A paper co-authored by Yamada and the University of Maryland team was accepted at Transactions of the Association for Computational Linguistics (TACL), a leading journal in computational linguistics and natural language processing.
In this paper, a new method for developing question answering datasets that are difficult for current state-of-the-art AI to solve is proposed.
Abou the paper
Trick Me If You Can: Human-in-the-loop Generation of Adversarial Examples for Question Answering
Eric Wallace (U. Maryland), Pedro Rodriguez (U. Maryland), Shi Feng (U. Maryland), Ikuya Yamada, Jordan Boyd-Graber (U. Maryland) Transactions of the Association for Computational Linguistics (TACL), 2019
Access to paper: https://www.transacl.org/ojs/index.php/tacl/article/view/1711