Teaching LLMs for Step-Level Automatic Math Correction via Reinforcement Learning
This addresses the need for detailed feedback in math education, offering an incremental improvement over existing problem-level correction methods.
The paper tackles the problem of providing step-level feedback in automatic math correction by proposing a reinforcement learning method to enhance large language models, achieving superior performance over eleven baselines on two benchmark datasets.
Automatic math correction aims to check students' solutions to mathematical problems via artificial intelligence technologies. Most existing studies focus on judging the final answer at the problem level, while they ignore detailed feedback on each step in a math problem-solving process, which requires abilities of semantic understanding and reasoning. In this paper, we propose a reinforcement learning (RL)-based method to boost large language model (LLM) for step-level automatic math correction, named StepAMC. Particularly, we convert the step-level automatic math correction within the text classification task into an RL problem to enhance the reasoning capabilities of LLMs. Then, we design a space-constrained policy network to improve the stability of RL. Then, we introduce a fine-grained reward network to convert the binary human feedback into a continuous value. We conduct extensive experiments over two benchmark datasets and the results show that our model outperforms the eleven strong baselines.