Are NLP Models really able to Solve Simple Math Word Problems?

TL;DR

This paper reveals that current NLP models for elementary math word problems often rely on superficial heuristics, and introduces a new challenging dataset, SVAMP, showing room for improvement in solving simple MWPs.

Contribution

The paper demonstrates the limitations of existing NLP solvers on simple MWPs and introduces SVAMP, a challenging dataset to evaluate true problem-solving capabilities.

Findings

Existing solvers can solve MWPs without understanding the question.

Bag-of-words models achieve high accuracy on benchmark datasets.

Performance drops significantly on the SVAMP dataset.

Abstract

The problem of designing NLP solvers for math word problems (MWP) has seen sustained research activity and steady gains in the test accuracy. Since existing solvers achieve high performance on the benchmark datasets for elementary level MWPs containing one-unknown arithmetic word problems, such problems are often considered "solved" with the bulk of research attention moving to more complex MWPs. In this paper, we restrict our attention to English MWPs taught in grades four and lower. We provide strong evidence that the existing MWP solvers rely on shallow heuristics to achieve high performance on the benchmark datasets. To this end, we show that MWP solvers that do not have access to the question asked in the MWP can still solve a large fraction of MWPs. Similarly, models that treat MWPs as bag-of-words can also achieve surprisingly high accuracy. Further, we introduce a challenge dataset, SVAMP, created by applying carefully chosen variations over examples sampled from existing datasets. The best accuracy achieved by state-of-the-art models is substantially lower on SVAMP, thus showing that much remains to be done even for the simplest of the MWPs.