中文
相关论文

相关论文: Depth-Breadth Synergy in RLVR: Unlocking LLM Reaso…

200 篇论文

Reinforcement Learning with Verifiable Rewards (RLVR) has proven effective for Large Language Model (LLM) reasoning, yet current methods face key challenges in resource allocation and policy optimization dynamics: (i) uniform rollout…

机器学习 · 计算机科学 2026-04-24 Yangyi Fang , Jiaye Lin , Xiaoliang Fu , Cong Qin , Haolin Shi , Chaowen Hu , Lu Pan , Ke Zeng , Xunliang Cai

Reinforcement learning with verifiable rewards (RLVR) has delivered impressive gains in mathematical and multimodal reasoning and has become a standard post-training paradigm for contemporary language and vision-language models. However,…

机器学习 · 计算机科学 2025-10-28 Hoang Phan , Xianjun Yang , Kevin Yao , Jingyu Zhang , Shengjie Bi , Xiaocheng Tang , Madian Khabsa , Lijuan Liu , Deren Lei

The ongoing debate on whether reinforcement learning with verifiable rewards (RLVR) expands or shrinks the reasoning capabilities of large language models (LLMs) remains unresolved. Some studies contend that RLVR mainly improves sampling…

机器学习 · 计算机科学 2025-10-07 Xinhao Yao , Lu Yu , Xiaolin Hu , Fengwei Teng , Qing Cui , Jun Zhou , Yong Liu

It is widely recognized that reinforcement learning (RL) fine-tuning of large language models often leads to diversity collapse, where outputs lack variety. Prior work has proposed a range of heuristics to counteract this effect, but these…

机器学习 · 计算机科学 2025-12-12 Jingchu Gai , Guanning Zeng , Huaqing Zhang , Aditi Raghunathan

Reinforcement Learning with Verified Reward (RLVR) has emerged as a critical paradigm for advancing the reasoning capabilities of Large Language Models (LLMs). Most existing RLVR methods, such as GRPO and its variants, ensure stable updates…

Recent studies posit that Reinforcement Learning with Verifiable Rewards (RLVR) primarily amplifies behaviors inherent to the pre-training distribution rather than inducing new capabilities, but these insights are predominantly limited to…

人工智能 · 计算机科学 2026-04-15 Minghe Shen , Zhuo Zhi , Chonghan Liu , Shuo Xing , Zhengzhong Tu , Che Liu

Policy-based reinforcement learning currently plays an important role in improving LLMs on mathematical reasoning tasks. However, existing rollout-based reinforcement learning methods (GRPO, DAPO, GSPO, etc.) fail to explicitly consider…

机器学习 · 计算机科学 2025-09-25 Guochao Jiang , Wenfeng Feng , Guofeng Quan , Chuzhan Hao , Yuewei Zhang , Guohua Liu , Hao Wang

Reinforcement learning with verifiable rewards (RLVR) succeeds in reasoning tasks (e.g., math and code) by checking the final verifiable answer (i.e., a verifiable dot signal). However, extending this paradigm to open-ended generation is…

计算与语言 · 计算机科学 2026-01-27 Yuxin Jiang , Yufei Wang , Qiyuan Zhang , Xingshan Zeng , Liangyou Li , Jierun Chen , Chaofan Tao , Haoli Bai , Lifeng Shang

Reinforcement Learning with Verifiable Rewards (RLVR) offers a robust mechanism for enhancing mathematical reasoning in large models. However, we identify a systematic lack of emphasis on more challenging questions in existing methods from…

人工智能 · 计算机科学 2026-01-29 Yanqi Dai , Yuxiang Ji , Xiao Zhang , Yong Wang , Xiangxiang Chu , Zhiwu Lu

When applying reinforcement learning--typically through GRPO--to large vision-language model reasoning struggles to effectively scale reasoning length or generates verbose outputs across all tasks with only marginal gains in accuracy. To…

计算与语言 · 计算机科学 2025-10-24 Wenyi Xiao , Leilei Gan

Recent Large Language Models (LLMs) and Large Vision-Language Models (LVLMs) increasingly use Reinforcement Learning (RL) for post-pretraining, such as RL with Verifiable Rewards (RLVR) for objective tasks and RL from Human Feedback (RLHF)…

计算机视觉与模式识别 · 计算机科学 2025-09-29 Ziyu Liu , Yuhang Zang , Shengyuan Ding , Yuhang Cao , Xiaoyi Dong , Haodong Duan , Dahua Lin , Jiaqi Wang

We propose VL Norm (Variance-reduced Length-dependent Normalization), a simple yet effective loss aggregation method tailored to the characteristic of dynamic generation lengths in Reinforcement Learning with Verifiable Rewards (RLVR).…

机器学习 · 计算机科学 2025-10-14 Zhiyuan He , Xufang Luo , Yike Zhang , Yuqing Yang , Lili Qiu

While Large Language Models (LLMs) have demonstrated strong math reasoning abilities through Reinforcement Learning with *Verifiable Rewards* (RLVR), many advanced mathematical problems are proof-based, with no guaranteed way to determine…

计算与语言 · 计算机科学 2026-02-20 Haotong Yang , Zitong Wang , Shijia Kang , Siqi Yang , Wenkai Yu , Xu Niu , Yike Sun , Yi Hu , Zhouchen Lin , Muhan Zhang

Reinforcement Learning with Verifiable Rewards (RLVR) is commonly based on group sampling to estimate advantages and stabilize policy updates. In practice, computational limits often rule out very large groups, so training proceeds with…

The growing disparity between the exponential scaling of computational resources and the finite growth of high-quality text data now constrains conventional scaling approaches for large language models (LLMs). To address this challenge, we…

Reinforcement Learning with Verifiable Rewards (RLVR) has become the dominant paradigm for advanced reasoning in Large Language Models (LLMs), but rollout samples are expensive to obtain, making sample efficiency a critical bottleneck. A…

机器学习 · 计算机科学 2026-05-20 Yuchun Miao , Sen Zhang , Yuqi Zhang , Yaorui Shi , Qi Gu , Xunliang Cai , Lefei Zhang

Large language models (LLMs) exhibit enhanced capabilities in language understanding and generation. By utilizing their embedded knowledge, LLMs are increasingly used as conversational recommender systems (CRS), achieving improved…

信息检索 · 计算机科学 2026-04-14 Zhenrui Yue , Honglei Zhuang , Zhen Qin , Zhankui He , Huimin Zeng , Julian McAuley , Dong Wang

Reinforcement Learning with Verifiable Rewards (RLVR) has significantly advanced the reasoning capabilities of Large Language Models (LLMs) by optimizing them against factual outcomes. However, this paradigm falters in long-context…

计算与语言 · 计算机科学 2026-03-03 Guanzheng Chen , Michael Qizhe Shieh , Lidong Bing

Large Reasoning Language Models (LRLMs or LRMs) demonstrate remarkable capabilities in complex reasoning tasks, but suffer from significant computational inefficiencies due to overthinking phenomena. Existing efficient reasoning methods…

人工智能 · 计算机科学 2025-10-13 Dongqi Zheng

Reinforcement Learning from Verifier Rewards (RLVR) has emerged as a widely used approach for post-training large language models on reasoning tasks, with group-based methods such as GRPO and its variants gaining broad adoption. These…