Expansion-by-analogy: A vector symbolic approach to semantic search

Abstract

In this paper, we develop an approach to semantic search that utilizes high-dimensional vector representations to infer the nature of the relationship between query concepts and other concepts in relevant documents. We do so by incorporating outside knowledge drawn from tens of millions of concept-relation-concept triplets, known as semantic predications, extracted from the biomedical literature using a Natural Language Processing (NLP) system called SemRep. Inference is accomplished in high-dimensional space using Expansion-by-Analogy, a novel analogical approach to pseudo-relevance feedback, in which the relationships between query concepts and other concepts in documents they occur in guide the query expansion process. The semantic vector based approaches developed in this work show improvements in performance over a baseline bag-of-concepts model, and these improvements are most pronounced on queries that are not conducive to keyword-based search.