Recently, attempting to model texts as graph structure and introducing graph neural networks to deal with it has become a trend in many NLP research areas. In this paper, we investigate whether the graph structure is necessary for multi-hop question answering. Our analysis is centered on HotpotQA. We construct a strong baseline model to establish that, with the proper use of pre-trained models, graph structure may not be necessary for multi-hop question answering. We point out that both graph structure and adjacency matrix are task-related prior knowledge, and graph-attention can be considered as a special case of self-attention. Experiments and visualized analysis demonstrate that graph-attention or the entire graph structure can be replaced by self-attention or Transformers.
@article{arxiv.2004.03096,
title = {Is Graph Structure Necessary for Multi-hop Question Answering?},
author = {Nan Shao and Yiming Cui and Ting Liu and Shijin Wang and Guoping Hu},
journal= {arXiv preprint arXiv:2004.03096},
year = {2020}
}