Hearst Patterns Revisited: Automatic Hypernym Detection from Large Text Corpora
This work addresses hypernym detection for natural language processing, but it is incremental as it revisits and compares existing paradigms without introducing new methods.
The paper tackled the problem of unsupervised hypernym detection by comparing pattern-based and distributional methods, finding that simple pattern-based approaches consistently outperform distributional methods on benchmark datasets, highlighting their superior contextual constraints.
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. Our results show that pattern-based models provide important contextual constraints which are not yet captured in distributional methods.