Unsupervised Word Sense Disambiguation Rivaling Supervised Methods

1 Introduction

1. What does Yarowsky (1995) achieve?
   An unsupervised word sense disambiguation algorithm, trained on unannotated text, that rivals supervised methods needing hand-tagged data (accuracy above 96%), by bootstrapping from a few seeds with two constraints: one sense per collocation and one sense per discourse.
   • Yarowsky, 1995, Unsupervised Word Sense Disambiguation Rivaling Supervised Methods (Proceedings of the 33rd Annual Meeting of ACL), §1
2. State the two properties of language Yarowsky’s algorithm exploits.
   One sense per collocation: nearby words give strong, consistent clues to a word’s sense, depending on distance, order and syntactic relation. One sense per discourse: a word’s sense is highly consistent within a given document.
   • Yarowsky, 1995, Unsupervised Word Sense Disambiguation Rivaling Supervised Methods (Proceedings of the 33rd Annual Meeting of ACL), §1
3. How does sense disambiguation differ from sense induction?
   Disambiguation assigns each instance of a word to established senses (e.g. a dictionary’s); induction partitions instances into clusters by distributional similarity, which may not match standard senses.
   • Yarowsky, 1995, Unsupervised Word Sense Disambiguation Rivaling Supervised Methods (Proceedings of the 33rd Annual Meeting of ACL), §1
4. Is one sense per discourse used as a hard constraint?
   No: it affects classification probabilistically and can be overridden when local evidence is strong.
   • Yarowsky, 1995, Unsupervised Word Sense Disambiguation Rivaling Supervised Methods (Proceedings of the 33rd Annual Meeting of ACL), §1