Hashing with Mutual Information

TL;DR

This paper introduces a supervised hashing method that optimizes mutual information within deep neural networks to produce high-quality binary embeddings for fast nearest neighbor retrieval in large, high-dimensional datasets.

Contribution

The paper proposes a novel supervised hashing approach based on mutual information optimization, improving neighborhood structure clarity and retrieval accuracy.

Findings

Effective on four image retrieval benchmarks including ImageNet.

Outperforms existing hashing methods in retrieval tasks.

Enhances neighborhood structure in learned embeddings.

Abstract

Binary vector embeddings enable fast nearest neighbor retrieval in large databases of high-dimensional objects, and play an important role in many practical applications, such as image and video retrieval. We study the problem of learning binary vector embeddings under a supervised setting, also known as hashing. We propose a novel supervised hashing method based on optimizing an information-theoretic quantity: mutual information. We show that optimizing mutual information can reduce ambiguity in the induced neighborhood structure in the learned Hamming space, which is essential in obtaining high retrieval performance. To this end, we optimize mutual information in deep neural networks with minibatch stochastic gradient descent, with a formulation that maximally and efficiently utilizes available supervision. Experiments on four image retrieval benchmarks, including ImageNet, confirm the effectiveness of our method in learning high-quality binary embeddings for nearest neighbor retrieval.