Relations such as Hypernymy: Identifying and Exploiting Hearst Patterns in Distributional Vectors for Lexical Entailment
This work addresses lexical entailment prediction for natural language processing, offering an incremental improvement by leveraging known patterns in distributional vectors.
The paper tackled the problem of predicting lexical entailment by analyzing an existing model and discovering it learns hypernyms via Hearst patterns, then introduced a novel model that extracts these features and combines them with other methods, achieving matching or superior performance on multiple datasets.
We consider the task of predicting lexical entailment using distributional vectors. We perform a novel qualitative analysis of one existing model which was previously shown to only measure the prototypicality of word pairs. We find that the model strongly learns to identify hypernyms using Hearst patterns, which are well known to be predictive of lexical relations. We present a novel model which exploits this behavior as a method of feature extraction in an iterative procedure similar to Principal Component Analysis. Our model combines the extracted features with the strengths of other proposed models in the literature, and matches or outperforms prior work on multiple data sets.