Solving Math Word Problems with Reexamination

TL;DR

This paper introduces a pseudo-dual learning scheme for math word problem solving that enhances existing models by reexamining problems through a number infilling task, improving performance across various solvers.

Contribution

The paper proposes a model-agnostic pseudo-dual learning approach with a scheduled fusion strategy to improve math word problem solvers by reexamining problems during training.

Findings

Proven effective across multiple MWP solvers.

Enhances problem understanding via reexamination process.

Code and models publicly available.

Abstract

Math word problem (MWP) solving aims to understand the descriptive math problem and calculate the result, for which previous efforts are mostly devoted to upgrade different technical modules. This paper brings a different perspective of \textit{reexamination process} during training by introducing a pseudo-dual task to enhance the MWP solving. We propose a pseudo-dual (PseDual) learning scheme to model such process, which is model-agnostic thus can be adapted to any existing MWP solvers. The pseudo-dual task is specifically defined as filling the numbers in the expression back into the original word problem with numbers masked. To facilitate the effective joint learning of the two tasks, we further design a scheduled fusion strategy for the number infilling task, which smoothly switches the input from the ground-truth math expressions to the predicted ones. Our pseudo-dual learning scheme has been tested and proven effective when being equipped in several representative MWP solvers through empirical studies. \textit{The codes and trained models are available at:} \url{https://github.com/steven640pixel/PsedualMWP}. \end{abstract}