Learning Deep Embeddings with Histogram Loss

TL;DR

This paper introduces a parameter-free histogram loss for deep embedding learning, which estimates similarity distributions to improve embedding quality across various datasets using a simple, differentiable approach.

Contribution

The paper proposes a novel histogram-based loss function that does not require parameter tuning and effectively learns deep embeddings by modeling similarity distributions.

Findings

Outperforms recent alternative loss functions in experiments

Works well across multiple datasets and problems

Uses simple, differentiable histogram operations for optimization

Abstract

We suggest a loss for learning deep embeddings. The new loss does not introduce parameters that need to be tuned and results in very good embeddings across a range of datasets and problems. The loss is computed by estimating two distribution of similarities for positive (matching) and negative (non-matching) sample pairs, and then computing the probability of a positive pair to have a lower similarity score than a negative pair based on the estimated similarity distributions. We show that such operations can be performed in a simple and piecewise-differentiable manner using 1D histograms with soft assignment operations. This makes the proposed loss suitable for learning deep embeddings using stochastic optimization. In the experiments, the new loss performs favourably compared to recently proposed alternatives.