paper-with-me

Papers

EdVAE: Mitigating Codebook Collapse with Evidential Discrete Variational Autoencoders

2023-10-09 · Gulcin Baykal, Melih Kandemir, Gozde Unal

Codebook collapse is a common problem in training deep generative models with discrete representation spaces like Vector Quantized Variational Autoencoders (VQ-VAEs). We observe that the same problem arises for the alternatively designed discrete variational autoencoders (dVAEs) whose encoder directly learns a distribution over the codebook embeddings to represent the data. We hypothesize that using the softmax function to obtain a probability distribution causes the codebook collapse by assigning overconfident probabilities to the best matching codebook elements. In this paper, we propose a novel way to incorporate evidential deep learning (EDL) instead of softmax to combat the codebook collapse problem of dVAE. We evidentially monitor the significance of attaining the probability distribution over the codebook embeddings, in contrast to softmax usage. Our experiments using various datasets show that our model, called EdVAE, mitigates codebook collapse while improving the reconstruction performance, and enhances the codebook usage compared to dVAE and VQ-VAE based models. Our code can be found at https://github.com/ituvisionlab/EdVAE .

📄 PDF Abstract BibTeX arXiv:2310.05718

Code (1)

ituvisionlab/edvae 공식 구현 pytorch

Methods 이 논문이 사용한 방법론

VQ-VAE VQ-VAE is a type of variational autoencoder that uses vector quantisation to obtain a discrete latent representation. It differs from…
Softmax The Softmax output function transforms a previous layer's output into a vector of probabilities. It is commonly used for multiclass classification. Given an input vector $x$…

Similar Papers 제목 키워드 기반

Early Quantization Shrinks Codebook: A Simple Fix for Diversity-Preserving Tokenization

2026-03-17 · Wenhao Zhao, Qiran Zou, Rushi Shah, Yudi Wu 외 arxiv

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, diffu…

Representation Collapsing Problems in Vector Quantization

2024-11-25 · Wenhao Zhao, Qiran Zou, Rushi Shah, Dianbo Liu

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, diffu…

Quantization

NSVQ: Mitigating Codebook Collapse by Stabilizing Encoder Drift in Vector Quantization

2026-06-09 · Hao Lu, Yongxin Guo, Onur Koyun, Zhengjie Zhu 외 arxiv

Vector quantization is central to modern generative modeling pipelines, but large-codebook VQ models often suffer from codebook collapse. We identify encoder drift as a key driver of this failure: as the encoder moves th…

Codebook Transfer with Part-of-Speech for Vector-Quantized Image Modeling

2024-03-15 · CVPR 2024 1 · Baoquan Zhang, Huaibin Wang, Luo Chuyao, Xutao Li 외

Vector-Quantized Image Modeling (VQIM) is a fundamental research problem in image synthesis, which aims to represent an image with a discrete token sequence. Existing studies effectively address this problem by learning …

Image Generation

Continuous First, Discrete Later: VQ-VAEs Without Dimensional Collapse

2026-05-07 · Xinyu Zhao, Nikita Karagodin, Hamed Hassani, Sinan Hersek 외 arxiv

While many approaches to improve VQ-VAE performance focus on codebook size and utilization, the effect of dimensional collapse, where trained VQ-VAE representations live in an extremely low-dimensional subspace (1-2% of …