ReasonFuse: Reason Path Driven and Global–Local Fusion Network for Numerical Table-Text Question Answering
Yuancheng Xia,Feng Li,Qing Liu,Li Jin,Zequn Zhang,Xian Sun,Lixu Shao
DOI: https://doi.org/10.1016/j.neucom.2022.09.046
IF: 6
2023-01-07
Neurocomputing
Abstract:Numerical Table-Text Question Answering aims to predict the program over heterogeneous tabular and textual information, which has recently attracted strong attention in the NLP research community. Current state-of-the-art methods typically use a Seq2Seq-based model to predict programs, which may suffer from the inherent limitations of error propagation. Also, existing models just combine table-text context information and previously predicted operation when reasoning. This approach may lose information over long distances especially when the program is long, and fail to exploit the inter-dependencies between operations in the program. In this paper, we propose a joint Reason Path Driven and Global–Local Fusion Network (ReasonFuse) to alleviate the above problems. First of all, inspired by the intuition that operator prediction is a classification task while operand prediction is an extraction task, we propose a reason path driven architecture that predicts operators first and then guides the operands prediction jointly. After that, the global–local information fusion mechanism is designed to aggregate the context information and global–local program information to mitigate the loss of information over long distances. Last but not least, the soft-dropout mechanism is applied when training to improve the ability and generalization of the model. We conduct extensive experiments on the FinQA and MathQA datasets, our model achieves significant improvements as compared to all published results. The ablation study and case study verify the effectiveness, generalization, and interpretability of our model.
computer science, artificial intelligence