Learning at the Right Pace: Adaptive Data Scheduling Improves LLM Reinforcement Learning
For researchers and practitioners in LLM post-training, this work addresses the limitation of uniform data sampling by introducing a general data scheduling strategy that consistently improves RL methods.
The paper proposes Adaptive Data Scheduling (ADS), a dual-level framework that replaces uniform data sampling with adaptive distribution over semantic clusters and policy-boundary sample selection for LLM reinforcement learning post-training. ADS improves average accuracy by 5.2% over GRPO across three LLMs and seven reasoning benchmarks.
Large Language Models (LLMs) achieve remarkable reasoning capabilities through reinforcement learning (RL) post-training. However, existing RL post-training commonly relies on uniform data sampling, which ignores the semantic structure of the training data and the changing capability of the training policy. To address these limitations, we propose Adaptive Data Scheduling (ADS), a dual-level data scheduling framework for pacing RL post-training that replaces uniform sampling with an adaptive distribution over semantic clusters and policy-boundary sample selection. At the cluster level, ADS organizes samples according to semantic patterns and maintains an adaptive inter-cluster distribution to solidify current training progress. At the sample level, ADS performs intra-cluster scheduling to continuously sample policy-boundary samples, which provides informative relative advantages. Experimental results across three LLMs and seven reasoning benchmarks demonstrate that ADS improves average accuracy by 5.2% over Group Relative Policy Optimization (GRPO). Notably, ADS consistently improves RL methods with different objective designs, highlighting its potential as a general data scheduling strategy for LLM RL post-training. The source code is available at: https://github.com/Richard-zrx/ADS.