Trends · 研究趋势
数据来自 Hugging Face 论文的 AI 提取关键词,按月统计研究方向的增长与热度。
Kwai Keye Team, Biao Yang, Bin Wen +57 authors
Kwai Keye-VL, an 8-billion-parameter multimodal model, excels in short-video understanding and general vision-language tasks through a comprehensive pre-training and post-training process, including a five-mode data mixture and reinforcement learning.
Ganqu Cui, Yuchen Zhang, Jiacheng Chen +14 authors
Entropy dynamics in reinforcement learning with large language models are investigated to prevent policy entropy collapse and improve exploration.
Kimi Team, Angang Du, Bofei Gao +91 authors
A multi-modal LLM trained with reinforcement learning achieves state-of-the-art reasoning performance across various benchmarks by utilizing long context scaling and effective policy optimization methods.
Hongjin Su, Shizhe Diao, Ximing Lu +13 authors
A small orchestrator using ToolOrchestra method coordinates various intelligent tools with reinforcement learning, achieving higher accuracy and efficiency in solving complex tasks like Humanity's Last Exam compared to larger models.
Haoming Wang, Haoyang Zou, Huatong Song +103 authors
UI-TARS-2, a native GUI-centered agent model, addresses challenges in data scalability, multi-turn reinforcement learning, and environment stability, achieving significant improvements over its predecessor and strong baselines across various benchmarks.
Tianzhe Chu, Yuexiang Zhai, Jihan Yang +6 authors
Reinforcement learning demonstrates superior generalization compared to supervised fine-tuning across textual and visual domains, while supervised fine-tuning stabilizes the model output format essential for RL performance.
Xuekai Zhu, Daixuan Cheng, Dinghuai Zhang +20 authors
FlowRL enhances LLM reinforcement learning by matching the full reward distribution through flow balancing, improving diversity and performance over reward-maximizing methods.
Weijie Wang, Xiaoxuan He, Youping Gu +9 authors
World-R1 framework improves video generation by incorporating 3D constraints through reinforcement learning and specialized text datasets while maintaining visual quality and scalability.
Wenhan Liu, Xinyu Ma, Weiwei Sun +4 authors
A reasoning-intensive reranker, ReasonRank, achieves state-of-the-art performance in passage ranking tasks by using synthesized training data and a two-stage post-training approach with reinforcement learning.
Ling Team, Bin Han, Caizhi Tang +25 authors
The Ring-linear model series, including Ring-mini-linear-2.0 and Ring-flash-linear-2.0, uses a hybrid architecture combining linear and softmax attention to reduce inference costs and improve training efficiency.
Yongheng Zhang, Ziang Liu, Jiaxuan Zhu +17 authors
Large Language Models are evolving from conversational systems to integrated AI colleagues with enhanced reasoning capabilities and persistent work environments.
Wenhao Zeng, Yuling Shi, Xiaodong Gu +10 authors
A Dockerless environment-free agentic patch verifier improves code patch evaluation accuracy and enables effective post-training without execution-based verification costs.
Zhengxi Lu, Zhiyuan Yao, Zhuowen Han +8 authors
SDAR enhances reinforcement learning for multi-turn agent training by integrating self-distillation through a sigmoid gate that selectively strengthens positive token-level guidance while mitigating negative teacher rejections.
Yaorui Shi, Yuxin Chen, Zhengxi Lu +6 authors
Skill1 is a unified framework that trains a single policy to simultaneously evolve skill selection, utilization, and distillation capabilities using a shared task-outcome objective, demonstrating superior performance over existing baselines in complex task environments.
Luoxin Chen, Jinming Gu, Liankai Huang +33 authors
Seed-Prover, a lemma-style reasoning model using Lean, achieves high performance in formal theorem proving and automated mathematical reasoning through iterative refinement and specialized geometry support.
LASA Team, Weiwen Xu, Hou Pong Chan +16 authors
A medical-specialized multimodal large language model, Lingshu, is introduced with enhanced data curation and reinforcement learning to address limitations in medical applications.
Yifan Jiang, Yueying Wang, Rui Zhao +4 authors
VideoP2R, a process-aware reinforcement fine-tuning framework, improves video reasoning and understanding by modeling perception and reasoning separately, achieving state-of-the-art results on multiple benchmarks.
Zhuofeng Li, Haoxiang Zhang, Seungju Han +6 authors
AgentFlow, a trainable agentic framework with in-the-flow optimization, enhances reasoning in large language models by coordinating specialized modules and outperforms top baselines across various tasks.
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.
Songchun Zhang, Zeyue Xue, Siming Fu +6 authors
Astrolabe is an efficient online reinforcement learning framework for distilled autoregressive video models that improves generation quality through forward-process RL formulation and streaming training with multi-reward objectives.
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