From Logical to Distributional Models

TL;DR

This paper bridges logical and distributional semantic models by translating logical reasoning into algebraic vector operations, enabling detailed comparison of sentence meanings at the word level.

Contribution

It introduces a method to transfer logic-based reasoning into vector space models, facilitating a unified approach to semantic modeling.

Findings

Reformulation of quantum logic operations as algebraic vector operations

A mapping from functional to vector space models for sentence comparison

Enhanced ability to compare sentence meanings word by word

Abstract

The paper relates two variants of semantic models for natural language, logical functional models and compositional distributional vector space models, by transferring the logic and reasoning from the logical to the distributional models. The geometrical operations of quantum logic are reformulated as algebraic operations on vectors. A map from functional models to vector space models makes it possible to compare the meaning of sentences word by word.