Explaining, Analyzing, and Probing Representations of Self-Supervised Learning Models for Sensor-based Human Activity Recognition

TL;DR

This paper compares self-supervised and supervised models for sensor-based human activity recognition, revealing that SSL models are more noise-robust while supervised models better encode activity features.

Contribution

The study provides a comprehensive analysis of SSL representations in HAR, highlighting their robustness and differences from supervised models using explainability and probing methods.

Findings

SSL representations are more robust to input noise

Supervised features are more homogeneous across subjects

Supervised models better encode activity-specific features

Abstract

In recent years, self-supervised learning (SSL) frameworks have been extensively applied to sensor-based Human Activity Recognition (HAR) in order to learn deep representations without data annotations. While SSL frameworks reach performance almost comparable to supervised models, studies on interpreting representations learnt by SSL models are limited. Nevertheless, modern explainability methods could help to unravel the differences between SSL and supervised representations: how they are being learnt, what properties of input data they preserve, and when SSL can be chosen over supervised training. In this paper, we aim to analyze deep representations of two recent SSL frameworks, namely SimCLR and VICReg. Specifically, the emphasis is made on (i) comparing the robustness of supervised and SSL models to corruptions in input data; (ii) explaining predictions of deep learning models using saliency maps and highlighting what input channels are mostly used for predicting various activities; (iii) exploring properties encoded in SSL and supervised representations using probing. Extensive experiments on two single-device datasets (MobiAct and UCI-HAR) have shown that self-supervised learning representations are significantly more robust to noise in unseen data compared to supervised models. In contrast, features learnt by the supervised approaches are more homogeneous across subjects and better encode the nature of activities.