Rethinking the Sampling Criteria in Reinforcement Learning for LLM Reasoning: A Competence-Difficulty Alignment Perspective
Reinforcement learning exhibits potential in enhancing the reasoning abilities of large language models, yet it is hard to scale for the low sample efficiency during the rollout phase. Existing methods attempt to improve efficiency by scheduling problems based on problem difficulties. However, these approaches suffer from unstable and biased estimations of problem difficulty and fail to capture the alignment between model competence and problem difficulty in RL training, leading to suboptimal results. To tackle these limitations, this paper introduces $\textbf{C}$ompetence-$\textbf{D}$ifficulty $\textbf{A}$lignment $\textbf{S}$ampling ($\textbf{CDAS}$), which enables accurate and stable estimation of problem difficulties by aggregating historical performance discrepancies of problems. Then the model competence is quantified to adaptively select problems whose difficulty is in alignment with the model's current competence using a fixed-point system. Experimental results across a range of challenging mathematical benchmarks show that CDAS achieves great improvements in both accuracy and efficiency. CDAS attains the highest average accuracy against baselines and exhibits significant speed advantages compared to Dynamic Sampling, a competitive strategy in DAPO, which is 2.33 times slower than CDAS.
Code (0)
등록된 구현이 없습니다.
Tasks
SchedulingMethods 이 논문이 사용한 방법론
Similar Papers 제목 키워드 기반
CARE: Competence-Aware Reward Shaping for Adaptive Reasoning Length in Video-MLLMs
In multimodal video reasoning, reinforcement learning-based methods typically rely on simplistic and inflexible reasoning-length control strategies that fail to adapt to the model's evolving competence. This mismatch may…
Reinforcement LearningMultimodal ReasoningDiagLoop: A Counterfactual Data Flywheel with Stage-Localized Reinforcement for Diagnostic LLMs
Causal diagnostic models must explain how conclusions follow from evidence because diagnoses guide repairs and treatments. Yet serious cases are scarce, records rarely contain reasoning paths, and data transfer poorly ac…
Reinforcement LearningProRL: Prolonged Reinforcement Learning Expands Reasoning Boundaries in Large Language Models
Recent advances in reasoning-centric language models have highlighted reinforcement learning (RL) as a promising method for aligning models with verifiable rewards. However, it remains contentious whether RL truly expand…
Reinforcement Learning (RL)Rethinking Reward Supervision: Rubric-Conditioned Self-Distillation
Post-training of reasoning language models is commonly driven by supervised distillation and reinforcement learning with verifiable rewards. Distillation often relies on chain-of-thought annotations that are expensive to…
Reinforcement LearningCurriculum Learning for Graph Neural Networks: A Multiview Competence-based Approach
A curriculum is a planned sequence of learning materials and an effective one can make learning efficient and effective for both humans and machines. Recent studies developed effective data-driven curriculum learning app…
Link PredictionNode ClassificationScheduling