Using logical prompts to automatically solve math word problems

Automatic solving of math word problems (MWP) has been a popular research direction in recent years. Due to the complex mathematical logic relationships involved, many studies have been updating methods to seek the best way to solve problems. In this scenario, this article proposes an end-to-end model LPSAU, guided by logical prompts, to enable the model to better understand the mathematical semantic information within it. The model proposed in this article is used to solve MWP with two unknowns. To verify the effectiveness of the solution, we introduced a new challenging dataset LF and conducted rich quantitative and comparative experiments. The experimental results showed that our model improved significantly in accuracy compared to the methods proposed by previous researchers.