Unsupervised Domain Adaptation for Action Recognition via Self-Ensembling and Conditional Embedding Alignment
This work addresses domain discrepancies in wearable human action recognition, which is crucial for real-world deployment where expert annotations are scarce and user variations cause performance drops, though it is incremental as it builds on existing UDA techniques.
The paper tackles the problem of unsupervised domain adaptation for wearable human action recognition by proposing μDAR, a joint optimization architecture that integrates consistency regularization, temporal ensembling, and conditional distribution alignment, resulting in an average macro-F1 score improvement of approximately 4-12% over six state-of-the-art methods across four benchmark datasets.
Recent advancements in deep learning-based wearable human action recognition (wHAR) have improved the capture and classification of complex motions, but adoption remains limited due to the lack of expert annotations and domain discrepancies from user variations. Limited annotations hinder the model's ability to generalize to out-of-distribution samples. While data augmentation can improve generalizability, unsupervised augmentation techniques must be applied carefully to avoid introducing noise. Unsupervised domain adaptation (UDA) addresses domain discrepancies by aligning conditional distributions with labeled target samples, but vanilla pseudo-labeling can lead to error propagation. To address these challenges, we propose $μ$DAR, a novel joint optimization architecture comprised of three functions: (i) consistency regularizer between augmented samples to improve model classification generalizability, (ii) temporal ensemble for robust pseudo-label generation and (iii) conditional distribution alignment to improve domain generalizability. The temporal ensemble works by aggregating predictions from past epochs to smooth out noisy pseudo-label predictions, which are then used in the conditional distribution alignment module to minimize kernel-based class-wise conditional maximum mean discrepancy ($k$CMMD) between the source and target feature space to learn a domain invariant embedding. The consistency-regularized augmentations ensure that multiple augmentations of the same sample share the same labels; this results in (a) strong generalization with limited source domain samples and (b) consistent pseudo-label generation in target samples. The novel integration of these three modules in $μ$DAR results in a range of $\approx$ 4-12% average macro-F1 score improvement over six state-of-the-art UDA methods in four benchmark wHAR datasets