Auto-Augmentation Contrastive Learning for Wearable-based Human Activity Recognition

2023-12-05

IEEE International Conference on Bioinformatics and Biomedicine

AI Summary

This work addresses the limitations of current human activity recognition (HAR) systems in wearable devices, where sensor signals exhibit semantic sparsity and contrastive learning relies on handcrafted data augmentation strategies that hinder generalization. To overcome these challenges, we propose AutoCL, an end-to-end automated augmentation framework for contrastive learning in HAR. AutoCL integrates an adaptive data augmentation mechanism into a siamese network architecture by embedding a generator that learns augmentation policies in the latent space. It further incorporates stop-gradient operations and feature decorrelation to effectively suppress sensor noise and redundant information. Evaluated on four mainstream HAR datasets, AutoCL significantly outperforms existing unsupervised methods, achieving substantial improvements in accuracy, generalization, and robustness.

Technology Category

• Machine Learning: Deep Generative Models & Autoencoders
• Humans and AI: Human-in-the-loop Machine Learning
• Computer Vision: Video Understanding & Activity Analysis

Application Category

• Search and Retrieval-Augmented AI: Web learning to rank, online learning, and counterfactual learning for ranking
• Responsible Web: Machine-in-the-loop, human agency and autonomy
• User Modeling, Personalization and Recommendation: Attacks and countermeasures in recommendation systems

Abstract

For low-semantic sensor signals from human activity recognition (HAR), contrastive learning (CL) is essential to implement novel applications or generic models without manual annotation, which is a high-performance self-supervised learning (SSL) method. However, CL relies heavily on data augmentation for pairwise comparisons. Especially for low semantic data in the HAR area, conducting good performance augmentation strategies in pretext tasks still rely on manual attempts lacking generalizability and flexibility. To reduce the augmentation burden, we propose an end-to-end auto-augmentation contrastive learning (AutoCL) method for wearable-based HAR. AutoCL is based on a Siamese network architecture that shares the parameters of the backbone and with a generator embedded to learn auto-augmentation. AutoCL trains the generator based on the representation in the latent space to overcome the disturbances caused by noise and redundant information in raw sensor data. The architecture empirical study indicates the effectiveness of this design. Furthermore, we propose a stop-gradient design and correlation reduction strategy in AutoCL to enhance encoder representation learning. Extensive experiments based on four wide-used HAR datasets demonstrate that the proposed AutoCL method significantly improves recognition accuracy compared with other SOTA methods.

Problem

• contrastive learning
• data augmentation
• human activity recognition
• self-supervised learning
• wearable sensors

Innovation

• auto-augmentation
• contrastive learning
• wearable-based HAR
• Siamese network
• self-supervised learning