Knowledge guidance combined with end-to-end method to solve math word problem without implicit knowledge

Over the past few years, there has been an increasing fascination with the automated resolution of mathematical challenges. Many researchers have developed their own algorithm models, achieving success on the public dataset Math23k. Specifically, end-to-end algorithm models based on deep learning have shown a significant improvement in solving accuracy compared to other methods. However, these end-to-end deep learning models do not perform well in handling questions that lack implicit information. The reason for the poor performance of existing models is the lack of relevant background knowledge in solving such problems. Therefore, this paper proposes a solving strategy that combines knowledge guidance with an end-to-end approach for mathematical application questions that lack implicit information. The experimental results demonstrate the effectiveness of this method in improving the solving accuracy of various models for mathematical application questions that lack implicit information.