MultiLingPoT: Enhancing Mathematical Reasoning with Multilingual Program Fine-tuning

TL;DR

This paper introduces MultiLingPoT, a multilingual program reasoning approach that fine-tunes language models on multilingual data, enabling better mathematical reasoning by selecting the most suitable programming language for each problem.

Contribution

It proposes a novel multilingual program-of-thought method with hybrid language selection, improving mathematical reasoning performance over single-language approaches.

Findings

MultiLingPoT improves reasoning accuracy by about 2.5%.

Proper language mixing boosts performance by 6%.

Hybrid language selection effectively chooses suitable programming languages.

Abstract

Program-of-Thought (PoT), which aims to use programming language instead of natural language as an intermediate step in reasoning, is an important way for LLMs to solve mathematical problems. Since different programming languages excel in different areas, it is natural to use the most suitable language for solving specific problems. However, current PoT research only focuses on single language PoT, ignoring the differences between different programming languages. Therefore, this paper proposes an multilingual program reasoning method, MultiLingPoT. This method allows the model to answer questions using multiple programming languages by fine-tuning on multilingual data. Additionally, prior and posterior hybrid methods are used to help the model select the most suitable language for each problem. Our experimental results show that the training of MultiLingPoT improves each program's mathematical reasoning by about 2.5\%. Moreover, with proper mixing, the performance of MultiLingPoT can be further improved, achieving a 6\% increase compared to the single-language PoT with the data augmentation.Resources of this paper can be found at https://github.com/Nianqi-Li/MultiLingPoT.