Using dependency parsing for few-shot learning in distributional semantics

TL;DR

This paper investigates leveraging dependency parsing information to improve few-shot learning of word meanings, introducing new dependency-based embedding models and methods to enhance existing additive models.

Contribution

It introduces novel dependency-based embedding models and two methods that incorporate dependency information to advance few-shot learning in distributional semantics.

Findings

Dependency-based embeddings improve few-shot learning performance.

Two new methods effectively utilize dependency information.

Enhanced additive baseline models outperform traditional approaches.

Abstract

In this work, we explore the novel idea of employing dependency parsing information in the context of few-shot learning, the task of learning the meaning of a rare word based on a limited amount of context sentences. Firstly, we use dependency-based word embedding models as background spaces for few-shot learning. Secondly, we introduce two few-shot learning methods which enhance the additive baseline model by using dependencies.