Compressing Word Embeddings via Deep Compositional Code Learning

TL;DR

This paper introduces a novel method for compressing word embeddings using deep compositional code learning, achieving high compression rates with minimal performance loss across NLP tasks.

Contribution

It proposes a multi-codebook quantization approach with end-to-end learning via Gumbel-softmax, enabling language-independent embedding compression without altering model architecture.

Findings

Achieves up to 98% compression rate in sentiment analysis

Attains 94-99% compression in machine translation

Improves performance slightly by reducing compression rate

Abstract

Natural language processing (NLP) models often require a massive number of parameters for word embeddings, resulting in a large storage or memory footprint. Deploying neural NLP models to mobile devices requires compressing the word embeddings without any significant sacrifices in performance. For this purpose, we propose to construct the embeddings with few basis vectors. For each word, the composition of basis vectors is determined by a hash code. To maximize the compression rate, we adopt the multi-codebook quantization approach instead of binary coding scheme. Each code is composed of multiple discrete numbers, such as (3, 2, 1, 8), where the value of each component is limited to a fixed range. We propose to directly learn the discrete codes in an end-to-end neural network by applying the Gumbel-softmax trick. Experiments show the compression rate achieves 98% in a sentiment analysis task and 94% ~ 99% in machine translation tasks without performance loss. In both tasks, the proposed method can improve the model performance by slightly lowering the compression rate. Compared to other approaches such as character-level segmentation, the proposed method is language-independent and does not require modifications to the network architecture.