CUA-Universe is a scalable environment-to-data pipeline that transforms real desktop software into hybrid GUI+CLI environments, enabling agents to coordinate visual inspection with command-line operations. It includes App-Forge for reproducible VMs and CLI surfaces, Task-Weave for generating diverse hybrid tasks, and Path-Steer for efficient rollouts and trajectory harvesting. Training on this data improves agent success and efficiency across multiple benchmarks, demonstrating the value of hybrid interaction.
By Haoting Shi, Wenhao Wang, Weicheng Fang, Yaozhong Liang, Tian Jin, Pengxiang Zhao, Guangyi Liu, Siheng Chen, Yanfeng Wang
The paper introduces KOPA-Bench, a benchmark of 145 real-world tasks that evaluate multi-step tool‑calling over Korean open public APIs. It also presents EDGE, a data‑synthesis method that builds an execution‑grounded dynamic graph to generate executable multi‑step trajectories, and shows that a fine‑tuned 9B model performs nearly as well as an untuned 27B model on KOPA‑Bench and the BFCL benchmark.
By Dain Kim, Eungi Cho, Kyumin Kim, Shinyeong Noh, Kyuseong Lim
The paper introduces llmovoice, a middleware that explicitly models voice context for large language model (LLM) serving in voice AI applications. By incorporating speaking rate, background noise, packet loss, and other paralinguistic factors into a bounded context, llmovoice guides the LLM to generate more aligned responses. Experiments show significant reductions in speaking‑rate errors, false interruptions, and model usage costs, especially in long voice sessions.
By Linyi Jiang, Silvery D. Fu, Yifei Zhu
The paper proposes a new interpretable machine learning approach for discovering unknown nonlinear ordinary differential equations from a single state trajectory. It differs from existing methods by deriving its formulation from Functional Analysis and Operator Theory and by defining a cost function as an integral distance between functions rather than a discrete error sum. An incremental learning algorithm enables online updates, allowing simultaneous identification of both system dynamics and external time‑varying forces, with numerical examples illustrating its benefits.
By Seyyed Shaho Alaviani, Yongzhi Qu, Gregory W. Vogl
arXiv:2609. 04552v1 Announce Type: cross Abstract: Unattended interactive autonomy - machines that step into danger in place of humans and complete tasks with human tools - remains a missing capability in mission-critical operations.
By Amarjot Singh, Tanmay R. Pancholi, Jainam Kothari, Shrirang Mahajan, Ketan Bansal, Zackory Erickson, Giuseppe Loianno, Alexandre M. Bayen, Jeff Schneider, Vince Nakayama
The paper reports on building a research-software catalog using a coding agent, starting from a three‑day hackathon prototype and moving to public deployment. It details the engineering work needed—adversarial review, data‑quality checks, browser validation, and publication safeguards—to ensure reliable operation, noting that silent failures were more problematic than crashes. The authors then examine applying these lessons to a larger, human‑curated portal (MateriApps) that combines curated metadata, external documentation, vector search, and local language‑model generation, finding that explicit validation, monitoring, and repeated review remain essential for AI‑assisted software portals.
By Kazuyoshi Yoshimi, Satoshi Terasaki, Gotai Yamada
The paper investigates how large language models balance helpfulness and safety by refusing harmful queries while responding to benign ones. It decomposes safety-tuning responses into a boilerplate refusal statement and a rationale, finding that the statement causes false refusals by relying on superficial cues. Training on rationales alone reduces false refusals without compromising safety performance, suggesting that fine‑grained safety supervision is essential for better alignment.
By Minji Kim, Hyounghun Kim
Persistent Teacher Anchoring (PTA) is a method that extends on‑policy knowledge distillation by ensuring that a teacher verifies entire turns before any tool calls are executed. PTA builds on chunk‑level verification with an added turn‑level commitment, treating verified chunks as atomic units and introducing persistent lookahead to keep rollout capacity full. Experiments on Search‑R1 and DeepEyes show that PTA improves macro best@4 by 2.5–2.8 points over standard OPKD and boosts throughput by 24%.
By Hyun Bin Park (Sogang University), Kyungho Song (University of Michigan, Ann Arbor), Sangmin Lee (Sogang University), Du-Seong Chang (Sogang University)
MABPD (Multi‑Agent Bias Probing & Detection) is a training‑free pipeline that uses three specialized large language model agents to analyze news articles from complementary perspectives and resolve disagreements via a Structured Argument Debate (SAD) protocol. SAD imposes an asymmetric burden of proof—biased claims lacking grounded textual evidence receive zero weight—along with role‑weighted voting and post‑consensus verification, replacing task‑specific supervised decision boundaries. Ablation studies show that the debate module alone accounts for up to a 10.6‑point F1 gain, and on the BABE benchmark MABPD attains 83.4% macro F1, within 0.7 percentage points of the supervised state‑of‑the‑art, while achieving 75.0% zero‑shot accuracy on the SemEval 2019 HyperPartisan corpus.
By Garvit Joshi (Graphic Era University, Dehradun, India), Stavya Dhyani (Graphic Era University, Dehradun, India), Jasmine (Graphic Era University, Dehradun, India), Arun Chauhan (Graphic Era University, Dehradun, India)
arXiv:2609. 04875v1 Announce Type: cross Abstract: Long-running LLM agents are stateful: beyond the transcript they accrete compressed summaries, plaintext memory, pending tool plans, and, under every serving API, a KV cache.
By Chao Yao, Yangbo Wei, Zhen Huang, Junhong Qian, Chenle Chen, Shaoqiang Lu, Chen Wu, Lei He
The paper addresses the failure of large language models (LLMs) in code generation when routine correctness relies on execution-dependent coupling—situations where the meaning of one routine depends on another’s runtime behavior. It introduces a dynamic context adaptation framework that iteratively validates generated code, extracts diagnostic information from execution traces, and guides generation using a knowledge graph and simulated annealing. Experiments show the method surpasses zero‑shot, Reflexion, and OpenEvolve on most benchmark problems, especially where runtime coupling is critical.
By Gnaneswar Villuri, Hashmath Shaik, Alex Doboli
The paper introduces Debate-Mixture-of-Agents (DMoA), a multi‑agent framework that structures role‑based interactions to mimic iterative diagnostic reasoning in clinical settings. Evaluated on 297 rare disease cases and 1,719 challenging cases, DMoA outperformed a GPT‑4o baseline, improving most likely diagnosis accuracy by 10.21 percentage points and safety rate by 11.36 percentage points. Ablation studies and further analyses revealed that these gains stem from the structured workflow rather than merely adding more models or longer outputs, and that performance benefits are influenced by the chosen structure, base model strength, and token budget.
By Chang Xia, Leilei Ouyang, Huimin Wang, Yong Zhao, Kang Li
RL-VLA$^3$ is a fully asynchronous distributed reinforcement learning framework designed for Vision‑Language‑Action (VLA) model training. It allows fine‑grained asynchronous interaction between simulation, inference, and training via dynamic batching schedulers and flexible environment sharding, addressing the variable, resource‑intensive latencies of physical simulators. Experiments across multiple simulation backends, VLA architectures, and RL algorithms show throughput gains of up to 85.2% over synchronous baselines while preserving sample efficiency, and the system scales from 8 to 256 GPUs.
By Haoran Sun, Yongjian Guo, Zhong Guan, Shuai Di, Xiaodong Bai, Jing Long, Tianyun Zhao, Mingxi Luo, Hongke Zhao, Likang Wu, Xiaotie Deng, Xu Chu, Xi Xiao, Sheng Wen, Yicheng Gong, Junwu Xiong
MemCoRe is a memory system for large language model agents that organizes factual knowledge into a compression hierarchy, progressively reducing redundancy while preserving retrieval structure. The hierarchy compresses detailed records into keywords and then into topic groups, allowing evidence to be located by searching across levels. Experiments show that MemCoRe outperforms current state‑of‑the‑art baselines in retrieving relevant evidence for downstream reasoning.
By Zhenyuan Zhang, Xianzhang Jia, Zhiqin Yang, Zhenbo Song, Wei Xue, Sirui Han, Yike Guo
The paper explores how to trace the functional role of AI in natural language generation, distinguishing between AI acting as an assistive editor or a creative generator. It proposes a methodology that infers the latent role from prompts, embeds it during generation, and recovers the role from the output. Experiments demonstrate that the approach can discriminate roles, remains robust to perturbations, and preserves linguistic quality.
By Ching-Chun Chang, Yuchen Guo, Hanrui Wang, Timo Spinde, Isao Echizen
The paper introduces Agentic Context Cracking, a technique that adaptively and speculatively structures unstructured data during the reasoning process of large language model agents. By creating a sub-agent that extracts useful structure from documents as they are opened, the method reduces the need to repeatedly read large files, cutting token usage by 53% on the FanOutQA benchmark while maintaining accuracy. Over time, more queries are answered using the accumulated structured data, approaching the efficiency of a database lookup.
By Milad Rezaei Hajidehi, Qitong Wang, Stratos Idreos
The paper introduces Post Fusion Stabilizer (PFS), a lightweight module that refines intermediate bird’s‑eye view (BEV) feature maps in existing camera‑LiDAR fusion detectors. PFS stabilizes feature statistics under domain shift, suppresses regions affected by sensor degradation, and adaptively restores weakened cues via residual correction, acting as a near‑identity transformation. On the nuScenes benchmark, PFS achieves state‑of‑the‑art robustness, notably improving camera dropout robustness by +1.2% and low‑light performance by +4.4% mAP while adding only 3.3 M parameters.
By Trung Tien Dong, Dev Thakkar, Arman Sargolzaei, Xiaomin Lin
Octopus Protocol is a hardware onboarding framework that allows an AI coding agent to automatically discover, identify, and integrate hardware devices into an AI system. Using a single bootstrap command, the agent runs a five-stage pipeline to enumerate visible hardware, infer device capabilities, generate typed Model Context Protocol tools, produce the necessary code, and activate a live endpoint. The system maintains a persistent daemon that repairs deployment failures, enabling consistent, platform‑agnostic interfaces across diverse hosts without manual integration code.
By Quilee Simeon, Justin M. Wei, Yile Fan
A new study shows that conformal certificates can become invalid when a large language model (LLM) is influenced by peers who unanimously provide a wrong answer, even though the question itself remains unchanged. This phenomenon, termed a score‑mechanism shift, reveals that a model’s calibration for single‑agent scoring does not hold in multi‑agent settings, leading to a drop in coverage from 90% to 74% under unanimous‑wrong peers. The shift also allows attackers to target low‑confidence items, nearly halving coverage for that subgroup while keeping overall averages deceptively high, and can cause systems to act confidently on incorrect answers.
whyItMatters":"The findings expose a critical vulnerability in conformal prediction for multi‑agent LLM systems, undermining their reliability and safety in real‑world applications."
By Yibo Hu, Hanyu Su
The paper presents an energy‑based reduced‑order model for micromagnetic magnetization dynamics that couples a convolutional autoencoder with a latent neural ODE. The latent dynamics are driven by a learned scalar potential via an antisymmetric operator and a symmetric dissipative operator, ensuring monotonic energy decrease while allowing motion along level sets. The model is trained solely on short trajectory windows without explicit physical labels, and shows that antisymmetric‑dissipative and deep‑quadratic energy formulations yield superior long‑term trajectory predictions compared to purely dissipative models.
By Sebastian Schaffer, Lukas Exl