Bag-of-Vector Embeddings of Dependency Graphs for Semantic Induction

TL;DR

This paper introduces bag-of-vector embeddings for dependency graphs, enabling flexible, structure-agnostic semantic representations that improve unsupervised semantic tasks in NLP.

Contribution

It proposes a novel nonparametric embedding method for arbitrary linguistic graphs using tensor factorisation, extending vector space models.

Findings

Effective embedding of dependency graphs demonstrated

Improved performance on semantic similarity tasks

Flexible representation not tied to specific graph structures

Abstract

Vector-space models, from word embeddings to neural network parsers, have many advantages for NLP. But how to generalise from fixed-length word vectors to a vector space for arbitrary linguistic structures is still unclear. In this paper we propose bag-of-vector embeddings of arbitrary linguistic graphs. A bag-of-vector space is the minimal nonparametric extension of a vector space, allowing the representation to grow with the size of the graph, but not tying the representation to any specific tree or graph structure. We propose efficient training and inference algorithms based on tensor factorisation for embedding arbitrary graphs in a bag-of-vector space. We demonstrate the usefulness of this representation by training bag-of-vector embeddings of dependency graphs and evaluating them on unsupervised semantic induction for the Semantic Textual Similarity and Natural Language Inference tasks.