RQ-MoE: Residual Quantization via Mixture of Experts for Efficient Input-Dependent Vector Compression

TL;DR

The paper introduces RQ-MoE, a novel vector quantization framework that uses a mixture of experts to enable input-dependent codebook adaptation, resulting in faster decoding and improved performance.

Contribution

It proposes a new residual quantization method combining a two-level MoE with dual-stream quantization, allowing dynamic codebook construction and parallel decoding.

Findings

Achieves state-of-the-art or comparable performance in reconstruction and retrieval tasks.

Provides 6x-14x faster decoding than previous vector quantization methods.

Theoretically unifies standard Residual Quantization and QINCo as special cases.

Abstract

Vector quantization is a fundamental tool for compressing high-dimensional embeddings, yet existing multi-codebook methods rely on static codebooks that limit expressiveness under heterogeneous data geometry. While recent dynamic quantizers like QINCo adapt codebooks to individual inputs and improve expressiveness, their strict sequential dependencies create decoding bottlenecks. We propose Residual Quantization via Mixture of Experts (RQ-MoE), a framework combining a two-level MoE with dual-stream quantization to enable input-dependent codebook adaptation for efficient vector quantization. RQ-MoE enables dynamic codebook construction and decouples instruction from quantization, facilitating parallel decoding. Theoretically, we show that standard Residual Quantization and QINCo can be recovered as constrained special cases of RQ-MoE, and derive a guideline for setting expert dimensionality in RQ-MoE. Extensive experiments show that RQ-MoE achieves state-of-the-art or on-par performance in reconstruction and retrieval, while providing 6x-14x faster decoding than prior vector quantization methods. The implementation is available at https://github.com/KDEGroup/RQ-MoE.