Explore · 每周精选
发现最受关注的研究论文,追踪研究趋势,订阅感兴趣的期刊与关键词。
50 篇论文 · 按点赞排序
Yuling Shi, Jinghan Xu, Kelin Fu +12 authors
SWE-Bench ProMax is a rigorously curated multilingual benchmark of large-scale code refactoring tasks that reveals substantial unsolved challenges for current AI coding agents.
Kou Shi, Zun Wang, Qisheng Su +10 authors
FACET constructs executable terminal tasks by preserving source intent and grounding instructions, solutions, and verifiers in a shared repaired environment to enable scalable agent training.
Alexander Panfilov, David Schmotz, Ilia Shumailov +5 authors
Encrypted reasoning traces shared across sessions and models can be intercepted and injected into weaker models to extract proprietary reasoning, private data, hidden hazards, and hidden prompts.
Yanlun Tu, Huacan Wang, Ziyue Zhou +10 authors
SemaPLC is a verification-gated agent harness that validates generated PLC logic through external compilation and live runtime execution, achieving higher verified pass rates than baseline methods.
Cheng Qian, Wenting Zhao, Liangwei Yang +6 authors
Stronger models can build inference-time harnesses that substantially improve weaker models' task performance without parameter updates by offloading reasoning into structured code and routing.
Yifan Zhang, Yutong Dai, Juntao Tan +9 authors
DarwinX evolves agent harnesses via population selection with frozen models, improving verified performance across benchmarks without benchmark-specific patches.
Tianjie Ju, Zheng Wu, Yueqing Sun +15 authors
UrbanGround evaluates whether multimodal language model agents can sustain reliable navigation and spatial reasoning in a realistic 3D city replica, revealing that local perceptual skills fail to compose into extended goal-directed behavior.
Qiming Shi, Yulong Tao, Linbo Jin +10 authors
MerchantBench evaluates long-term coherence of LLM agents in a year-long e-commerce simulation requiring persistent decision-making across sourcing, pricing, and cash-flow management with mixed feedback delays.
Yunheng Li, Guohong Mu, Hao Li +4 authors
OraRL improves reinforcement learning post-training for video multimodal language models by integrating oracle rollouts with decoupled advantage estimation and sign-balanced pruning, achieving higher sample efficiency and scalability without chain-of-thought generation.
Hao Fei, Yiran Zhao
Mental World Modeling integrates hidden mental states into predictive world models to improve action forecasting by jointly simulating physical scenes and agent beliefs.
Zhi Zheng, Rongsheng Chen, Yunpeng Ba +3 authors
Agentic ESOpt uses evolution strategies for scalable full-parameter fine-tuning of long-horizon LLM agents via trajectory-level reward-weighted updates and parameter-context co-evolution.
Zongchuang Zhao, Xin Zhou, Tianyang Xu +5 authors
SimWAM trains a lightweight action planner using video generation as a training signal, enabling efficient trajectory prediction without future generation at inference and supporting reinforcement learning optimization.
Qinsi Wang, Jing Shi, Huazheng Wang +8 authors
RLSVR extends reinforcement learning with verifiable rewards to open-ended tasks by transforming them into self-verifiable proxy environments, demonstrated via a multi-agent spy game that yields automatic rewards for text generation and reasoning.
Yicheng Xiao, Wenxun Dai, Xinran Qin +22 authors
JoyAI-Video-Edit is a 16B-parameter autoregressive diffusion framework that enables real-time, open-ended video editing with high source fidelity and long-term temporal consistency on a single GPU.
Zi-Han Wang, Zhengxi Lu, Zhiyuan Yao +10 authors
AgentOPSD uses recursive Bayesian belief updates over turn-level log-probability gaps to assign credit in long-horizon agentic reinforcement learning without extra critics or rollouts.
Yunhao Yang, Yuexin Bian, Yunjie Tian +6 authors
Co-RL enables unsupervised reasoning via cooperative multi-agent reinforcement learning with peer-derived rewards, improving performance across text and vision tasks without ground-truth labels.
Senqiao Yang, Chengyao Wang, Yuxin Chen +13 authors
VLAct improves vision-language-action model performance by pre-training on diverse robot data with preserved vision-language priors and shared action semantics, achieving strong results across simulations and unseen embodiments with limited compute.
Junliang Ye, Kenkun Liu, Guocun Wang +12 authors
Hunyuan3D-Buffalo 1.0 unifies 3D understanding, generation, editing, and part generation via a multimodal architecture trained on a large-scale 3D corpus, achieving leading benchmark results.
Hangrui Xu, Jiarui Wang, Yang Yang +5 authors
DART-SD improves multi-turn tool-calling agents by modeling execution as a diamond-topology graph, identifying critical failure points, and applying localized self-distillation to preserve valid reasoning while correcting errors.
Anton Razzhigaev, Andrei Gritsaev, Andrei Kaznacheev +3 authors
We present Ouroboros, a self-developing agent harness whose tools, prompts, context assembly, and core implementation improve through reviewed commits that become the runtime for later work. Core evolution proceeds in two modes. In recursive free evolution, improvement is itself a task, and completing one evolution cycle can schedule the next. In experience-driven core evolution, ordinary work and social interaction expose bugs, rough edges, and inefficient context construction that lead to reviewed structural changes. On Terminal-Bench 2.1, an Opus 5 run scores 86.74%, the best result reported on the benchmark. On OSWorld-Verified, an Opus 5 run reaches 90.69%, exceeding the best previously reported score. A five-rollout CL-Bench campaign achieves a normalized reward of 0.2301, setting a new state of the art. Hope is the longest-running publicly documented Ouroboros deployment. It is a 161-day living agent experiment in free evolution under governed human communication across seven surfaces. Human interaction surfaces faults and generates proposals, but the agent decides which changes to pursue. Because a self-developing agent may rewrite its own code and select new model APIs, operational safety becomes a primary design problem: guardrails must remain authoritative under evolutionary and public social pressure. Benchmark campaigns use frozen system snapshots, while Hope continues live evolution on a separate lineage.
北京市昌平区探索星信息技术及软件开发工作室
京ICP备2026059466号