A Novel Bilingual Word Embedding Method for Lexical Translation Using Bilingual Sense Clique

TL;DR

This paper introduces a new bilingual sense clique approach for word embedding that captures deeper semantic relationships, improving translation accuracy without needing separate bilingual projections.

Contribution

It proposes a latent bilingual sense clique derived from maximal complete sub-graphs, enabling more effective bilingual embeddings without separate projection steps.

Findings

Outperforms existing bilingual word embedding methods in lexicon translation tasks.

Uses maximum complete sub-graphs of PMI-based graphs to derive bilingual sense units.

Demonstrates effectiveness across multiple dimension reduction techniques.

Abstract

Most of the existing methods for bilingual word embedding only consider shallow context or simple co-occurrence information. In this paper, we propose a latent bilingual sense unit (Bilingual Sense Clique, BSC), which is derived from a maximum complete sub-graph of pointwise mutual information based graph over bilingual corpus. In this way, we treat source and target words equally and a separated bilingual projection processing that have to be used in most existing works is not necessary any more. Several dimension reduction methods are evaluated to summarize the BSC-word relationship. The proposed method is evaluated on bilingual lexicon translation tasks and empirical results show that bilingual sense embedding methods outperform existing bilingual word embedding methods.