[KongchangAI]
Model残差量化变分自编码器 / Residual Quantization VAE

RQ-VAE

A deep learning model combining residual quantization with a variational autoencoder, used to compress continuous vectors into hierarchical discrete codeword sequences. It is commonly applied for semantic ID encoding in generative recommendation systems.

Timeline (last 90 days)

Oct 5

Semantic IDs compress item content features or collaborative filtering embeddings into discrete codeword sequences through hierarchical quantization (e.g., RQ-VAE), with content-similar items sharing a common prefix.

Unverified50%

All Facts (1)

Unverified

Semantic IDs compress item content features or collaborative filtering embeddings into discrete codeword sequences through hierarchical quantization (e.g., RQ-VAE), with content-similar items sharing a common prefix.

50%

Source Articles

HASTE: Hierarchical and Attribute Supervision Doubles Generative Recommendation Training SpeedOct 14Yandex Sona: How a Single Transformer Replaced 15+ Recommendation ComponentsOct 5