Teach-to-Crash is a closed‑loop testing framework that uses a dual‑LLM architecture to generate collision‑inducing scenarios for autonomous driving systems. A high‑reasoning Teacher LLM controls the search when collision metrics stagnate, while a low‑reasoning Student LLM produces simulator‑executable scenarios in JSON. In a CARLA case study, Teach‑to‑Crash achieved the highest collision hit rate (90.79 %), the shortest mean time‑to‑collision (18.31 s), and superior diversity and avoidability metrics compared to other methods.
By Zaid Ghazal, Khouloud Gaaloul, Bruce Maxim
The paper introduces TSGPD-IR, a network that disentangles weather-induced artifacts from true thermal signals in infrared images. It uses weather semantics and regional degradation severity to generate adaptive prompts, estimate severity without manual labels, and select appropriate expert modules for restoration. This approach aims to reduce artifacts and preserve weak thermal details across varying weather conditions.
By Xinyao Wang, Lijun He, Zhihan Ren, Fan Li
Quantum Reinforcement Learning for Cost and Delay Tradeoffs in Quantum Cloud Orchestration proposes QRLQ, a scheduling framework that integrates parameterised quantum circuits with a dueling double deep Q‑network to balance execution cost and delay in quantum‑as‑a‑service environments. Simulation results show QRLQ outperforms heuristic baselines, achieving 5‑11% lower mean cost and up to 82% lower mean delay while maintaining fidelity within 2% of a fidelity‑greedy policy. Compared to a classical deep reinforcement learning baseline, QRLQ delivers comparable performance with 72% fewer trainable parameters.
By An N. H. Phan, Dang Van Huynh, Muhammad Usman, Hoa T. Nguyen
The paper investigates how closed‑loop autonomous discovery systems can develop false‑science induction when physical objects and measurements are incorrectly paired. It demonstrates that such misbinding causes neural surrogates to learn spurious associations, diverting experimental effort toward low‑performing regions in both green fluorescent protein fitness and materials band‑gap prediction loops. The study shows that the coherence of these errors—not just their frequency—drives budget misallocation and proposes monitoring strategies to detect and quarantine corrupted hypothesis axes.
By Hanbing Liang, Fujun Liu
The paper introduces COMPASS, a decentralized architecture that uses spatial transformers to generate local feedback tokens for large collectives of AI agents in robotics. By aggregating multi‑hop messages, COMPASS enables scalable control of up to 1024 robots, achieving cohesive flocking formations and accurate execution of natural language commands. Experiments show that structured diversity in input commands improves performance and that learned feedback tokens outperform hand‑crafted raw state feedback.
By Frederic Vatnsdal, Roshan Gopal, Romina Garcia Camargo, Vijay Kumar, Alejandro Ribeiro
arXiv:2609. 28366v1 Announce Type: cross Abstract: Vision-language models (VLMs) offer a promising approach to long-tail autonomous driving, but existing driving datasets provide limited supervision for connecting decision-critical visual evidence with reasoning and planning.
By Zhipeng Bao, Wenjie Zhao, Tianle Zhu, Haohua Que, Chence Yang, Geng Yuan, Qianwen Li
MobileGym is a browser-hosted, lightweight simulation platform designed for mobile GUI agent research. It offers verifiable outcome signals via deterministic, JSON-based state judging and supports scalable online reinforcement learning with hundreds of parallel instances on a single server. The platform includes a declarative task-definition framework, a structured AnswerSheet protocol, and a benchmark of 416 parameterized tasks across 28 apps, demonstrating strong sim-to-real transfer in a case study.
By Dingbang Wu, Rui Hao, Haiyang Wang, Shuzhe Wu, Han Xiao, Zhenghong Li, Bojiang Zhou, Zheng Ju, Zichen Liu, Lue Fan, Zhaoxiang Zhang
FDE-Bench is a benchmark that tests large language model agents on 136 deployment‑configuration tasks involving Docker, Compose, and Kubernetes, in both greenfield and diagnose‑and‑repair scenarios. Agents submit declarative artifacts that are rebuilt and redeployed in a clean environment, and four binary check layers evaluate build, readiness, behavior, and specification conformance without an LLM judge. The benchmark includes a release gate, detailed check annotations, adversarial strategies, and reports that state‑of‑the‑art models resolve 52.9–75.0 % of tasks, while zero‑intelligence baselines solve none.
By Weihang Ding, Junfei Zhan, Yueting Li, Qirong Guo
The paper introduces a capability manifold, a multidimensional framework that maps downstream capabilities—such as reasoning, retrieval, planning, and adaptation—to pre‑training, post‑training, and test‑time resources via bounded scaling functions. It provides analytical Jacobians to quantify how sensitive each capability is to changes in resources and their interactions. By embedding existing Kaplan‑ and Chinchilla‑type scaling laws and test‑time compute into this manifold, the authors demonstrate that these scaling relationships can be unified as trajectories on a common capability manifold.
By Syed Ali Raza Zaidi, Maryam Hafeez
ChipMEM introduces a verification‑grounded memory layer for electronic design automation agents that combines cross‑task procedural memory with within‑trajectory statistical guidance. The procedural component stores a skill only after it passes synthesis, simulation, or formal checks, while a Bayesian component ranks recovery strategies based on tool‑call outcomes. Experiments on RTLRewriter‑Bench and CVDP tasks show that ChipMEM improves equivalence‑passing outputs and area metrics compared to agents without memory.
By Abdulrahman AlRabah, Joshua Mabry, Dilek Hakkani-T\"ur, Abdussalam Alawini, Hamid Shojaei, Kartik Hegde, Sandesh Adhikary
The paper introduces a compliance-first AI architecture for regulated finance, treating regulation as an orientation layer rather than a rigid rule set. It uses a regulatory intent matrix and a governed policy compiler to translate regulatory requirements into concrete prohibitions, obligations, and runtime budgets, while maintaining proportional committee activation and supervisory oversight. Evidence and decisions are recorded on a permissioned DAG with deterministic timestamps, enabling replay, provenance checks, and clear attribution of failures, and clause-level legal indexing ensures portability across the DACH region and the EU.
By Walter Kurz, Reinhard Magg
The paper investigates whether complex communication topologies are necessary for effective multi‑agent debate (MAD) among large language models. It demonstrates that a simple random-without-replacement routing policy—where each agent debates with two newly sampled peers each round—consistently improves the accuracy‑cost trade‑off in sparse MAD setups. Additionally, the study shows that lightweight deliberation stopping can further reduce inference costs without sacrificing accuracy.
By Boxuan Wang, Zhuoyun Li, Xiaowei Huang, Yi Dong
The paper introduces Direct Relational Set‑Risk Pruning (DRSR), a method for compressing the history of long‑horizon language‑model agents by selecting deletion sets based on risk constraints rather than independent unit scores. DRSR builds counterfactual supervision offline, then uses a lightweight scorer to predict set‑level harm during deployment, removing the largest safe set while respecting recency, protocol, and budget limits. Experiments on WorkBuddyBench Full260 and Eval40 show that DRSR improves mean reward from 0.699 to 0.802 and reduces token usage by over 20%, with further analyses highlighting the importance of decision‑conditioned relations, retained context, pair interactions, and abstention.
By Mingxuan Wang, Bo Wang, Fei Luo, Guorun Yao, Chao Ning, Yinglong Guo, Hongyue Chen, Yanbiao Ma, Jungong Han
TimeEvo is a new method for time‑series agents that autonomously evolves its tool library based on failures observed during runtime. By clustering diagnosed failures into capability gaps, planning measurements, synthesizing evidence‑only tools, and admitting candidates through a paired gate, the system starts from an empty library and improves accuracy across ten QA tasks and three backbones. Experiments show that even a library built on a cheap model benefits stronger models when installed.
By Jie Yang, Yan Zheng, Jiarui Sun, Xiran Fan, Junpeng Wang, Liang Wang, Zelin Xu, Qinghua Liu, Zhengyu Fang, Yiwei Cai, Philip S. Yu
The paper investigates how long‑horizon language model agents encode memory‑management signals before taking actions. By examining hidden states just prior to each action, the authors find that the model already signals the need for compression and recall, independent of context length or interaction progress, and that these signals vary across model depth. They propose the Preaction Memory with Evidence Retrieval (PaMER) framework, which uses state‑guided compression and selective evidence retrieval to reduce context consumption while preserving task performance.
By Mingxuan Wang, Guorun Yao, Fei Luo, Yinglong Guo, Chao Ning, Bo Wang, Hongyue Chen, Yanbiao Ma, Jungong Han
The paper introduces the Large Knowledge Model (LKM), a scientific knowledge infrastructure that converts research literature into shared, computationally accessible reasoning graphs. LKM aligns questions, claims, and reasoning chains across papers, creating a Scientific Reasoning Landscape with Question, Workflow, and Evidence views. The system enhances scientific search, evidence‑grounded QA, and research planning, achieving notable accuracy gains on ChemBench, PubMedQA, and SciBench.
By Yuan Huang, Sihan Hu, Hongyu Gu, Chao Ma, Jiaxing Zhang, Zhiyong Zou, Caiyu Fan, Yan Xiao, Mingjun Xu, Chenyu Xie, Mingzhen Ju, Zhehao Ma, Qi Zhang, Baozong Wang, Yu Li, Zhiyuan Yao, Ruoxue Liao, Xinyu Li, Linfeng Zhang, Kun Chen, Weinan E
StateComp introduces a method for long‑horizon agents to decide when to compress historical interactions based on the current agent state, rather than relying on fixed windows or periodic schedules. The framework uses a two‑stage annotation process to create KEEP and READY labels, trains an imbalance‑aware router on frozen language model representations, and groups adjacent READY interactions into compact summaries. Experiments on WorkBuddyBench show that StateComp cuts agent and summarization tokens by 52.27% and speeds up representation extraction 12.67‑fold while preserving task performance.
By Mingxuan Wang, Hongyue Chen, Yinglong Guo, Fei Luo, Chao Ning, Bo Wang, Guorun Yao, Yanbiao Ma, Jungong Han
The paper introduces GUI‑SD‑v2, an on‑policy self‑distillation framework that extends previous methods from GUI grounding to multi‑turn GUI interaction. It employs a two‑stage training process that first improves privilege following by jointly optimizing rollouts with and without privileged guidance, then selectively distills step‑specific reasoning and memory guidance via a privilege‑conditioned self‑teacher. Experiments on AndroidWorld and MobileWorld benchmarks demonstrate that GUI‑SD‑v2 outperforms existing OPSD baselines and state‑of‑the‑art methods in Pass@1 and Pass@3 success rates.
By Yan Zhang, Daiqing Wu, Huawen Shen, Liang Li, Gang Cao, Zhi Gong, Wei Dai, Xiaode Zhang, Can Ma, Yu Zhou
The paper introduces VHD-Play, a pipeline that first samples and solves a mathematical model before generating agentic reinforcement learning environments, ensuring that dynamics and evaluation are aligned from the outset. This approach yields 3,300 diverse environments at a low cost and significantly improves the performance of a large language‑model agent (Qwen3.6‑35B‑A3B) across multiple diagnostic families and external benchmarks. The study demonstrates that stateful interaction is a key factor in learning gains and that scaling the training substrate can further enhance performance.
By Xinjie Shen, Wei Fan, Xudong Guo, Jianhong Tu, Yang Su, Chuqiao Kuang, Yinger Zhang, Dayiheng Liu
The paper introduces Geometry Guided Evidence Preserving Memory (GEM), a training‑free compressor that prioritizes task evidence over geometric redundancy when compressing long‑horizon agent histories. Experiments show that GEM reduces token usage by 21.4% while keeping task reward comparable, whereas purely geometric methods fail to preserve critical action‑related information. The study highlights that efficient compression must focus on preserving task evidence rather than just maintaining global geometry.
By Mingxuan Wang, Fei Luo, Bo Wang, Guorun Yao, Yinglong Guo, Chao Ning, Hongyue Chen, Yanbiao Ma, Jungong Han