Data-Driven Methods for Solving Algebra Word Problems
This work addresses the problem of automated math problem-solving for educational or AI applications, but it is incremental as it builds on existing data-driven approaches without introducing a major breakthrough.
The paper tackled solving algebra word problems using data-driven methods, finding that well-tuned neural equation classifiers outperform more complex models like sequence-to-sequence and self-attention on large-scale datasets, though it notes that semantic and world knowledge is needed for further improvements.
We explore contemporary, data-driven techniques for solving math word problems over recent large-scale datasets. We show that well-tuned neural equation classifiers can outperform more sophisticated models such as sequence to sequence and self-attention across these datasets. Our error analysis indicates that, while fully data driven models show some promise, semantic and world knowledge is necessary for further advances.