Representation Collapsing Problems in Vector Quantization

TL;DR

This paper investigates the phenomenon of representation collapse in vector quantization within generative models, identifying causes and proposing mitigation strategies to improve the diversity and discriminative power of learned representations.

Contribution

It is the first comprehensive study to analyze representation collapsing problems in vector quantization, revealing causes and suggesting solutions.

Findings

Restricted initialization leads to codebook token collapse

Limited encoder capacity causes embedding collapse

Proposed mitigation strategies can reduce collapse severity

Abstract

Vector quantization is a technique in machine learning that discretizes continuous representations into a set of discrete vectors. It is widely employed in tokenizing data representations for large language models, diffusion models, and other generative models. Despite its prevalence, the characteristics and behaviors of vector quantization in generative models remain largely underexplored. In this study, we investigate representation collapse in vector quantization - a critical degradation where codebook tokens or latent embeddings lose their discriminative power by converging to a limited subset of values. This collapse fundamentally compromises the model's ability to capture diverse data patterns. By leveraging both synthetic and real datasets, we identify the severity of each type of collapses and triggering conditions. Our analysis reveals that restricted initialization and limited encoder capacity result in tokens collapse and embeddings collapse. Building on these findings, we propose potential solutions aimed at mitigating each collapse. To the best of our knowledge, this is the first comprehensive study examining representation collapsing problems in vector quantization.