Qinco2: Vector Compression and Search with Improved Implicit Neural Codebooks

TL;DR

QINCo2 advances vector compression and search by enhancing neural codebook-based quantization with pre-selection, beam-search, and optimized training, achieving state-of-the-art results on large datasets.

Contribution

It introduces QINCo2, an improved neural codebook-based vector quantization method with novel encoding, decoding, and training techniques for better compression and search accuracy.

Findings

34% reduction in MSE for 16-byte vectors on BigANN

24% improvement in search accuracy with 8-byte encodings on Deep1M

Outperforms previous methods in both compression and nearest neighbor search

Abstract

Vector quantization is a fundamental technique for compression and large-scale nearest neighbor search. For high-accuracy operating points, multi-codebook quantization associates data vectors with one element from each of multiple codebooks. An example is residual quantization (RQ), which iteratively quantizes the residual error of previous steps. Dependencies between the different parts of the code are, however, ignored in RQ, which leads to suboptimal rate-distortion performance. QINCo recently addressed this inefficiency by using a neural network to determine the quantization codebook in RQ based on the vector reconstruction from previous steps. In this paper we introduce QINCo2 which extends and improves QINCo with (i) improved vector encoding using codeword pre-selection and beam-search, (ii) a fast approximate decoder leveraging codeword pairs to establish accurate short-lists for search, and (iii) an optimized training procedure and network architecture. We conduct experiments on four datasets to evaluate QINCo2 for vector compression and billion-scale nearest neighbor search. We obtain outstanding results in both settings, improving the state-of-the-art reconstruction MSE by 34% for 16-byte vector compression on BigANN, and search accuracy by 24% with 8-byte encodings on Deep1M.