The paper presents a method for improving the robustness of deep reinforcement learning (DRL) policies used in planetary landing by employing adversarial reinforcement learning. A secondary agent generates physically constrained perturbations that challenge the landing controller, creating an asymmetric zero‑sum game. Training is conducted in three stages—baseline learning, stochastic robustness enhancement, and iterative adversarial optimization—showing that adversarial disturbances uncover vulnerabilities missed by conventional randomization and hint at better generalization to unseen perturbations.
By Miguel Leiva-V\'elez, Nicolas Gaston-Rozado, Andrea Scorsoglio, Hodei Urrutxua, V\'ictor Rodr\'iguez-Fern\'andez
OxyGen introduces a unified KV cache management system for Vision‑Language‑Action (VLA) models that run multiple tasks in parallel, such as manipulation, conversation, and memory construction. By treating the KV cache as a shared resource, it eliminates redundant prefill operations and enables cross‑frame continuous batching, leading to significant speedups—up to 3.7× faster than isolated execution—while maintaining high language throughput and action frequency on both RTX 4090 and Jetson AGX Thor platforms. The approach is validated across several backbones and expert configurations, including deployment on a humanoid robot with on‑board Jetson hardware.
By Xiangyu Li, Huaizhi Tang, Xin Ding, Weijun Wang, Ting Cao, Yunxin Liu
Agentic‑MatriBO is a new Bayesian optimization framework that treats BO components and selection rules as composable capabilities, allowing a language policy to propose optimization recipes that are then verified by a compiler and executed deterministically. The system stores oracle observations in research memory to inform subsequent rounds and was evaluated against eight baselines on four offline candidate pools (COF, PCE10, Buchwald–Hartwig, and Suzuki–Miyaura). Across these tasks, Agentic‑MatriBO achieved the best mean endpoint quality, reached the COF optimum in all five seeds, and showed higher cumulative quality and nAUC‑BO on COF, with ablations indicating the importance of the capability compiler and evidence gate.
By Yu Zeng, Yuchuan Chen, Ziyue Zhang, Zijian Wei, Yanyi Su, Guojiang Zhao, Hongshuai Wang, Zhifeng Gao
arXiv:2607.29237v2 Announce Type: replace
Abstract: LiDAR scene flow is the task of estimating point-wise motion from sequential point clouds. It is essential for autonomous driving, as it provides c...
By Minh-Quan Dao, Yancong Lin, Julie Stephany Berrio Perez, Holger Caesar
arXiv:2610.14539v1 Announce Type: new
Abstract: Real-world business requests often leave key requirements unstated because frontline practitioners struggle to articulate rules they know from experien...
By Huaiming Li, Can Huang, Chufan Wu, Danyu Liu, Xiaomin Yuan, Qinze Li, Yui Lo, Wenhui Bai, Qianlong Wang, Zengle Ge, Xiyu Yan, Yuheng Cheng, Jinhui Ren, Guokai Chen, Jiaqun Liu, Haoran Li, Qianhui Liu, Annan Li, Jianmin Wu, Dawei Yin, Dou Shen
arXiv:2610.14636v1 Announce Type: new
Abstract: Terminal rewards make multi-turn agents trainable without step-level annotations, but group-relative advantages provide little guidance about which int...
By Jie Yang, Mianqiu Huang, Yiwei Zhao, Bo Wang, Yang Liu, Yuxin Wang
arXiv:2610.14700v1 Announce Type: new
Abstract: Most text-to-SQL benchmarks provide the target database. In a large data catalog, however, choosing the sources is part of the problem: an agent must f...
By Jinlong Luo, Haoran Wu, Bo Xu
arXiv:2610.14736v1 Announce Type: new
Abstract: Optimization by PROmpting (OPRO) directly leverages Large Language Models (LLMs) to generate candidate solutions through natural-language prompting. Ho...
By Zhongnian Li, Shi Xu, Peng Ying, Weikang Gong, Zhixiao Wang, Xinzheng Xu
arXiv:2610.14877v1 Announce Type: new
Abstract: Experience-based learning enables large language model (LLM) agents to improve over time by reusing knowledge from past executions. However, existing m...
By Bingjie Zhang, Yibo Yang, Qiantong Liu, Benshuo Fu, Yi Chang, Dandan Guo
arXiv:2610.14843v1 Announce Type: new
Abstract: Claims of recursive self-improvement are easy to make and hard to falsify: did improvement come from re-using designer-provided mechanisms, or from mec...
By Wenbin Mao, Pei Zou
arXiv:2610.14918v1 Announce Type: new
Abstract: Evaluating an agent's ability to strictly satisfy diverse and procedurally complex requirements is essential to determining its practical reliability....
By Zhengxi Li, Junyi Zhou, Qiyuan Zhang, Chuhan Wu, Fuyuan Lyu, Le Tian, Xiao Zhou, Chen Ma, Yufei Wang
arXiv:2610.14915v1 Announce Type: new
Abstract: Developing terminal-based agents for long-horizon execution demands synthetic tasks that tightly couple complex instructions with authentic, interactiv...
By Junyi Zhou, Yufei Wang, Qiyuan Zhang, Chuhan Wu, Zhengxi Li, Wentao Zhao, Jie Zhou, Le Tian, Xiao Zhou, Chen Ma
arXiv:2610.14913v1 Announce Type: new
Abstract: Reinforcement learning (RL) is essential for adaptive decision-making, e.g., robust Meta-RL and domain randomization (DR), and efficient post-training...
By Jiarong Wen, Qi Tao, Kaiyu Zhang, Yun Qu, Lizhou Cai, Heming Zou, Yuhang Jiang, Qi Wang
arXiv:2610.14966v1 Announce Type: new
Abstract: Long-horizon agents must complete workflows in which later activities depend on earlier outputs and changes to shared state. A key challenge for evalua...
By Yuxuan Liu, Zhaochen Su, Hangyu Guo, Shijue Huang, Zhixiong Zhang, Kawai Chung, Zhongwei Xie, Tianshi Zheng, Wei Fan, Qing Zong, Haochen Shi, Baixuan Xu, Haoran Li, Yangqiu Song
arXiv:2610.14970v1 Announce Type: new
Abstract: With the rapid evolution of large language models (LLMs), there is an increasing demand for effective personalization techniques that can adapt model b...
By Kounianhua Du, Jianxing Liu, Kangning Zhang, Wenxiang Jiao, Yuan Lu, Jiarui Jin, Jianghao Lin, Weiwen Liu, Yong Yu, Weinan Zhang
arXiv:2610.14958v1 Announce Type: new
Abstract: Activation steering has recently been used as a memory for skills: past experiences of an agent are stored as steering vectors and retrieved later. Fac...
By Vedant Jumle
arXiv:2610.15084v1 Announce Type: new
Abstract: Visual agents that interleave reasoning with tool invocation have posed a promising direction for multimodal tasks. However, learning effective tool in...
By Wenxi Yang, Yuzhong Zhao, Baochang Zhang, Yunfan Liu, Qixiang Ye
arXiv:2610.15380v1 Announce Type: new
Abstract: Bridging the structural gap between biological brains and artificial neural networks presents a pathway to altered computational dynamics. This paper i...
By Huy Mai, Robert Clarke
arXiv:2610.15481v1 Announce Type: new
Abstract: Independently authored skills can impose incompatible requirements, competing preferences, and redundant instructions when jointly loaded by an LLM age...
By Enhao Huang, Xinye Zhao, Yiming Li, Bingrun Yang, Nanjun Zhou, Chunshu Xia, Hongwei Yao, Zhan Qin
arXiv:2610.15529v2 Announce Type: new
Abstract: Agent-to-agent protocols allow LLM agents to discover peers, exchange results, and coordinate tool use. A locally successful protocol attack, however,...
By Xiuming Liu, Le Gai, Zhiyuan Chang, Peiyang Li, Bin Zhu, Shuo Wang, Chao Feng