SWL-Adapt: An Unsupervised Domain Adaptation Model with Sample Weight Learning for Cross-User Wearable Human Activity Recognition

TL;DR

SWL-Adapt is an unsupervised domain adaptation model that learns sample weights for cross-user wearable human activity recognition, significantly improving performance by differentiating samples based on their importance.

Contribution

The paper introduces a novel sample weight learning approach with meta-optimization for unsupervised domain adaptation in WHAR, outperforming existing fixed-rule methods.

Findings

Achieves state-of-the-art accuracy and macro F1 scores on three public datasets.

Outperforms baseline models by 3.1% in accuracy and 5.3% in macro F1 score.

Demonstrates the effectiveness of sample weight learning in cross-user WHAR.

Abstract

In practice, Wearable Human Activity Recognition (WHAR) models usually face performance degradation on the new user due to user variance. Unsupervised domain adaptation (UDA) becomes the natural solution to cross-user WHAR under annotation scarcity. Existing UDA models usually align samples across domains without differentiation, which ignores the difference among samples. In this paper, we propose an unsupervised domain adaptation model with sample weight learning (SWL-Adapt) for cross-user WHAR. SWL-Adapt calculates sample weights according to the classification loss and domain discrimination loss of each sample with a parameterized network. We introduce the meta-optimization based update rule to learn this network end-to-end, which is guided by meta-classification loss on the selected pseudo-labeled target samples. Therefore, this network can fit a weighting function according to the cross-user WHAR task at hand, which is superior to existing sample differentiation rules fixed for special scenarios. Extensive experiments on three public WHAR datasets demonstrate that SWL-Adapt achieves the state-of-the-art performance on the cross-user WHAR task, outperforming the best baseline by an average of 3.1% and 5.3% in accuracy and macro F1 score, respectively.