July 16, 2018
Hearst Patterns Revisited: Automatic Hypernym Detection from Large Text Corpora
Association for Computational Linguistics (ACL)
Methods for unsupervised hypernym detection may broadly be categorized according to two paradigms: pattern-based and distributional methods. In this paper, we study the performance of both approaches on several hypernymy tasks and find that simple pattern-based methods consistently outperform distributional methods on common benchmark datasets.
By: Stephen Roller, Douwe Kiela, Maximilian Nickel