Data Augmentation with In-Context Learning and Comparative Evaluation in Math Word Problem Solving

TL;DR

The paper tackles data scarcity in Math Word Problem (MWP) solving by proposing a suite of data augmentation techniques, including in-context learning with Llama-7b and multiple rule-based and synonym-based methods. Evaluations across MAWPS-Single and SVAMP with nine baseline models show that augmentation improves both equation and answer accuracy, with the strongest gains when combining multiple augmentation strategies. The in-context learning approach preserves problem semantics while enabling controlled numerical modifications to enhance generalization, and results indicate practical benefits for MWP solvers. Overall, the work demonstrates robust improvements in MWP performance and suggests potential applicability to other languages and larger LLMs.

Abstract

Math Word Problem (MWP) solving presents a challenging task in Natural Language Processing (NLP). This study aims to provide MWP solvers with a more diverse training set, ultimately improving their ability to solve various math problems. We propose several methods for data augmentation by modifying the problem texts and equations, such as synonym replacement, rule-based: question replacement, and rule based: reversing question methodologies over two English MWP datasets. This study extends by introducing a new in-context learning augmentation method, employing the Llama-7b language model. This approach involves instruction-based prompting for rephrasing the math problem texts. Performance evaluations are conducted on 9 baseline models, revealing that augmentation methods outperform baseline models. Moreover, concatenating examples generated by various augmentation methods further improves performance.