Syntactic Patterns Improve Information Extraction for Medical Search

TL;DR

This paper shows that incorporating syntactic pattern features into sequence tagging models significantly enhances the extraction of medically relevant information from literature, capturing contextual nuances beyond individual words.

Contribution

It introduces a method to encode syntactic patterns into models, improving information extraction accuracy in medical literature search.

Findings

Syntactic pattern features improve model performance

Learned representations differ from unigram embeddings

Patterns capture contextual information

Abstract

Medical professionals search the published literature by specifying the type of patients, the medical intervention(s) and the outcome measure(s) of interest. In this paper we demonstrate how features encoding syntactic patterns improve the performance of state-of-the-art sequence tagging models (both linear and neural) for information extraction of these medically relevant categories. We present an analysis of the type of patterns exploited, and the semantic space induced for these, i.e., the distributed representations learned for identified multi-token patterns. We show that these learned representations differ substantially from those of the constituent unigrams, suggesting that the patterns capture contextual information that is otherwise lost.