Using Representation Expressiveness and Learnability to Evaluate Self-Supervised Learning Methods

TL;DR

This paper introduces CLID, a new evaluation method for self-supervised learning models based on representation expressiveness and learnability, which correlates well with model performance both in-distribution and out-of-domain.

Contribution

The paper proposes CLID, combining Intrinsic Dimension and Cluster Learnability, as a novel, label-free evaluation metric for SSL models that outperforms existing schemes.

Findings

CLID correlates better with in-distribution performance.

CLID effectively predicts out-of-domain transfer performance.

The approach is validated across diverse SSL algorithms.

Abstract

We address the problem of evaluating the quality of self-supervised learning (SSL) models without access to supervised labels, while being agnostic to the architecture, learning algorithm or data manipulation used during training. We argue that representations can be evaluated through the lens of expressiveness and learnability. We propose to use the Intrinsic Dimension (ID) to assess expressiveness and introduce Cluster Learnability (CL) to assess learnability. CL is measured in terms of the performance of a KNN classifier trained to predict labels obtained by clustering the representations with K-means. We thus combine CL and ID into a single predictor -- CLID. Through a large-scale empirical study with a diverse family of SSL algorithms, we find that CLID better correlates with in-distribution model performance than other competing recent evaluation schemes. We also benchmark CLID on out-of-domain generalization, where CLID serves as a predictor of the transfer performance of SSL models on several visual classification tasks, yielding improvements with respect to the competing baselines.