also differ in tool-mediated information access. The results provide a retrospective benchmark for interactive sensing investigation and motivate prospective evaluation at diary non-response moments.
arXiv ID: 2605.17679
Multi-Agent Systems (11 papers)
A Multi-Agent Framework for Automated Coarse-Grained Molecular Dynamics of Polymers - Joohee Choi, Junhyeong Lee, Seunghwa Ryu - [[pdf]](https://arxiv.org/pdf/2608.06694)
Abstract: Coarse-grained (CG) molecular dynamics extends polymer simulation beyond the scales accessible to all-atom (AA) methods, but bottom-up CG modeling is laborious. The CG resolution is a design choice, so a transferable parameter set is generally not available and the potentials are derived anew for each polymer mapping. Here we present CGMas, a multi-agent framework that automates topology construction, equilibration, mapping, potential derivation, and validation from a natural-language specification of the polymer and target resolution. A large-language-model (LLM) reasoning agent infers the AA topology from polymer name, while layered self-correction resolves physical errors common to unsaturated, heteroatom-containing, and polar polymers. Downstream agents equilibrate the system, map it onto CG representation, derive potentials through Boltzmann inversion, and benchmark the model against its atomistic reference. CGMas completed all 27 homopolymer and copolymer tasks, matched the AA density to within 5% in 22, and reduced simulation from 38-88 min to 1 min, establishing agentic LLMs as a route to automated polymer coarse-graining.
arXiv ID: 2608.06694
Deal Me Maybe: The Role of Emotions in Multi-Agent Negotiation - Massimiliano Luca, Apoorva Singh, Bruno Lepri - [[pdf]](https://arxiv.org/pdf/2608.06922)
Abstract: Negotiation is a demanding social task for LLM agents, requiring strategic reasoning, persuasion, and interpersonal adaptation. Yet existing benchmarks often treat agents as emotionally neutral, overlooking a key driver of human bargaining behavior. We study how prompt-conditioned emotions affect LLM-based price negotiation. In a controlled framework, buyer and seller agents are independently assigned one of six emotional states and negotiate over 350 real consumer products under two budget conditions. Across 36 emotion-pair settings and five widely used LLMs, we find that emotions strongly shape outcomes. Angry buyers almost never reach agreement (0.39% deal rate), while happy buyers agree most often (28.91%), but obtain worse prices than fearful buyers. Emotion effects are role-dependent: buyer emotion mainly drives acceptance and rejection, whereas seller emotion shapes concession dynamics. These effects influence not only language, but also termination behavior and price trajectories, raising concerns for emotion-conditioned agents in commerce.
arXiv ID: 2608.06922
Does Splitting a Triage Decision Across Agents Hide Bias or Help Catch It? A Multi-Agent Simulation Study of LLM-Based Resource Allocation Under Audit Capacity Constraints - Paul-Peter Arslan - [[pdf]](https://arxiv.org/pdf/2608.06949)
Abstract: Prior benchmarking work has shown that a single large language model (LLM), forced to make life-or-death resource-allocation decisions, exhibits measurable demographic bias. Real deployments, however, rarely use a single agent: they use pipelines, with review steps meant to catch exactly this kind of failure. We study what happens to bias when the same decision is distributed across a role-differentiated multi-agent pipeline (assessment, allocation, independent audit) instead of made and checked by one model alone. Using a synthetic disaster-triage simulator with paired cases that are clinically identical except for one demographic attribute, we run 192 episodes (2,304 resolved case pairs) on GPT-4o-mini comparing a single-agent control condition to a nine-agent pipeline under three independently varied pressure dimensions. We find no measurable difference in how often biased outcomes occur between the two conditions (6.9% vs. 6.1%, p = 0.498). We do find a large and significant effect of audit capacity on whether bias is caught: 30.0% of biased outcomes go entirely undetected, rising to 43.8% when the auditor is overloaded and falling to 18.4% when it is not. Decomposing this effect shows it is driven almost entirely by coverage (whether a case is reviewed at all, which collapses from 100.0% to 65.6% under load, p < 0.001) rather than by degraded judgment on the cases that are reviewed (81.6% vs. 85.7%, p = 1.000, direction reversed). A follow-up experiment shows that reordering the audit queue by estimated risk, rather than first-come-first-served, recovers most of the lost coverage under the same capacity constraint (65.6% to 91.7%, p = 0.028). We discuss the implications for any system that adds independent oversight to an LLM agent pipeline under resource constraints, and report the study's limitations honestly: one model, modest sample sizes, and no adversarial replication.
arXiv ID: 2608.06949
EMAS: Stabilizing Multi-Agent System Evolution through Evidence-Guided Revision - Chao Fei, Qingyi Si, Kaihua Liang, Yanghua Xiao, Panos Kalnis, Hongcheng Guo - [[pdf]](https://arxiv.org/pdf/2608.07196)
Abstract: Many methods for automated multi-agent system design optimize prompts and topologies during an initial design stage and then deploy the resulting system unchanged on subsequent samples. Experience from these samples is rarely consolidated into reusable system updates, while accuracy-oriented designs may incur high token costs. We introduce EMAS (Evolving Multi-Agent System), which uses this experience to revise MAS topology and prompts without updating LLM parameters, either to improve accuracy or to reduce cost. EMAS converts traces into structured diagnoses that specify a revision operation and target. It generates a candidate revision only when the same diagnosis recurs across samples and applies it only if paired validation against the current MAS meets the corresponding acceptance criterion. Across four benchmarks and two LLMs, EMAS attains the highest task-weighted overall accuracy for both backbones and is best or tied in six of eight model--benchmark settings. Within two evolution epochs, EMAS achieves relative gains of 6.30% and 20.10% in task-weighted accuracy on Kimi-K2-6 and Qwen3.6-27B, respectively. On MBPP with Qwen3.6-27B, EMAS raises accuracy from 55.09% to 89.12% while reducing token use per task by 62.2%. These results show that EMAS can turn experience from new samples into reusable updates to MAS topology and prompts.
arXiv ID: 2608.07196
Multi-Agent Forensic Reasoning for Generalizable Deepfake Video Detection - Xuechao Zou, Shun Zhang, Kai Li, Yi Zhou, Xinyu Sun, Yuhui Chen, Zhe Wu, Congyan Lang, Junliang Xing - [[pdf]](https://arxiv.org/pdf/2608.06865)
Abstract: The malicious use of generative artificial intelligence to create highly realistic deepfake videos raises serious ethical concerns and poses substantial challenges to AI safety. However, existing deepfake video benchmarks provide limited coverage of recent synthesis methods and generally lack reliable fine-grained textual annotations. Meanwhile, conventional detectors and multimodal large language models (MLLMs), whether operating as a single model or relying on a single analytical perspective, often fail to capture subtle forgery artifacts, limiting their generalization to emerging AI-generated methods. To address these limitations, we introduce FaceVid-Forensics-100K, a large-scale deepfake video dataset comprising 100,000 videos and spanning 33 synthesis methods across face swapping, face reenactment, and entire-face synthesis, including recent generators such as Seedance 2.0. The dataset provides fine-grained textual annotations of visual observations and verdict-consistent forensic explanations, automatically synthesized through a multi-model aggregation and conflict-resolution pipeline powered by advanced MLLMs. Building on this benchmark, we propose a multi-agent forensic reasoning framework that employs four specialized domain-expert agents to independently analyze forgery cues from four perspectives: texture, lighting, motion, and physics. A judge agent then reconciles their reports to produce a final prediction together with an explanation. Extensive evaluations on out-of-domain test sets show that, despite being composed entirely of small open-source MLLMs, our framework outperforms all methods including closed-source GPT and Gemini models and ranks first across all reported metrics on this benchmark. The project page is available at this https URL.
arXiv ID: 2608.06865
PHOENIX: Fine-Tuned SLM-Powered Autonomous Satellite Lifetime Extension via Predictive Self-Healing and Multi-Agent AI Recovery - Sumaiya Islam, Harsha Kumara Moraliyage - [[pdf]](https://arxiv.org/pdf/2608.07126)
Abstract: Most CubeSats, small and low-cost satellites roughly the size of a shoebox, do not survive as long as they were designed to: a study of 178 missions found that only 48-65% remain operational after two years, against a designed lifetime of 2-5 years. The deeper issue is that a CubeSat in low Earth orbit (LEO) is physically unreachable from the ground for roughly 85 minutes out of every 96-minute orbit, so faults that start during that window go unnoticed until the next contact pass, by which point recovery may no longer be possible. We propose PHOENIX (Predictive Health On-orbit Edge Neural Intelligence eXtension) to give the satellite its own fault reasoning capability. A fine-tuned Small Language Model (SLM) compact enough to run on embedded hardware is deployed onboard the CubeSat, running on the flight-proven Aethero NxN-ECM computer, monitoring all sensor readings continuously, and resolving recurring faults using a memory system that stores past repairs so the same inference does not need to run twice. Once per orbit it sends a short structured health report to the ground instead of a raw data dump; six specialized AI agents on the ground read that report and generate validated satellite commands within the 5-10 minute contact window. A generative diffusion model (DDPM) creates synthetic training data because real fault examples make up only 0.57-1.80% of the dataset. We report preliminary results on the ESA Anomaly Detection Benchmark (14 years, 76 channels, 118 labeled faults).
arXiv ID: 2608.07126
AutoMOOSE: An Agentic AI for Autonomous Phase-Field Simulation - Sukriti Manna, Henry Chan, Subramanian K.R.S. Sankaranarayanan - [[pdf]](https://arxiv.org/pdf/2603.20986)
Abstract: Phase-field modeling links thermodynamics and kinetics to microstructural evolution, but multiphysics frameworks such as MOOSE require expertise to construct inputs, manage campaigns, diagnose failures, and validate results. We introduce AutoMOOSE, an open-source multi-agent framework that orchestrates the simulation lifecycle from a single natural-language prompt. Six specialized agents--Architect, Input Writer, Runner, Reviewer, Visualization, and a physics-grounded Skeptic--generate, execute, analyze, and adversarially test simulations against conservation laws, asymptotic limits, and scaling relations. We validate AutoMOOSE in two domains: non-conserved copper grain growth governed by Allen-Cahn dynamics and conserved Fe-Cr spinodal decomposition governed by Cahn-Hilliard dynamics. On a prospectively specified 25-task grain-growth benchmark spanning temperature, grain count, resolution, and model formulation, AutoMOOSE generates valid inputs for all 25 tasks, completes 19 simulations with measurable coarsening, and delivers 15 results that satisfy Burke-Turnbull kinetics (
arXiv ID: 2603.20986
Why Study Emergent Behavior When You Can Regulate It? Aligning Multi-Agent Systems with Reward Prediction - Assaf Caftory, Almog Zemach, Moshe Butman, Doron Friedman - [[pdf]](https://arxiv.org/pdf/2608.07280)
Abstract: Multi-agent simulations are widely used to study complex social and ecological systems, where rich and often unexpected emergent behaviors arise from local interactions. A large body of prior work has focused on analyzing such emergent dynamics across domains. In this paper, we move beyond analyzing emergent behavior and introduce a learning-based mechanism for actively shaping it via social reward modeling. We introduce Multi-Agent Reward Prediction (MARP), a simple framework that extends preference-based reward modeling to multi-agent reinforcement learning. While the framework is designed to be applicable across multi-agent settings, the present empirical validation is limited to a single environment, and we therefore present MARP as a proof of concept within the studied domain. Rather than relying on handcrafted rewards, MARP learns a shared reward model from episode-level evaluations of collective outcomes, enabling decentralized agents to align their behavior with global social objectives. We study MARP in the Harvest Game, a canonical sequential social dilemma modeling common-pool resource management and related real-world challenges. Our results show that MARP can be tuned to produce behavior that is more closely aligned with target social metrics than standard reward-based baselines, while the learned reward model captures subtle environmental structure without explicit programming. Crucially, MARP supports multiple and composite social objectives within a single training regime. By modifying only the high-level evaluation metric, the same framework seamlessly aligns agent behavior with diverse goals, including sustainability, equality, and peace, as well as combinations of individual and group-level objectives. These findings demonstrate that emergent multi-agent behavior can be treated not only as a phenomenon to study, but as a target of principled, data-driven regulation.
arXiv ID: 2608.07280
Online Security Learning in Cooperative Multi-Agent Systems under Hidden Byzantine Attacks - Ximing Sun, Yue Wang - [[pdf]](https://arxiv.org/pdf/2608.06520)
Abstract: We study online cooperative control of a multi-agent system under Byzantine attacks. Namely, an unknown, fixed subset of agents are Byzantine comprised and can stealthily overwrite its own coordinates of the team's planned joint action after observing that plan. The learner observes planned actions, public rewards, and public states, but neither the overwrite nor the executed joint action. Our objective is security: to optimize the team performance against the worst overwrites and achieve the optimal security value. We first show that the attacker's information determines the geometry. An attacker that observes the planned action induces an exact
arXiv ID: 2608.06520
Plan-and-Avoid: Real-Time Aircraft Trajectory Coordination in a Multi-Agent Environment - Huseyin Emre Tekaslan, Ella M. Atkins, Natasha Neogi - [[pdf]](https://arxiv.org/pdf/2608.06648)
Abstract: This paper presents a real-time Plan-and-Avoid (PAA framework for coordinating cooperative multi-agent airspace operations around a declared priority trajectory. The priority trajectory represents an aircraft flight plan that must be preserved because of constrained maneuverability, an emergency, a mission-critical task, or assigned operational priority. The framework predicts uncertainty-aware, well-clear separation violations with surrounding traffic and, when the priority plan alone cannot maintain separation, generates vehicle-constrained unilateral advisories that modify nearby aircraft trajectories to maintain well-clear separation for all traffic. The approach is applicable to any declared priority trajectory. This paper demonstrates the Plan component using a contingency landing planner to generate candidate priority trajectories. PAA then identifies nearby aircraft passing too close to this priority trajectory and issues Avoid resolution advisories to these aircraft. The framework is tested using real-world Automatic Dependent Surveillance-Broadcast (ADS-B) traffic from the Washington, D.C., airspace across more than 900 forced-landing cases, totaling over 140 hours of simulated flight. The PAA framework generates feasible cooperative advisories for all 575 unique conflict encounters, with a worst-case end-to-end response time of 5.7 s on a personal computer, including priority trajectory planning, advisory generation, and 1 s two-way datalink delay. In total, 93.5% of generated advisories satisfy the 35 s RTCA DO-365 Detect-and-Avoid temporal threshold. These results demonstrate low-latency coordination for preserving priority trajectories while maintaining well-clear separation through real-time automated advisory generation. Future work will quantify advisory-induced delays and their operational impacts.
arXiv ID: 2608.06648
Robot guide with multi-agent control and automatic scenario generation with LLM - Elizaveta D. Moskovskaya, Anton D. Moscowsky - [[pdf]](https://arxiv.org/pdf/2509.10317)
Abstract: The article describes the development of a hybrid social robot control architecture to overcome the limitations of traditional approaches, where behavior scripts manually synchronize the robot's actions and text, and existing methods focus primarily on short dialogue responses. The architecture of the proposed system combines a multi-agent resource management system with automatic generation of behavior scenarios based on large language models. This system automates the preparation of text and commands for the robot's non-verbal behavior for extended narratives and resolves resource conflicts between multiple execution mechanisms. The system was tested on the MENTOR-1 tour guide robot, for which it successfully generated scenarios automatically and demonstrated more natural and rich behavior compared to existing approaches. The proposed approach provides full automation of both scenario preparation and execution through efficient resource management, enhancing the quality of social robot interaction in long-term storytelling tasks.
arXiv ID: 2509.10317
KNOWPLAN: Knowledge-Driven AI Agents for Smart Degree Pathway Planning - Shuheng Cao, Weijia Zhang, Jiaqi Wu, Xiyun Hu, Yat Yang, Juqy Chen, Zhaoxiang Feng - [[pdf]](https://arxiv.org/pdf/2608.06530)
Abstract: Planning a degree from official university sources requires solving two problems in order. The institution's curriculum must first be reconstructed from catalogs, departmental pages, JSON endpoints, and PDFs that share no schema, and only then can a student-specific path be optimized under prerequisite logic and overlapping requirement constraints. Coupling the two lets each failure mode hide the other, because a planner that drives its own crawling never learns facts its current plan does not need. We present KnowPlan, which enforces an extraction-first boundary and measures the interface between the stages rather than assuming it. CatalogBrowse explores with no access to any user profile. It scores legal actions by lower-confidence expected marginal gain over a finite set of atomic catalog obligations per unit of source access, parses deterministically through platform adapters with a span-constrained clause-to-AST model fallback, and terminates on a closure certificate over index, schema, provenance, and reference completeness instead of a reward threshold. Its output contract is three provenance-linked JSON documents. DegreeMap consumes only those documents. It compiles them into a typed requirement hypergraph and optimizes lexicographically with CP-SAT over hard feasibility, completion horizon, load and risk, personalized utility, and option value, so that each stage optimizes inside the previous stage's proven optimum and stays certifiable within the solver budget. Across a 100-university broad track and a six-school dense track, CatalogBrowse reaches 96.2% inventory recall and 88.7% masked-source recovery at 47% less source access than an exhaustive crawler, DegreeMap holds 100.0% hard feasibility while improving personalized utility by +0.066 over the strongest baseline, and the full pipeline certifies 99.5% of requests with a utility gap to the privileged gold graph of 0.015.
arXiv ID: 2608.06530
The Optimizer Is the Agent: Reasoning-Driven Search across Prompts, Programs, and ML Workflows - Junbo Li, Boyi Liu, Canwen Xu, Yite Wang, Yuxiong He, Zhangyang Wang, Qiang Liu, Zhewei Yao - [[pdf]](https://arxiv.org/pdf/2608.06714)
Abstract: Recent systems for optimizing prompts, programs, and ML workflows typically rely on explicit outer-loop controllers such as evolutionary search, bandits, or textual-gradient methods. We ask a fundamentally different question: how much of this search policy can be internalized by a single tool-using agent? We present ReASearch, a unified framework for reasoning-driven optimization in which the agent autonomously decides what to evaluate, how to diagnose failures, which edits to make, and when to verify or restart. Rather than serving only as a proposal generator guided by hand-designed heuristics, the agent actively analyzes outcomes, allocates budget, and refines its strategy over long horizons through persistent memory. With a shared agent loop and domain-specific tools, ReASearch instantiates the exact same scaffold to optimize prompts, programs, and ML workflows. Across 14 diverse tasks, it is competitive with and mostly better than specialized optimization systems, achieving gains of 2% to 40% over strong domain-specific baselines, and in some cases discovering solutions that improve on prior human best-known results. Crucially, we observe that complex search behaviors, which are typically implemented by explicit controllers, emerge naturally from the agent's reasoning process.
arXiv ID: 2608.06714
CEDAR: Agent-Orchestrated Tree Search for Goal-Directed Optimization of Complex Systems - Yingtao Tian - [[pdf]](https://arxiv.org/pdf/2608.06871)
Abstract: Complex systems, core objects of study in artificial life, model diverse phenomena through nonlinear, feedback-driven interactions that produce emergent behavior, with applications from population dynamics and biology to economic policy and strategic decision-making. Yet the difficulty of predicting how feedback structure gives rise to emergent behavior, a central open problem in artificial life, makes goal-directed design exceptionally challenging. In established practice, system structures are written in specialized modeling languages such as DYNAMO or STELLA, compounding the challenge with labor-intensive workflows that limit adoption and hinder timely decision-making. To address these challenges, we introduce CEDAR, an autonomous method that uses Large Language Model (LLM) agents to discover complex systems satisfying user-specified behavioral goals. Our key innovation is an LLM-driven Monte Carlo Tree Search (MCTS) deeply coupled with complex systems: at each iteration, an LLM Judge evaluates emergent behavior against specified goals and an LLM Editor proposes improved variants, with the Judge acting as a fitness function and the Editor as a variation operator, akin to a generate-and-evaluate loop in evolutionary computation. We represent complex systems as a restricted, runnable subset of Python with domain-specific primitives, letting LLMs modify system dynamics directly. CEDAR formalizes this as an MCTS variant with an LLM-parameterized transition kernel and value function, enabling goal-directed discovery of complex system behaviors while preserving solution diversity, and its LLM-based interpretability reveals how structural changes drive emergent behavior. CEDAR reduces human effort while enabling capabilities difficult to achieve with existing approaches, facilitating broader adoption of complex systems across domains.
arXiv ID: 2608.06871
LMM Modality Transfer: A Pre-requisite for Autonomous GIS Agents - Ivan Majic, Zexian Huang, Franziska Hübl, Krzysztof Janowicz, Meilin Shi, Mina Karimi, Zilong Liu, Alexandra Fortacz-Lazan - [[pdf]](https://arxiv.org/pdf/2608.06948)
Abstract: AI models are becoming increasingly adept at understanding and processing spatial information, thereby facilitating agentic problem-solving in spatial tasks and workflows. However, most of the research on their spatial capabilities (e.g., spatial reasoning) has focused on the textual modality as input and output. This contrasts with the human approach to GIS workflows, where text and visual modalities are often used together, interchangeably, and in a complementary manner. Thus, to truly achieve an automated GIS analysis pipeline or carry out human-designed GIS workflows, AI models --- Large Multimodal Models (LMMs) in particular --- need to be able to seamlessly transition between image- and text-based modalities that are traditionally used in such workflows. We present a modality transfer task that (1) asks an LMM to first describe an input image of colored squares in a regular grid, and (2) asks a new LMM instance to re-generate an image of the original spatial scene using the textual description output by the former model. This task quantifies the ability of LMMs to transfer spatial information between image and text modalities. Ultimately, by examining the modality transfer capability of LMMs through the lens of spatial information theory, this work highlights a critical bottleneck: achieving strong and robust geospatial understanding in LMMs requires rigorous, multi-modal alignment. Our results indicate that recent LMMs (here from OpenAI) still struggle with modality transfer, when tasked with re-generating an image of a simple spatial grid of color squares.
arXiv ID: 2608.06948
TEPA: Revoking Stale Memories for Conflict-Robust Language Agents - Yan Zhou, Yue Ouyang, Kaiyang Zheng, Suncheng Xiang - [[pdf]](https://arxiv.org/pdf/2608.07429)
Abstract: Long-term memory enables language agents to reuse past facts, preferences, and task experience. Persistence also creates a central falsifiability problem: when the world changes, stale memories can remain retrievable and pollute the prompt. We characterize this failure mode as memory pollution: degradation caused by active memories that newer conflicting evidence has superseded. We introduce TEPA, a revocable evidence-memory mechanism that makes validity an explicit state of memory. TEPA represents observations as keyed precedents and revokes active precedents when fresh evidence contradicts them under the same key, allowing retrieval to draw from current evidence while preserving revoked history for audit. Across controlled hidden-regime drift, real file-backed executable drift, and preference-update streams, revocation prevents stale active memory from remaining in the retrieval set after reversal. In controlled drift over 50 seeds, append-only and last-write-wins memory fell below no memory during full reversal (append-only and last-write-wins both 0.210, no memory 0.309, TEPA 0.950), and the same pattern reproduced under real file execution (append-only 0.203, no memory 0.298, TEPA 0.950). On clean MemoryAgentBench SH-6k, TEPA matches a strong last-write-wins cache, confirming that current-key replacement is the decisive operation for single-hop fact consolidation. Boundary tests on multi-hop and very long-context MemoryAgentBench settings expose retrieval-chain and context-selection bottlenecks beyond fact-level validity tracking. Together, these results establish lifecycle revocation as a core memory operation for agents that must falsify, audit, and later re-promote evolving knowledge.
arXiv ID: 2608.07429
Risk-Aware Decision Policies for Agents Under Noisy Perception - David Szczecina - [[pdf]](https://arxiv.org/pdf/2608.06420)
Abstract: Perception in biological systems is inherently noisy, requiring organisms to make decisions under uncertainty where misclassification can be costly or fatal. We present an Artificial Life predator-prey model of foraging under noisy perception, and compare agent performance when using various policies that take into account their noisy predictions. Through controlled experiments under both symmetric and asymmetric perceptual noise, we show that blindly trusting perceptual labels leads to catastrophic failure as noise increases, while uncertainty-aware strategies significantly improve survival and reduce fatal errors. We further observe qualitative regime shifts in behaviour, with agents transitioning from exploratory to conservative strategies as uncertainty increases. Our model links risk-sensitive foraging, ecological information use, and Artificial Life by showing that explicit information gathering can improve robustness when perception is unreliable. These results highlight the importance of uncertainty-aware decision-making and provide an interpretable artificial life analogue to robust learning with noisy labels.
arXiv ID: 2608.06420
CyberForge: Verified Vulnerability Injection at Repository Level for Cybersecurity Agent Training - Amine Lbath, Manan Suri, Aurelien Delaitre, Vadim Okun, Massih-Reza Amini, Ram D. Sriram, Dinesh Manocha - [[pdf]](https://arxiv.org/pdf/2608.06471)
Abstract: Despite recent advances, frontier large language model (LLM) agents remain limited in discovering and patching complex vulnerabilities in real-world software. Generally available agents can already aid attackers, who only need to find one exploitable weakness, while defenders must continuously identify and patch all vulnerabilities across fast-growing codebases. Stronger defensive agents would help close this gap, yet the scarcity of security training data with reproducible build and execution environments remains a bottleneck. We present CyberForge, a framework that synthesizes executable, repository-level security training data by injecting vulnerabilities into real C/C++ projects. It validates each instance dynamically: the injected build must pass the project's unit tests, and generated proof-of-vulnerability (PoV) must trigger on the injected build and not on the clean one. CyberForge is not limited by the availability of disclosed vulnerabilities, therefore it can scale in comparison to data augmentation techniques which rely on historic CVE data. The resulting corpus holds 1034 validated vulnerabilities across 80 projects and 63 weakness categories, with edit locality similar to real CVE patches under a real-versus-real noise floor. Fine-tuning on trajectories collected over this corpus improves SEC-bench patch repair by +3.3 to +14.7 points, in all six configurations of three model scales and two teachers, with the 31B student reaching its GPT-5.4-mini teacher, 72.7% against 74.0%. These gains generalize out of distribution to PatchEval, a corpus containing other programming languages, where every configuration also improves and the 31B student passes its teacher.
arXiv ID: 2608.06471
SyncSBC: Decentralized Swarm Behavior Prediction for Synchronized Autonomous Control - Varun Raveendra, Connor Mattson, Daniel S. Brown - [[pdf]](https://arxiv.org/pdf/2608.06587)
Abstract: Robot swarms utilize many independent limited-sensing agents to produce complex emergent behaviors without requiring centralized control. However, little research explores how agents can infer swarm-level behavior from purely local perception, a capability critical for detecting faults and behavior changes. In this paper, we introduce Synchronized Swarm Behavior Classification (SyncSBC), which combines improvements in machine learning and distributed consensus to classify collective swarm behavior and synchronize swarm decision-making in an entirely decentralized manner. We show that SyncSBC achieves high classification accuracy and low synchronization delay, making it suitable for real-world deployment. Finally, we use SyncSBC to demonstrate two promising swarm applications on real robots where we show that swarms utilizing SyncSBC can accurately identify anomalies in robot behavior and autonomously coordinate collective changes in swarm behavior. Videos, code and supplemental experiments are available at this https URL.
arXiv ID: 2608.06587
Online Monitoring and Corrective Steering of Programming Agents - Shuyang Liu, Saman Dehghan, Ji Young Kim, Jatin Ganhotra, Martin Hirzel, Reyhaneh Jabbarvand - [[pdf]](https://arxiv.org/pdf/2608.06701)
Abstract: Fixing GitHub issues in large-scale projects is a long-horizon task, especially when a fix requires changes across multiple locations or the issue description lacks the information needed to localize and repair it. As a result, agents traverse long trajectories that are prone to inefficiency and error: they drift away from their intended plan, repeat failed actions, or terminate without a working patch. This paper proposes LivePlan to monitor, detect, and correct such behavioral inefficiencies and drifts in real time. LivePlan decouples judging from advising: a deterministic, rule-based monitor examines general signals over the trajectory to detect issues without invoking an LLM, and only when an issue is detected does it consult an advisor LLM for a high-level, next-step correction. This design avoids the misleading re-planning and costly interventions of prior approaches. We implement LivePlan on top of SWE-agent and evaluate it using five LLMs (three as executor agents and two as advisors) across SWE-bench Verified and SWE-bench Pro. Compared to vanilla SWE-agent, LivePlan notably improves issue resolution rates, achieving consistent gains of up to 15.2% (average: 9.9%), while incurring only an additional cost of $0.08 per instance. The additional solutions concentrate on medium and hard instances. LivePlan consistently outperforms alternative approaches in resolution rate, with minimal regression on already successful runs and new successes on problems that no baseline solves.
arXiv ID: 2608.06701
Same physical state, different collective dynamics: state encodings select synchronization outcomes in language-model agents - Takahiro Ezaki, Naoto Imura, Katsuhiro Nishinari - [[pdf]](https://arxiv.org/pdf/2608.06968)
Abstract: Language-model agents act on state encodings of their environment, yet these are treated as interchangeable interfaces. Using pretrained language models, we designed a circular-synchronization experiment applying a state-encoding intervention while holding the physical system fixed: each agent sees only a summary of its neighbours' relative phases and chooses to advance, stay or retard. Encoding that state as low-order circular moments rather than as a histogram selected different collective outcomes. In GPT the moment encoding synchronized the population in 6/6 seeds and the histogram encodings in 0/6; the effect replicated in Claude but reversed direction. Replaying identical fields shifted each agent's advance/stay/retard probabilities far beyond within-encoding repeat variation, in GPT, Claude and Gemini; in GPT, presentation alone shifted the operator with the moment values fixed. State encodings therefore form part of a model-dependent effective interaction law, not a neutral interface.
arXiv ID: 2608.06968
An Agentic Hybrid Top-Down and Bottom-Up Approach to Knowledge Graph Generation - Emma Jouffroy, Warren Jouanneau, Marc Palyart - [[pdf]](https://arxiv.org/pdf/2608.07023)
Abstract: Organizing thousands of unstandardized, multilingual expertise declarations is a persistent challenge for Human Resources (HR) platforms, directly impacting downstream tasks like accurate talent matching. To address this, we propose a hybrid knowledge graph generation pipeline that grounds a Large Language Model (LLM) in the Wikidata multilingual Knowledge Graph (KG) while employing an agentic reflexion pattern to synthesize emerging concepts and their associated metadata. Unlike rigid top-down methods or fragmented bottom-up approaches, our system anchors recognized concepts to stable Knowledge Graph entities while dynamically creating new nodes and relational metadata for unrecognized skills. Executed across five stages, entity reconciliation, multilingual canonicalization, active curation, deduplication, and the iterative recovery of unmapped concepts, the system autonomously adapts to rapidly evolving, noisy skill mentions across five European languages. Ultimately, this pipeline provides a highly scalable, explicable, and self-healing framework for generating a comprehensive skills knowledge graph, from which a structured taxonomy is derived, using unstructured, noisy text.
arXiv ID: 2608.07023
Autonomous discovery of accelerator commissioning algorithms - Thorsten Hellert - [[pdf]](https://arxiv.org/pdf/2608.07138)
Abstract: Simulated commissioning has become essential for de-risking modern light-source design and commissioning, but the procedures being simulated are still designed entirely by human experts. Their labor-intensive redevelopment after lattice changes makes such studies hard to repeat and limits their use during early design iteration. This Letter demonstrates a closed research loop in which a language-model agent writes commissioning code, tests it in simulation, and improves the algorithm from the results. Applied to RF beam capture in the ALS-U accumulator-ring model, the loop substantially improves a working expert procedure and can construct a working one from a minimal starting point, with more capable models succeeding from less initial code. Extending the same framework to multiple objectives produces 16 non-dominated algorithms spanning physically distinct trade-offs between rapid beam capture and correction of seeded machine errors. This reframes commissioning studies from evaluating human-designed procedures toward a mode in which agents participate directly in discovering accelerator algorithms.
arXiv ID: 2608.07138
Trustworthy Agent Network: Trust in Agent Networks Must Be Baked In, Not Bolted On - Yixiang Yao, Yuhang Yao, Xinyi Fan, Jiechao Gao, Jie Wang, Minjia Zhang, Srivatsan Ravi, Carlee Joe-Wong - [[pdf]](https://arxiv.org/pdf/2605.19035)
Abstract: The rapid advancement of Large Language Models has given rise to autonomous LLM-based agents capable of complex reasoning and execution. As these agents transition from isolated operation to collaborative ecosystems, we witness the emergence of the Agent-to-Agent (A2A) network, a paradigm where heterogeneous agents autonomously coordinate to solve multi-step tasks. While these networks may offer better task performance compared to simply using one agent to complete the entire task, they introduce systemic vulnerabilities, such as adversarial composition, semantic misalignment, and cascading operational failures, that existing agent alignment techniques cannot address. In this vision paper, we argue that the trustworthiness of A2A networks cannot be fully guaranteed via retrofitting on existing protocols that are largely designed for individual agents. Rather, it must be architected from the very beginning of the A2A coordination framework. We present a comprehensive conceptual framework that situates trust in A2A systems through four design pillars.
arXiv ID: 2605.19035
TokTier: Exact Stateful CPU+GPU Tokenization for Agentic LLM Serving - Zhenyu Zhang, Zhichao Cao - [[pdf]](https://arxiv.org/pdf/2607.29678)
Abstract: LLM serving caches prompt KV state, yet most front ends still re-tokenize the full request on every call. Coding agents pay most: sessions repeatedly submit a long transcript after a small append, which can shift token boundaries near the end of the prior sequence. Across 153,951 calls the median append is ~1.4K characters; only 1.0-3.6% of calls start or rebuild a session, yet those carrymulti-million-character contexts. Fleet prompt-cache hit rate is 94.1%, and as it approaches 0.99, tokenization grows from 10% to 64% of time to first token (TTFT) in component measurements. TokTier is a stateful CPU+GPU tokenization service for this two-mode workload, under one contract: emitted token IDs are always identical to full reference tokenization. For session continuations it re-tokenizes a small window around the append and splices only when a per-request check finds a stable pre-tokenization boundary; failed checks widen the window or fall back to full reference tokenization. For calls without a reusable prefix it runs exact GPT-family regex pre-tokenization and BPE on a GPU. A sampled shadow verifier re-checks live traffic. Across 17 production tokenizer families, differential campaigns cover 1.5x10^10 split checks, a 12.4 TB real-text corpus, and 93,000+ replayed agent steps, with zero divergence. Incremental repair takes 0.5-1.1 ms from 100K to 3M characters, up to 437x faster than HF tokenization and 2.1x faster at 1M characters than the strongest cache-based baseline (Gigatoken) fully prewarmed. GPU tokenization encodes a 1M-character request in 0.87 ms, up to 491x below HF and 23.4x below the fastest published CPU method on the same protocol. With vLLM, median TTFT drops 16-34% and P99 TTFT 23% under recorded bursts. Under a 50 ms P99 objective, a four-core repair pool plus one GPU sustains 1,821 requests/s, where a 16-core stateless front end saturates at 40 requests/s.
arXiv ID: 2607.29678
Toward a Causal Data Management Ecosystem for Decision Making and Agentic AI - Dazhuo Qiu, Yingli Zhou, Amedeo Pachera, Angela Bonifati, Andrea Mauri - [[pdf]](https://arxiv.org/pdf/2608.07214)
Abstract: Modern AI is no longer a single model but an ecosystem: classical ML predictors, deep and multimodal models, large language models, and agents, each trained and tuned over different data sources and each producing outputs at scale that become inputs to the others. Operating such an ecosystem is fundamentally a data integration problem - the knowledge it depends on is fragmented across dozens of heterogeneous, independently governed sources that must be reconciled and continually maintained. Yet integration alone is not enough. The predictions these systems make are shaped by many interacting factors, and the events, decisions, and variables that drive an outcome are routinely entangled with the ones that merely accompany it; treated as a basis for action, such correlational signals invite confounded decisions. This becomes acute once agents act autonomously: to be trustworthy and reliable, an agent must anticipate the consequences of its actions, not merely extrapolate from what has co-occurred before. Causal reasoning is what closes this gap, distinguishing the drivers of an outcome from its correlates, and enabling prescriptive and counterfactual analysis over the ecosystem's data. We therefore argue that the integrated ecosystem needs an explicit causal layer, and we propose to build it as a shared, persistent, queryable Causal World System (CWS).
arXiv ID: 2608.07214
Reinforcement Learning (29 papers)
ADIAS: Automated Design of Interactive Agentic Systems - Lekang Jiang, Bohan Tang, Stephan Goetz, Yiwen Guo - [[pdf]](https://arxiv.org/pdf/2608.06410)
Abstract: Automated agent design improves agent harnesses through iterative revision, evaluation, and feedback summarization. Existing methods are largely candidate-centric: cross-round experience is organized around candidate agents, which leaves the repair progress implicit. This causes inefficient repair targeting, slow consolidation of partial progress, and propagation of ineffective interventions across rounds. Therefore, we formulate issue-centric agent optimization, in which repair progress is carried forward as an explicit persistent issue state to guide optimization, rather than re-derived from candidate history in each round. We instantiate the formulation in ADIAS, a framework for automated full-code agent design with two mechanisms. A persistent issue state maintains stable issue identities, lifecycle status, supporting evidence, and intervention-outcome histories. Issue-guided optimization uses this state to jointly propose repair targets and revision directions for subsequent focused full-code modification. Across five interactive benchmarks, ADIAS outperforms the strongest baseline by 25.2% on average and achieves consistent gains across four backbone models. Controlled ablations further show that removing persistent issue state or replacing issue-centric revision with candidate-centric policies leads to performance drops of up to 40.7%.
arXiv ID: 2608.06410
Vehicle routing problem using deep reinforcement learning - A case study about truck planning in the industry - Siliang Lu, Dan Hu, Lili Wu - [[pdf]](https://arxiv.org/pdf/2608.06668)
Abstract: As an important component of the supply chain industry, transportation has experienced rapid development in the past decade with the assistance of digital platforms and intelligent algorithms. Within the field of transportation research, Vehicle Routing Problem (VRP) has remained a persistent and enduring challenge. In the realm of management science, experts, and scholars from both the industrial and academic sectors have continuously explored optimization models and algorithms to effectively address routing problems, from the classical Traveling Salesman Problem to the more general Vehicle Routing Problem. These models and algorithms are applied in real-world industrial scenarios to achieve cost optimization and reduce carbon footprints. However, due to the complexity of real-world problems, numerous specific constraints are often added, and challenges such as information opacity, uncertainty, and irrational human behavior may arise. Therefore, deploying and optimizing mathematical models for VRP in practical scenarios while maintaining optimal results poses numerous challenges. This paper discusses and provides solutions for three different logistic use cases involving external truck network design. Through these industrial case study, the paper introduces how deep reinforcement learning-based vehicle routing optimization has been implemented. As a result, it can be observed that the routes optimized by reinforcement learning agent have over 10% total cost compared to baseline results. Furthermore, the paper proposes that in future research, DRL algorithms for vehicle routing problems could be generalized into more variations of VRP.
arXiv ID: 2608.06668
Capek 0.5: An Execution-Centric Vision-Language Model for Embodied Intelligence - Ying Chen, Weizhen Li, Zhe Hu, Zhenjiang Li, Rui Jiang, Zhifeng Gu, Lihuang Fang, Jiangping Liu, Lei Yi, Jie Chen - [[pdf]](https://arxiv.org/pdf/2608.06756)
Abstract: Vision-language models are increasingly serving as the reasoning core of embodied agents. Robot execution is inherently iterative: each action reshapes the scene and physical state, continually renewing what must be perceived, reasoned about, and verified. Meeting these demands requires complementary capabilities that differ in supervision signals, prediction formats, and verification criteria. Existing approaches typically develop these capabilities against isolated, task-specific objectives, leaving open how they should be organized and integrated around execution as a whole. We present Capek 0.5, an embodied vision-language model built around an execution-centric capability taxonomy. Rather than organizing training by datasets or tasks, the taxonomy groups embodied capabilities according to their functional roles throughout execution and comprises four capability families: Spatial Reasoning, Temporal Understanding, Action Guidance, and State Verification. Each capability is first acquired by a dedicated specialist through reinforcement learning with verifiable rewards from a shared backbone, and the specialists are then consolidated into a single inference-time model through weight-space merging followed by routed policy-space distillation. We instantiate Capek 0.5 at the 2B and 35B-A3B scales and evaluate it from three complementary perspectives: comprehensive benchmark suites including Capek-StateBench, a new benchmark for state verification; a controlled study of capability retention from specialists to the unified model; and closed-loop evaluation in simulated embodied environments. Capek 0.5 improves the large majority of matched benchmark rows over its initialization, retains all four specialized capabilities in one checkpoint with quantified losses, and transfers to closed-loop embodied task execution.
arXiv ID: 2608.06756
Long-Horizon Agent Trajectory Attribution: A Unified Benchmark and Fine-Grained Annotation Framework - Jing Chen, Yang Sun, Li Zhang, Lin Xu, Jie Shi - [[pdf]](https://arxiv.org/pdf/2608.06909)
Abstract: Large language model (LLM) agents increasingly operate through long-horizon trajectories involving user instructions, tool use, external observations, and memory. Existing benchmarks primarily evaluate behavioral outcomes but provide limited support for fine-grained attribution analysis. We introduce trajectory attribution and develop a benchmark and annotation framework for this task. The benchmark organizes heterogeneous trajectories under a unified component schema and provides annotations of the primary attribution component, together with attack and execution chains where applicable. Instantiating the benchmark with trajectories from AgentDojo and the Stage and Canary settings of Agent3Sigma yields more than 1,300 annotated trajectories covering task-aligned actions, unsafe actions, and safety refusals. The benchmark defines two evaluation tasks, primary attribution localization and attribution-chain recovery, and provides reference baselines based on incremental trajectory contribution and component-level leave-one-out perturbation. It captures diverse attribution settings, including local and long-range attribution as well as structured attribution chains. Reference baseline results exhibit substantial performance differences across these settings, providing an initial characterization of the benchmark's attribution challenges. Beyond this initial instantiation, we release a reusable annotation skill that enables trajectories generated by new agent models to be standardized, annotated, and evaluated under the same framework. Project resources and future releases are available at this https URL.
arXiv ID: 2608.06909
MemOPD: On-Policy Distillation through Memory State Alignment for Long-Horizon Agents - Zhiyuan Liu, Tinghong Ye, Chenghao Liu, Yizhuo Li, Songfang Huang - [[pdf]](https://arxiv.org/pdf/2608.07068)
Abstract: Long-horizon agents accumulate growing contexts during interaction, impairing performance and stability. Compact memory mitigates this problem by compressing and rewriting the history retained between model invocations. Learning what to retain typically relies on proximal policy optimization (PPO) with final task rewards, but sparse rewards provide little guidance for individual memory updates. This limitation motivates on-policy distillation (OPD), which supplies dense teacher supervision on student rollouts. For such supervision to be valid, the teacher must evaluate each sampled action under the same state in which it was generated. However, the context rewriting performed during memory compression can break this alignment. When sampled responses are retained and re-encoded for later invocations, flattening the interaction into a persistent history may cause the teacher to score the action under a state that the student never visited during rollout. The action therefore remains on-policy by provenance, but not necessarily by state. We therefore propose Memory-Aligned On-Policy Distillation (MemOPD). MemOPD records the inputs and sampled outputs of each model invocation, restores its original token positions and causal visibility, and packs the reconstructed invocations for efficient teacher scoring. The teacher provides full-vocabulary supervision at the sampled action positions, while PPO preserves the final task objective. Experiments verify state alignment across several context updates and show that it improves F1 by 7.0% over persistent-history teacher scoring in a matched control. Overall, MemOPD-3B improves F1 over PPO by up to 416.2%, while packing yields up to a 1.63x speedup in actor computation during training. The code for this work is publicly available at: this https URL.
arXiv ID: 2608.07068
How Much, Then Where: Credit-Conserving Action-to-Token Allocation for Multi-Turn Agent Reinforcement Learning - Lichao Ma, Yang Sun, Shuaitao Zhao, Yangyi Fang, Cong Qin, Xiaoliang Fu, Yuhang Tian, Yuchen Wei, Junbo Zhu, Yang Wei, Lu Pan, Jiaye Lin - [[pdf]](https://arxiv.org/pdf/2608.07118)
Abstract: Credit assignment in multi-turn agent reinforcement learning operates at two levels: assigning trajectory-level credit to actions and distributing each action's credit across its tokens. In this paper, we introduce FACTOR, which separates these decisions. FACTOR uses checkpoint-calibrated TD residuals to assign per-action credits that telescope to the trajectory advantage, and feedback-conditioned teacher-student likelihood gaps to allocate each credit across the realized action tokens. Per-action normalization preserves the action-average coefficient and prevents token-level sign flips. We pair this construction with an action-mean reduction, removing the implicit dependence of an action's scalar surrogate weight on its token length. At the behavior policy and before clipping, each action's inner action-mean surrogate equals its TD credit. FACTOR consistently improves over competitive baselines across ALFWorld, WebShop, and ScienceWorld, with every environment-seed comparison favoring FACTOR and the largest gains emerging on the longest-horizon environment. The same hyperparameters transfer without retuning to a larger backbone and to a different model family. Ablations identify TD action credit as the dominant driver of the improvement, with hindsight token allocation contributing complementary gains.
arXiv ID: 2608.07118
DiDPO: Diff-in-Diff Policy Optimization for Coding Agent Training - Xucong Wang, Zhe Zhao, Liheng Yu, Di Wu, Xiaofeng Cao, Pengkun Wang - [[pdf]](https://arxiv.org/pdf/2608.07147)
Abstract: Reinforcement learning with Verifiable Reward (RLVR) has emerged as a powerful paradigm for training coding agents, where the execution feedback from compilation and tests provides objective verification. However, unlike agent tasks, coding agents face a unique and finer-grained credit assignment challenge: at each step, coding actions simultaneously pack varying changes into different regions of a code version, which makes the contribution of independent change indistinguishable. Existing RLVR methods mostly leverage the outcome reward or step-level reward, which fails to dive into a code diff and makes unique properties of coding actions invisible to training. In this paper, we propose Diff-in-Diff Policy Optimization (DiDPO), a critic-free RL method that constructs fine-grained credit units directly from the structure of code diffs. DiDPO organizes multi-turn coding interactions into multiple thought--action steps and discovers code diffs across sampled trajectories. It then selects anchors by aggregating highly similar sub-diffs split from each whole diff by our ``groupability score'', which provides the splitting schema that optimally balances the semantic scope of anchors and the group mass they may form. Finally these anchors form advantage groups and project the diff-level advantage back to individual response tokens. Experiments on long-horizon coding and reasoning benchmarks show that DiDPO significantly outperforms strong agentic RL baselines. On Qwen2.5-7B-Coder, DiDPO exceeds comparable methods by over 10% and narrows the gap with far larger models, offering a principled framework for fine-grained credit assignment in coding agent training. We also open-source verl-code, an agentic rl codebase that supports various RL methods and coding benchmarks.
arXiv ID: 2608.07147
ResidencyRL: Reinforcement Learning in Simulated Clinical Environments - Valentin Liévin, Samuel Schmidgall, Tim Strother, Alex Bijamov, Akshay Goel, Anil Palepu, Chunjong Park, Vahid Balazadeh, Min Woo Sun, Marius Guerard, Justin Chen, Dave Steiner, Vikram Dhillon, Ibrahim Azar, Akhil Mehta, Nicholas Spetsieris, Shilpan Shah, Maen Abdelrahim, Amit Dahiya, Yun Liu, Katherine Chou, Yossi Matias, Avinatan Hassidim, Dale R. Webster, Quoc V. Le, Raia Hadsell, Joelle Barral, Carey Radebaugh, Aleksandra Faust, Shekoofeh Azizi, Mike Schaekermann, Po-Hsuan Cameron Chen, Tao Tu, David Racz, Lin Yang - [[pdf]](https://arxiv.org/pdf/2608.07418)
Abstract: In medical education, physicians convert academic knowledge into clinical expertise through residency: years of training across thousands of encounters, with diverse sources of feedback and progressively greater autonomy. Much of clinical reasoning relies on the patient encounter, a dialogue in which a clinician elicits history, refines diagnostic hypotheses, and decides management under uncertainty. While large language models (LLMs) excel on static medical benchmarks, methods to optimize the full sequence of clinical decisions remain underdeveloped. We present ResidencyRL, a reinforcement learning (RL) method for training clinical artificial intelligence (AI) agents through simulated multi-turn clinical encounters (up to 60 dialogue turns and 8 tool calls per trajectory). ResidencyRL pairs the policy agent with LLM simulators capable of complex, adversarial behaviors, training against a structured reward aligned to diagnostic accuracy, management quality, communication, documentation, and safety. On held-out evaluations, the ResidencyRL agent improves diagnostic accuracy by 7.0% under adversarial conditions (88.0% vs. 81.0%) and reduces missed red flag rates by 31%, demonstrating rigorous mitigation of premature closure. Blinded expert clinicians validated these gains, preferring the trained agent in 87.6% of side-by-side comparisons. The procedural competencies transfer to unseen benchmarks: the agent outperforms the base model across all six clinical axes of the AMIE multi-visit benchmark, and shows consistent directional improvements on AgentClinic and CRAFT-MD. Our findings demonstrate that sequential clinical decision-making can be effectively learned through multi-turn RL in simulation, yielding robust, generalizable capabilities, paving the way towards clinical mastery. Prospective validation with real-world workflows remains necessary to establish clinical utility.
arXiv ID: 2608.07418
Multimodal Drivers' Emotion Recognition and Safety-Oriented Intervention for Intelligent Transportation Systems - Chang Liu, Dalai Mengke, Hanbo Zhou, Jia Hu, Peter Mihajlik, Tamas Sziranyi - [[pdf]](https://arxiv.org/pdf/2608.06378)
Abstract: Driver emotions can affect risk perception, decision-making, and vehicle control under complex road conditions. Existing studies mainly focus on driver emotion recognition, while limited attention has been given to context-aware intervention that jointly considers driver emotion and road perception. This paper proposes a safety-prioritized multimodal driver assistance framework that analyzes speech-derived emotional cues and visual road conditions to generate structured driving interventions. The framework first provides road safety reminders and then generates emotion-aligned verbal support. We construct a multimodal dataset by aligning emotional speech signals with structured road environment descriptors and introduce the CARE (Context-Aware Road-Emotion Evaluation) score to jointly evaluate emotion recognition, risk identification, and intervention generation. Experimental results show that the proposed framework balances environmental risk reporting and emotion-aware verbal regulation, providing a feasible safety-driven direction for intelligent transportation systems.
arXiv ID: 2608.06378
Evaluating XAI Support From A Hierarchical Reinforcement Learning Policy in Human-Agent Collaboration - Mateus Levi Simões Fernandes, Alberto Sardinha - [[pdf]](https://arxiv.org/pdf/2608.06381)
Abstract: Explainable AI (XAI) has shown promise for human-agent collaboration, yet results rely on hand-crafted policies in custom environments, limiting generalizability to state-of-the-art teaming research. We provide the first systematic evaluation of XAI support generated from an intrinsically explainable learned policy in an established benchmark. Using the Hierarchical Ad Hoc Agents (HA$^2$) architecture in Overcooked-AI, we generate real-time explanations from hierarchical subtask selections, delivered through text or audio via a novel trigger-based system. Our between-subjects experiment (n=38) found no significant performance effects, though participants with explanations showed trends toward faster performance improvement. More notably, audio explanations produced a significant reduction in participants' working-alliance bond with the agent -- an effect absent under the text modality -- suggesting that spoken explanations activate partnership expectations the underlying reactive policy cannot meet. We provide the first modality comparison in real-time human-agent collaboration and establish a baseline methodology for evaluating intrinsically explainable reinforcement learning architectures in benchmark environments. Results point to matching explanation modality to the underlying policy's capacity of sustaining the partnership its delivery implies as a potential path for more effective collaborative XAI.
arXiv ID: 2608.06381
Agentic AI: User Empowerment or Enclosure? - David Gamba, Daniel M. Romero, Grant Schoenebeck - [[pdf]](https://arxiv.org/pdf/2608.06510)
Abstract: Agentic AI promises a more flexible form of digital agency: systems that can act on users' behalf, from filtering content to negotiating prices to selecting services. Whether it will empower users is an open question, and we argue that the answer depends on more than the technology. We conduct a comparative case analysis of four more mature domains where similar forms of agency arose: browser-based ad blockers, platform recommender systems, financial robo-advisors, and email spam governance. Across the cases, decisions about whose interests agents would serve were resolved through technical arrangements: API choices, protocol governance, industry standards, and default configurations. Beyond their technical form, these were political decisions. We identify this as depoliticization, a concept from political theory, here at work in technological systems. Its most consequential effect is that individual outcomes and collective contestation capacity can move in opposite directions: spam inbox quality improved substantially while the organized capacity to contest spam governance collapsed. Where intermediary institutions sustained adversarial challenge, user-aligned agency proved more durable; where proprietary infrastructure and closed standard-setting absorbed contestation, displacement compounded. We apply this to agentic AI, where governance arrangements consolidating around the Model Context Protocol and the Agentic AI Foundation are settling these configurations before the choices that define what agents can do move outside the reach of users and the public.
arXiv ID: 2608.06510
Flowing Through States: Neural ODE Regularization for Reinforcement Learning - Mohamed Ghanem, Bernd Finkbeiner - [[pdf]](https://arxiv.org/pdf/2608.06595)
Abstract: Neural networks applied to sequential decision-making tasks typically rely on latent representations of environment states. While environment dynamics dictate how semantic states evolve, the corresponding latent transitions are usually left implicit, creating a potential misalignment between the two. We propose to model latent dynamics explicitly by drawing an analogy between Markov decision process (MDP) trajectories and ordinary differential equation (ODE) flows: in both cases, the current state fully determines its successors. Building on this view, we introduce a neural ODE-based regularization method that enforces latent embeddings to follow consistent ODE flows, thereby aligning representation learning with environment dynamics. Although broadly applicable to deep learning agents, we demonstrate its effectiveness in reinforcement learning by integrating it into Actor-Critic algorithms. Our approach yields major performance gains across various standard Atari benchmarks for A2C and gridworld environments for PPO.
arXiv ID: 2608.06595
Beyond Isolation: Unlocking Reinforcement Learning Component Synergy for Sample-Efficient Continuous Control - Qi Zhao, Guozheng Ma, Yilun Kong, Lu Li, Haoyu Wang, Zilin Wang, Tiantian Zhang, Yuxing Wang, Jian Sha, Yongzhe Chang, Xueqian Wang, Dacheng Tao - [[pdf]](https://arxiv.org/pdf/2608.07086)
Abstract: Reinforcement learning systems are significantly more complex than other machine learning paradigms due to inherent properties, causing RL system design to jointly account for many tightly coupled factors. Despite advances in individual algorithmic components, their functional interdependencies remain underexplored: do they exhibit mutual synergy or counterproductive interference? To bridge this gap, we conduct a systematic investigation and find that the efficacy of different components exhibits significant task-dependency, and naively stacking state-of-the-art techniques does not necessarily yield performance gains; instead, it often triggers emergent challenges, such as compounded non-stationarity. Building upon these findings, we distill a suite of actionable insights into the principled coordination of these components. Guided by these insights, we propose ROSER, an RL framework that coordinates three critical dimensions: Model-based Representation, Optimization Stability, and Experience Replay. Across diverse continuous-control benchmarks, ROSER consistently outperforms vanilla baselines and achieves 17.60% gains over naive stack. Our findings underscore the necessity of a holistic perspective in RL system design and paves the way for developing sample-efficient agents.
arXiv ID: 2608.07086
Interpretable reinforcement learning with decision-tree pruning - Mark Leon Ringer, Michel Tokic - [[pdf]](https://arxiv.org/pdf/2608.07151)
Abstract: Reinforcement learning policies are difficult to inspect, but interpreting them is a prerequisite for trustworthiness. Converting a trained policy into explicit decision-tree rules improves transparency and the resulting artifacts often remain too complex for human understanding. We present a pruning process that simplifies such rule-based policies while preserving task performance and making edits to the policy auditable. The process defines a small set of structural and usage-aware operators and evaluates candidate edits by re-executing the policy to measure return and interpretability proxies. This exposes an transformation process from complex to compact policy structures. We investigate this approach on classic control and MuJoCo benchmarks, where pruning traces reveal consistent interpretability improvements while maintaining high performance.
arXiv ID: 2608.07151
Momba: Network Modernization Improves Multi-Objective Reinforcement Learning - Adam Štafa, Santeri Heiskanen, Petr Novotný, Joni Pajarinen - [[pdf]](https://arxiv.org/pdf/2608.07180)
Abstract: Recent advances in deep reinforcement learning (RL) have shown that improving neural network architectures can yield substantial gains in sample efficiency and asymptotic performance without altering the underlying algorithms. In contrast, work on multi-objective reinforcement learning (MORL), which aims to discover a set of policies that balance trade-offs among conflicting objectives, has predominantly focused on algorithmic innovations, leaving the area of architectures underexplored. While the optimal policies and value functions can differ significantly depending on the trade-offs, MORL algorithms commonly represent them with simple feedforward networks conditioned on the trade-off. This raises the question of whether the performance of the algorithms could be improved with more expressive function approximators. In this paper, we integrate recent advances in neural network design: (i) observation and feature normalization, (ii) weight normalization, and (iii) modeling of distributional returns with an entropy-regularized MORL algorithm. The empirical results across standard continuous control benchmarks demonstrate that these changes substantially improve the quality of the produced solution sets without requiring major changes to the underlying algorithm.
arXiv ID: 2608.07180
Homebot: A Personal AI Agent for Conversational Home Assistance and Automation - Shengyuan Ye, Yixin Zhang, Han Liang, Liekang Zeng, Jiangsu Du, Mu Yuan - [[pdf]](https://arxiv.org/pdf/2608.02254)
Abstract: \texttt{Homebot} is a locally deployable AI agent for conversational household assistance and automation. It accepts voice and instant-messaging requests through a shared runtime that combines language-model responses with registered tools and task-specific skills. The design separates common request processing from session ownership: messaging history remains scoped to a channel and chat, whereas voice interaction is bounded by wake-word activation. For hands-free use, \texttt{Homebot} combines local wake-word detection, streaming speech recognition and synthesis, and an explicit dialogue-state protocol for ending, following up, or continuing a conversation. Clear channel, tool, and skill contracts support practical customization for household use.
arXiv ID: 2608.02254
SkillTrace: Multi-Trace Provenance Auditing for LLM-Agent Skill Reuse - Jialuo Chen, Minghe Wang, Lingqi Jiang, Jianan Ma, Xinhao Deng, Xiaohu Du, Ruixiao Lin, Yunhao Feng, Linkang Du, Jingyi Wang - [[pdf]](https://arxiv.org/pdf/2608.05204)
Abstract: LLM-agent ecosystems are rapidly growing around reusable skills: mixed-modality packages of metadata, natural-language instructions, code, tools, references, and operational workflows. As skills become marketplace artifacts, auditing their reuse is no longer the same problem as ordinary code clone detection. Existing detectors target single-modality source code or whole-package similarity, yet skill reuse evidence is distributed across authored text, implementation fragments, and operational structure. As a result, they can miss reuse that preserves only one part of a skill. We present SKILLTRACE, a multi-trace provenance auditing framework for LLM-agent skill reuse. SKILLTRACE extracts three provenance traces: Expression, Implementation, and Operational. It represents the Operational Trace as a Skill Operational Graph (SOG) that captures activation, procedure, and resource-flow structure. An LLM assists only the Operational-trace extraction, once at ingestion; at audit time SKILLTRACE compares cached traces deterministically, calibrates each trace against same-function strict negatives, and reports which trace supports a reuse decision. On SKILLTRACE-BENCH, with 820 transformed reuse positives over 100 marketplace anchors and 751 negative controls, SKILLTRACE achieves AUROC 0.938 and F1 0.898. A 36,446-skill wild audit further shows that trace-attributed evidence surfaces actionable reuse review queues beyond repository-level baselines.
arXiv ID: 2608.05204
Contextual Information Policy Optimization for Search Agents - Xingyu Guo, Wei Chen, Linlin Yang, Baochang Zhang - [[pdf]](https://arxiv.org/pdf/2608.06128)
Abstract: Search agents extend large language models beyond static parametric memory by enabling them to acquire and use external evidence during multi-step reasoning. For knowledge-intensive tasks involving complex or evolving information, their reliability depends not only on retrieving relevant evidence but also on using it to guide subsequent reasoning. However, existing methods primarily reward final-answer correctness or intermediate progress, without directly assessing whether post-retrieval actions are grounded in the retrieved evidence. This misalignment encourages prior-driven reasoning: agents form conclusions based on internal knowledge and use retrieval mainly to confirm them, resulting in confirmation bias and inefficient evidence use. To address this issue, we propose Contextual Information Policy Optimization (CIPO), an evidence-oriented reinforcement learning framework that explicitly aligns policy optimization with external evidence use. CIPO assigns dense, turn-level credit to reasoning actions influenced by retrieved information, while combining this evidence-use signal with a global outcome reward to preserve answer correctness. With this manner, CIPO discourages evidence-detached guesses and promotes reasoning trajectories in which retrieved facts can guide or revise subsequent reasoning. Importantly, CIPO requires neither human process annotations nor an additional reward model. Extensive experiments on seven in-domain and out-of-domain benchmarks show that CIPO reduces the prevalence of prior-driven reasoning and achieves excellent performance on most tasks.
arXiv ID: 2608.06128
Beyond Top-K: Replacing Black-Box Retrieval with Interpretable Agentic Operations - Sagar Tamang, Ayush Vyas, Tabarakul Hazarika - [[pdf]](https://arxiv.org/pdf/2608.06305)
Abstract: Retrieval-augmented generation over long documents is dominated by one design: chunk the text, embed the chunks, and surface the top-k nearest neighbours of the query. We argue that for an important class of documents -- financial statements, audit reports, regulatory returns -- this design is structurally unsound, and we make the argument measurable. On a 780-page government financial report, 86.8% of content lines are table rows, thousands of near-identical figures compete in one embedding space, and a figure inherits its unit from a header a median of 13 lines above it -- so a chunk boundary routinely separates a number from whether it is in lakh or crore, an error of two orders of magnitude. A table-aware chunker built as a steelman fixes the unit problem but leaves 27-30% of numeric chunks with no fiscal-year header at every chunk size we tried. We propose READ (Reliable Embedding-free Agentic Document-search), in which an agent reads the raw document through three deterministic operations -- normalized lexical search, structural navigation, and bounded span reads -- exposed over the Model Context Protocol, so a trajectory is a replayable audit trail, not an opaque similarity score. On 51 verified questions READ answers 58.8% against dense retrieval's 15.7% (p_Holm = 2 x 10^-5) -- or 35.3% tuned, which READ still leads by 23.5 points (p_Holm = 0.017). An agent given the same loop but a top-k tool reaches only 27.5%, locating the gain in the interface rather than in iteration. We also report what the evidence does not support: BM25 is statistically indistinguishable from READ, so our result separates embedding-based from embedding-free retrieval, not agentic from lexical search.
arXiv ID: 2608.06305
Kimi K2.5: Visual Agentic Intelligence - Kimi Team, Tongtong Bai, Yifan Bai, Yiping Bao, S.H. Cai, Yuan Cao, Ziwei Chai, Y. Charles, H.S. Che, Cheng Chen, Guanduo Chen, Huarong Chen, Jia Chen, Jianlong Chen, Jun Chen, Kefan Chen, Liang Chen, Ruijue Chen, Xinhao Chen, Yanru Chen, Yanxu Chen, Yicun Chen, Yimin Chen, Yingjiang Chen, Yuankun Chen, Yujie Chen, Yutian Chen, Zhirong Chen, Ziwei Chen, Dazhi Cheng, Yean Cheng, Minghan Chu, Jialei Cui, Jiaqi Deng, Muxi Diao, Hao Ding, Mengfan Dong, Mengnan Dong, Yuxin Dong, Yuhao Dong, Angang Du, Chenzhuang Du, Dikang Du, Lingxiao Du, Yulun Du, Yu Fan, Shengjun Fang, Qiulin Feng, Yichen Feng, Garimugai Fu, Kelin Fu, Hongcheng Gao, Tong Gao, Yuyao Ge, Shangyi Geng, Chengyang Gong, Xiaochen Gong, Zhuoma Gongque, Qizheng Gu, Xinran Gu, Yicheng Gu, Longyu Guan, Shuhao Guan, Yuanying Guo, Xiaoru Hao, Dailan He, Tianhong He, Weiran He, Wenyang He, Yibo He, Yunjia He, Chao Hong, Hao Hu, Jiaxi Hu, Yangyang Hu, Zhenxing Hu, Ke Huang, Ruiyuan Huang, Weixiao Huang, Zhiqi Huang, Chaobo Jia, Tao Jiang, Zhejun Jiang, Xinyi Jin, Yu Jing, Guokun Lai, Aidi Li, C. Li, Cheng Li, Fang Li, Guanghe Li, Guanyu Li, Haitao Li, Haoyang Li, Jia Li, Jingwei Li, Junxiong Li, Lincan Li, Mo Li, Weihong Li - [[pdf]](https://arxiv.org/pdf/2602.02276)
Abstract: We introduce Kimi K2.5, an open-source multimodal agentic model designed to advance general agentic intelligence. K2.5 emphasizes the joint optimization of text and vision so that two modalities enhance each other. This includes a series of techniques such as joint text-vision pre-training, zero-vision SFT, and joint text-vision reinforcement learning. Building on this multimodal foundation, K2.5 introduces Agent Swarm, a self-directed parallel agent orchestration framework that dynamically decomposes complex tasks into heterogeneous sub-problems and executes them concurrently. Extensive evaluations show that Kimi K2.5 achieves state-of-the-art results across various domains including coding, vision, reasoning, and agentic tasks. Agent Swarm also reduces latency by up to
arXiv ID: 2602.02276
Trajectory-Relative Hindsight Distillation for Agentic Reinforcement Learning - Haoyu Zheng, Yun Zhu, Qing Wang, Wenqiao Zhang - [[pdf]](https://arxiv.org/pdf/2608.07371)
Abstract: Recent agentic reinforcement learning methods use hindsight to complement sparse outcome rewards. However, a completed rollout can yield many such signals, leaving their appropriate allocation across turns unclear. We introduce TRIAL, a trajectory-relative hindsight distillation framework with a unified turn-aligned scoring protocol. For each decision turn, TRIAL extracts an outcome view of that decision's realized consequence and evaluates the same response under ordinary and hindsight-conditioned contexts. The signed log-probability gap determines the direction and local strength of token-level supervision, while turn-level magnitudes are normalized jointly over the realized trajectory. The resulting allocation multipliers have an eligible-token-weighted mean of one, redistributing dense supervision across turns while fixing its average multiplier. Experiments on WebShop and ALFWorld with different backbones show that TRIAL outperforms GRPO across all eight combinations of backbone, environment, and evaluation metric, while achieving the best or tied-best performance among six methods on six of them. On WebShop with Qwen3-1.7B, TRIAL improves the success rate from 56.4% to 75.2% and the task score from 78.7% to 85.7%. Controlled ablations further show that trajectory-relative turn allocation provides substantial gains beyond those of dense hindsight distillation alone.
arXiv ID: 2608.07371
Joint Optimization of Reasoning and Dual-Memory for Self-Learning Diagnostic Agent - Bingxuan Li, Simo Du, Yue Guo - [[pdf]](https://arxiv.org/pdf/2604.07269)
Abstract: Clinical expertise improves not only by acquiring medical knowledge, but by accumulating experience that yields reusable diagnostic patterns. Recent LLMs-based diagnostic agents have shown promising progress in clinical reasoning for decision support. However, most approaches treat cases independently, limiting experience reuse and continual adaptation. We propose SEA, a self-learning diagnostic agent with cognitively inspired dual-memory module. We design a reinforcement training framework tailored to our designed agent for joint optimization of reasoning and memory management. We evaluate SEA in two complementary settings. On standard evaluation with MedCaseReasoning dataset, SEA achieves 92.46% accuracy, outperforming the strongest baseline by +19.6%, demonstrating the benefit of jointly optimizing reasoning and memory. On the long-horizon with ER-Reason dataset, SEA attains the best final accuracy (0.7214) and the largest improvement (+0.35 Acc@100), while baseline methods show limited or unstable gains. Expert evaluation further indicates that rules consolidated from SEA show strong clinical correctness, usefulness and trust, suggesting that the induced rules in dual-memory module are reliable and practically meaningful. Overall, SEA improves both diagnostic reasoning ability and continual learning by effectively transforming experience into reusable knowledge.
arXiv ID: 2604.07269
CrystalGRPO: Target-Aligned and Coverage-Preserving Reinforcement Learning for Flow-Based Crystal Structure Prediction - Kaixiang Su, Hongfei Xue, Qiang Zhu - [[pdf]](https://arxiv.org/pdf/2608.06582)
Abstract: Flow-based generative models can efficiently produce candidate structures for crystal structure prediction (CSP), but their pretrained objectives do not directly optimize downstream target recovery. Reinforcement-learning post-training offers a flexible solution, yet existing approaches rely primarily on energy rewards and coordinate-only stochastic policies. Predicted energy does not identify the reference polymorph, while reward-driven concentration can reduce the candidate coverage required for Top-N recovery. We introduce CrystalGRPO, a CSP-aligned post-training framework that extends existing ODE-to-SDE policy constructions to the joint coordinate--lattice state. CrystalGRPO combines MACE-predicted energy with a StructureMatcher-based recovery score and provides two operating modes: CrystalGRPO-Q, which prioritizes single-draw recovery, and CrystalGRPO-C, which combines full-trajectory reference regularization with a coverage-aware group advantage to preserve finite-budget target recovery. Across MP-20 and MPTS-52 with PXRDGen and OMatG backbones, both variants reduce one- and twenty-sample RMSE relative to coordinate-only reinforcement in all four backbone--dataset settings. CrystalGRPO-Q consistently improves Top-1, whereas CrystalGRPO-C achieves a higher Top-20 across all settings.
arXiv ID: 2608.06582
Game-Theoretic Inverse Reinforcement Learning for Modeling Competitive Human Driving: A Cut-in Prediction Study - Yu Song - [[pdf]](https://arxiv.org/pdf/2608.06445)
Abstract: Capturing the strategic decision-making inherent in competitive human driving is critical for autonomous vehicle safety and traffic simulation. This study demonstrates that game-theoretic Inverse Reinforcement Learning (IRL) provides a robust framework for this challenge. We present a comprehensive analysis comparing data-driven IRL models against an established physics-based game-theoretic approach for predicting aggressive, safety-critical cut-in lane changes. Using the high-fidelity highD dataset, we systematically develop and evaluate a series of IRL models with increasing feature complexity. Our results reveal significant advantages: the best-performing IRL models achieve an overall prediction accuracy exceeding 75 percent while maintaining a Cut-In precision up to 51 percent and recall up to 49 percent. This represents a significant improvement over the established physics-based benchmark, which achieved only 4.4 percent precision in these high-stakes scenarios. The analysis reveals a clear trade-off: incorporating granular, instantaneous features yields higher precision, while adding temporal consistency features maximizes recall. These findings suggest that IRL-based models can effectively bridge the gap between microscopic driver intent and macroscopic safety outcomes, providing a more reliable foundation for modeling interactions in mixed-autonomy environments.
arXiv ID: 2608.06445
Spatiotemporal Agility: Time-Constrained Reinforcement Learning for Vision-Guided Dynamic Quadrupedal Interception - Yidong Zhu, Zibo Dai, Tongning Zhang, Leixin Chang, Hua Chen - [[pdf]](https://arxiv.org/pdf/2608.06907)
Abstract: Legged robots require robust agility to perceive and interact with complex and dynamic environments within a constrained time. However, most existing quadruped locomotion works rely on velocity-tracking policy, which struggle to reach precise targets within strict temporal constraints. Moreover, integrating real-time perception with agile locomotion for highly dynamic targets remains challenging due to sensor latency and processing delays. To concretely study and benchmark such agility in dynamic settings, we introduce a challenging ball-catching task for legged robots. This paper proposes an integrated framework that combines a vision module for landing point and time prediction with a direct position and time conditioned RL locomotion policy, instead of intermediate velocity commands. Beyond the method design, this work presents a system-level contribution that completes real-time robotic interception system that integrates multi-camera perception, online trajectory prediction, low-latency target communication, and sim-to-real locomotion control into a closed-loop deployment pipeline. By explicitly predicting the future spatial-temporal target, our approach mitigates perception latency during dynamic interception. We conducted extensive ball-catching experiments for the legged robot. Through comparative experiments against a velocity-tracking baseline, our direct target-conditioned approach achieves a higher success rate in catching balls with predicted landing spots within 2 meters and flight times between 0.8 and 1.2 seconds. This shows that the robot has successfully completed the dynamic ball-catching task under our tested setup. Furthermore, our policy exhibits a smaller performance gap after deployment, suggesting improved sim-to-real behavior in these trials.
arXiv ID: 2608.06907
Progress-Certified Reversible Simplex Supervision of Goal-Reaching Reinforcement Learning - Mehdi Heydari Shahna, Jouni Mattila - [[pdf]](https://arxiv.org/pdf/2601.19499)
Abstract: Task completion is difficult to certify when state aggregation, model mismatch, and disturbances invalidate nominal RL transitions. We present a reversible Simplex framework combining a frozen finite-state policy with goal-level robust-adaptive recovery and explicit accounting for post-action progress debt. Outward-rounded reachability verifies disturbed transitions, bounds debt from rejected learned actions, and constructs robust recovery and re-entry sets. If an independent checker accepts both certificates and the recovery decrement exceeds the verified debt, every completed switching cycle decreases storage, yielding finite switching and finite-sample goal entry. In a restricted end-to-end instance, the checker resolves all retained obligations, certifies
arXiv ID: 2601.19499
Social Facilitation of Creative Reflection: AI-agents and Humans - Olga Sutskova, Corey Ford - [[pdf]](https://arxiv.org/pdf/2608.06980)
Abstract: Social collaboration can support people's reflection and is a crucial component of creativity. Creative technologies have been designed to support more collaborative ways of working, including using AI to simulate social partners. As human-AI creative collaborations increase, further investigation is needed into how different social interactions influence creative reflection and at which stage a social intervention is crucial to improve creative outcomes. Considering that non-verbal communication is the bedrock of human cognition and influence, non-verbal social dynamics should be examined in detail in the age of AI-companionship. For example, during social interaction, the social facilitation effect describes how the mere presence or observation of others influences how a person behaves and feels in the context. Whether changes in technology-mediated social environments influence how people reflect on their creative work needs further exploration, as does whether social AI-companionship elicits similar effects as humans. This paper discusses how theoretical mechanisms relating to social facilitation could influence creative practice and reflection, proposing ways of further testing these effects.
arXiv ID: 2608.06980
MobileForge: Annotation-Free Adaptation for Mobile GUI Agents with Hierarchical Feedback-Guided Policy Optimization - Guangyi Liu, Pengxiang Zhao, Gao Wu, Yiwen Yin, Mading Li, Liang Liu, Congxiao Liu, Zhang Qi, Mengyan Wang, Liang Guo, Jiangning Zhang, Yong Liu - [[pdf]](https://arxiv.org/pdf/2606.19930)
Abstract: MLLM-based mobile GUI agents have made substantial progress in UI understanding and action execution, but adapting them to real target apps remains costly because mobile apps are numerous, frequently updated, and hard to cover with human-written tasks, demonstrations, or reward labels. Existing annotation-free GUI learning reduces manual supervision, yet lacks a unified substrate connecting target-app exploration, curriculum mining, rollout execution, and feedback, while policy optimization often relies on isolated rollouts and coarse rewards that are hard to convert into reliable improvement signals. We present MobileForge, an annotation-free adaptation system for mobile GUI agents. MobileForge consists of MobileGym, which grounds task generation and rollout evaluation in real mobile app interaction, and Hierarchical Feedback-Guided Policy Optimization (HiFPO), which turns trajectory outcomes, step-level process feedback, and corrective hints into hint-contextualized step-level GRPO updates. Using only automatically generated annotation-free adaptation data, MobileForge adapts Qwen3-VL-8B to 67.2% Pass@3 on AndroidWorld, close to the closed-data GUI-specialized GUI-Owl-1.5-8B base model at 69.0%. The MobileForge-adapted ForgeOwl-8B further reaches 77.6% Pass@3 on AndroidWorld and 41.0% success on the out-of-domain MobileWorld GUI-only split, establishing the strongest open-data mobile GUI agent in our evaluation. Code, data, and trained models will be released at this https URL.
arXiv ID: 2606.19930
The Perils of Agency: How Developers Perceive, Prioritize, and Address Risks in Agentic AI Products - Hao-Ping Lee, Jessica He, David Piorkowski, Thomas Serban von Davier, Jodi Forlizzi, Sauvik Das - [[pdf]](https://arxiv.org/pdf/2606.15485)
Abstract: Agentic AI systems act autonomously, use tools, adapt to context, and operate in complex real-world environments. However, these same characteristics can create or exacerbate product risks. We studied how industry developers (n=35) perceive, prioritize, and address the risks in their agentic AI products. We found that developers' perceptions of risk were closely tied to the qualities that made the product agentic, such as autonomy, tool use, and usage in a real-world context. Developers prioritized product and business risks before considering downstream societal risks like job displacement and end-user privacy. This prioritization also impacted developers' ability and motivation to mitigate agentic risks. Finally, developers lacked mature controls for containing agentic risks, often relying on constraining the same characteristics that make agents useful: e.g., autonomy and goal complexity. These findings reveal a capability vs. risk control tension in agentic AI development: developers need to address risks that emerge from agentic capabilities, yet they currently have limited support for doing so without constraining agentic functionality.
arXiv ID: 2606.15485
This list is automatically generated daily using arXiv web scraping