Question Directed Graph Attention Network for Numerical Reasoning over Text
This addresses a challenging machine reading comprehension task requiring natural language understanding and arithmetic computation, but appears incremental as it builds on existing graph-based methods for numerical reasoning.
The paper tackles numerical reasoning over text by proposing a heterogeneous graph representation and a question directed graph attention network, achieving results on a benchmark dataset (e.g., DROP or similar, though not specified with numbers in the abstract).
Numerical reasoning over texts, such as addition, subtraction, sorting and counting, is a challenging machine reading comprehension task, since it requires both natural language understanding and arithmetic computation. To address this challenge, we propose a heterogeneous graph representation for the context of the passage and question needed for such reasoning, and design a question directed graph attention network to drive multi-step numerical reasoning over this context graph. The code link is at: https://github.com/emnlp2020qdgat/QDGAT