2000 character limit reached
Question Directed Graph Attention Network for Numerical Reasoning over Text (2009.07448v2)
Published 16 Sep 2020 in cs.AI
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
- Kunlong Chen (11 papers)
- Weidi Xu (10 papers)
- Xingyi Cheng (20 papers)
- Zou Xiaochuan (1 paper)
- Yuyu Zhang (24 papers)
- Le Song (140 papers)
- Taifeng Wang (22 papers)
- Yuan Qi (85 papers)
- Wei Chu (118 papers)