Commonsense Justification for Action Explanation

Shaohua Yang, Qiaozi Gao, Sari Sadiya, Joyce Chai
2018 Proceedings of the 2018 Conference on Empirical Methods in Natural Language Processing  
To enable collaboration and communication between humans and agents, this paper investigates learning to acquire commonsense evidence for action justification. In particular, we have developed an approach based on the generative Conditional Variational Autoencoder (CVAE) that models object relations/attributes of the world as latent variables and jointly learns a performer that predicts actions and an explainer that gathers commonsense evidence to justify the action. Our empirical results have
more » ... hown that, compared to a typical attention-based model, CVAE achieves significantly higher performance in both action prediction and justification. A human subject study further shows that the commonsense evidence gathered by CVAE can be communicated to humans to achieve a significantly higher common ground between humans and agents. • The bread is next to the bread. • The bread is on the rack. • The bread is on the pan. • The man has keys. • The man has the band. • The bread is on the rack. • The bread is on the pan. • The bread is on the tray. • The bread is next to the bread. • The bread is baked. • The bread is baked. • The bread is next to the bread. • The person is pushing the tray. • The bread is on the pan. • The bread is on the rack. • The bread is on the tray. • The person is pushing the tray.
doi:10.18653/v1/d18-1283 dblp:conf/emnlp/YangGSC18 fatcat:tujk3fb32jdptfzw6vythpalay