On Large-Scale Retrieval: Binary or n-ary Coding?

TL;DR

This paper compares binary and n-ary coding methods for large-scale data retrieval, introducing a new n-ary coding technique and analyzing their effectiveness in different retrieval strategies through experiments.

Contribution

It introduces Linear Subspace Quantization (LSQ), a novel n-ary coding method that preserves similarity, and provides a comprehensive comparison of binary and n-ary coding for retrieval tasks.

Findings

n-ary LSQ outperforms others in Distance Estimation.

Binary coding is more effective for Subset Indexing.

Binary LSQ achieves the best accuracy in Subset Indexing.

Abstract

The growing amount of data available in modern-day datasets makes the need to efficiently search and retrieve information. To make large-scale search feasible, Distance Estimation and Subset Indexing are the main approaches. Although binary coding has been popular for implementing both techniques, n-ary coding (known as Product Quantization) is also very effective for Distance Estimation. However, their relative performance has not been studied for Subset Indexing. We investigate whether binary or n-ary coding works better under different retrieval strategies. This leads to the design of a new n-ary coding method, "Linear Subspace Quantization (LSQ)" which, unlike other n-ary encoders, can be used as a similarity-preserving embedding. Experiments on image retrieval show that when Distance Estimation is used, n-ary LSQ outperforms other methods. However, when Subset Indexing is applied, interestingly, binary codings are more effective and binary LSQ achieves the best accuracy.