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。
Ziheng Qin, Yaxin Lu, Zhangyang Atlas Wang +1 authors
StateM is a runtime system that improves long-horizon agent execution through durable states, recoverable runbooks, and enforceable procedural controls without altering model weights.
50 篇论文 · 按点赞排序
Weiliang Chen, Haowen Sun, Jun Gao +40 authors
HarnessEval-W uses hierarchical sub-agents to decompose world-model evaluations into verifiable reasoning chains that justify scores with transparent evidence.
Shuo Liang, Yixing Ma, Pengfei Zhou +32 authors
A new benchmark for AI-generated video detection reveals that current detectors fail to generalize across realistic crisis-related videos and become less reliable as content spreads socially.
Chengsong Huang, Zifeng Wang, Rujun Han +14 authors
EnvHarness and EnvRigger dynamically reshape static environments via programmable plugins to target agent weaknesses and improve reinforcement learning co-evolution.
Zhiyuan Jiang, Fangrui Huang, Hanwen Xing +6 authors
Skills enhance LLM agents primarily by stabilizing execution through procedural anchoring rather than injecting missing knowledge, though retrieval bottlenecks and brittle assumptions limit their effectiveness.
Yijiang Li, Yijun Liang, Yunjie Tian +6 authors
Self-supervised visual on-policy distillation improves small vision-language models by distilling from original images into strongly augmented student views without privileged annotations or larger teachers.
Keyu Tu, Zhuowei Chen, Mengqi Huang +4 authors
Semantic task completion video generation evaluates whether generated videos achieve intended outcomes with semantic grounding, supported by a curated dataset and vision-language model-based benchmark.
Yixuan Wang, Yifei Chen, Haichao Zhang +6 authors
SA-MRPO independently standardizes multi-objective rewards and adaptively discounts saturated objectives to redirect optimization toward under-optimized goals.
Xin Ding, Liang Mi, Mingzhe Huang +12 authors
Zetta is a closed-loop embodied harness that evolves runtime critics and recovery skills online to govern physical execution at action frequency, achieving high success on robot benchmarks with faster inference and scaling self-exploration.
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.
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.
Shuo Yang, Xiaoze Fan, Melissa Pan +8 authors
FreeToken is an edge-native Mixture-of-Experts serving system that dynamically maps computation and model state onto heterogeneous local hardware to run large open-weight models on personal machines.
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.
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.
Bobo Li, Hao Fei, Tianjie Ju +2 authors
OmniScientist is an end-to-end omni-modal AI scientist that performs multidisciplinary research directly from heterogeneous raw evidence using autonomous agents and lifecycle-wide perception, improving evidence-grounded discovery across diverse scientific modalities.
Yudong Jin, Tao Xie, Qihang Zhang +6 authors
4DAnyone reconstructs 4D humans from monocular video by generating multiview-consistent videos and lifting them into 4D Gaussian Splatting, using reference and target context designs to overcome scaling bottlenecks.
Brian Wang, Bin Feng, Xiaoman Pan +26 authors
Apodex Discovery introduces a framework for verifiable, extended AI investigations using a heavy-duty solver and structured evaluation across real-world scientific tasks.
Zhipeng Xu, Jiahao Lu, Yining Zheng +2 authors
SWE-bench Science benchmarks coding agents on scientific software repair, revealing failure mechanisms and mixed effects of scientific guidance.
Junwei Zhou, Zhen Sun, Binyu Li +39 authors
Artificial superintelligence (ASI) requires AI to move beyond mastering existing knowledge toward exploring the unknown, creating new knowledge, and turning new ideas into verifiable results. However, the capabilities of today's AI systems are still largely built on learning, compressing, and applying existing human knowledge. Accordingly, existing benchmarks primarily test whether AI can produce correct answers based on learned knowledge, or whether it can complete tasks under extensive human guidance. We therefore introduce ASI-Bench, the first benchmark to jointly evaluate AI systems' capabilities of innovative exploration and autonomous scientific execution across general research domains, and the first to progressively withdraw human methodological guidance within the same research project to test how far AI can proceed on its own. Built by over 40 experts with the cost of 31,000+ human hours, ASI-Bench contains 60 project-level research tasks across 11 scientific domains and progressively reduces methodological guidance to test whether AI can independently select methods, conduct research, and produce verifiable results. All tasks undergo expert review, AI-assisted auditing, sandbox execution, and scorer validation. Across 18 state-of-the-art agent--model configurations, the average score drops from 50.91 with full methodological guidance to 29.10 with only the method specified and 26.62 when agents must determine the method themselves. This sharp decline shows that current systems remain heavily dependent on human guidance and are still far from autonomously conducting end-to-end, project-level scientific research. ASI-Bench is open to the world. We invite researchers and builders everywhere to contribute new tasks, challenge the limits of today's AI, and help accelerate humanity's collective path toward artificial superintelligence at https://asibench.apexin.ai/submit.
Zhongwei Yu, Yan Song, Xue Yan +9 authors
A recurrent Large Discovery Model couples generative proposal with a Bayesian non-parametric reward surrogate to guide uncertainty-aware search across molecules, proteins, and programs.
Yansong Ning, Jingwen Ye, Zhongkai Wu +5 authors
A unified framework benchmarks and trains multimodal agents that infer intent, plan 3D scenes, invoke tools, and reflect on feedback, revealing that reinforcement learning improves open-source models beyond closed-source frontiers.
Yiwei Li, Wanli Yang, Hexiang Tan +10 authors
Frontier autonomous agents excel at engineering optimization but show unstable performance, limited novelty, and variable experience reuse across long-horizon tasks.
Bo Liu, Simon Yu, Yiding Jiang +15 authors
SPADE is a self-play reinforcement learning framework where a language model designs adaptive executable training environments and learns to solve them, improving reasoning and tool-use performance through regret-based environment targeting.
Hongyan Feng, Sunlai Chen, Xuanyu Liu +9 authors
TAMP-Nav improves embodied navigation by aligning vision-language models with 2D visual prompting, selective reasoning with compressed memory, and dense policy optimization.
Pengyu Wang, Chenkun Tan, Shaojun Zhou +29 authors
MOSS-VL is an open vision-language model family enabling real-time interaction by attending to vision via gated cross-attention during generation, using a synthesized interaction corpus and staged curriculum to achieve strong streaming performance with reduced time-to-first-token latency.
Zihan Ding, Longxu Dou, Qi Gao +26 authors
UI-Mate is a foundation GUI agent that uses environment-grounded training and in-context demonstration learning to improve reliability on long-horizon office tasks, achieving state-of-the-art results on computer-use benchmarks.
Chuyue Li, Jinpeng Yu, Haozhe Wang +7 authors
AVA-Encoder learns structured video representations via agentic auto-encoding using knowledge graphs and textual-gradient optimization to improve cinematic video generation and reasoning.
Hengyuan Xu, Qixun Wang, Yiji Cheng +5 authors
WithEveryone enables reliable identity-preserving group image generation for up to ten people by grounding identities to layout plans and using region-based identity losses.
Huatong Song, Fei Bai, Ming Yang +17 authors
A unified black-box reinforcement learning framework enables stable, scalable optimization of general agents through complex harnesses via sandbox execution, trajectory reconstruction, and mix-harness training.
Dengyang Jiang, Ruoyi Du, Zhennan Chen +10 authors
Researchers propose a latent-to-pixel training strategy that accelerates convergence and improves inference speed for large-scale pixel-space diffusion models.
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