Pairwise Similarity Learning is SimPLE

TL;DR

This paper introduces SimPLE, a simple proxy-free method for pairwise similarity learning that outperforms current state-of-the-art techniques across various open-set recognition tasks.

Contribution

The paper proposes a novel, simple, proxy-free approach called SimPLE for pairwise similarity learning, applicable to multiple challenging recognition tasks.

Findings

SimPLE outperforms existing methods on large-scale benchmarks.

It generalizes well without feature normalization or angular margin.

The approach is effective across face recognition, image retrieval, and speaker verification.

Abstract

In this paper, we focus on a general yet important learning problem, pairwise similarity learning (PSL). PSL subsumes a wide range of important applications, such as open-set face recognition, speaker verification, image retrieval and person re-identification. The goal of PSL is to learn a pairwise similarity function assigning a higher similarity score to positive pairs (i.e., a pair of samples with the same label) than to negative pairs (i.e., a pair of samples with different label). We start by identifying a key desideratum for PSL, and then discuss how existing methods can achieve this desideratum. We then propose a surprisingly simple proxy-free method, called SimPLE, which requires neither feature/proxy normalization nor angular margin and yet is able to generalize well in open-set recognition. We apply the proposed method to three challenging PSL tasks: open-set face recognition, image retrieval and speaker verification. Comprehensive experimental results on large-scale benchmarks show that our method performs significantly better than current state-of-the-art methods.