Deconfounding Causal Inference through Two-Branch Framework with Early-Forking for Sensor-Based Cross-Domain Activity Recognition

2025-07-05

AI Summary

Existing domain generalization (DG) methods for sensor-based human activity recognition (HAR) often neglect causal mechanisms and struggle with cross-domain distribution shifts. To address this, we propose a causally inspired representation disentanglement framework. Methodologically, we design an early-fork dual-branch network that implicitly separates category-informative causal features from domain-specific non-causal features using the Hilbert–Schmidt Independence Criterion (HSIC). A category-aware domain perturbation layer is introduced to prevent representation collapse, and heterogeneous domain sampling enhances disentanglement robustness. Extensive experiments on multiple benchmark datasets demonstrate that our method significantly outperforms 11 state-of-the-art DG approaches across canonical scenarios—including cross-subject, cross-device, and cross-location generalization. Ablation studies and visualization analyses further validate the effectiveness and generalizability of our causal disentanglement strategy.

Technology Category

• Knowledge Representation and Reasoning: Action, Change, and Causality
• Machine Learning: Causal Learning
• Reasoning under Uncertainty: Causality

Application Category

• Search and Retrieval-Augmented AI: Vertical and domain-specific search
• Graph Algorithms and Modeling for the Web: Representation, reconstruction, and subgraph or motif discovery in Web-related graphs
• Economics, Online Markets and Human Computation: Data quality aspects of human-annotated datasets

Abstract

Recently, domain generalization (DG) has emerged as a promising solution to mitigate distribution-shift issue in sensor-based human activity recognition (HAR) scenario. However, most existing DG-based works have merely focused on modeling statistical dependence between sensor data and activity labels, neglecting the importance of intrinsic casual mechanism. Intuitively, every sensor input can be viewed as a mixture of causal (category-aware) and non-causal factors (domain-specific), where only the former affects activity classification judgment. In this paper, by casting such DG-based HAR as a casual inference problem, we propose a causality-inspired representation learning algorithm for cross-domain activity recognition. To this end, an early-forking two-branch framework is designed, where two separate branches are respectively responsible for learning casual and non-causal features, while an independence-based Hilbert-Schmidt Information Criterion is employed to implicitly disentangling them. Additionally, an inhomogeneous domain sampling strategy is designed to enhance disentanglement, while a category-aware domain perturbation layer is performed to prevent representation collapse. Extensive experiments on several public HAR benchmarks demonstrate that our causality-inspired approach significantly outperforms eleven related state-of-the-art baselines under cross-person, cross-dataset, and cross-position settings. Detailed ablation and visualizations analyses reveal underlying casual mechanism, indicating its effectiveness, efficiency, and universality in cross-domain activity recognition scenario.

Problem

• Addresses domain generalization in sensor-based activity recognition
• Separates causal and non-causal features for accurate classification
• Enhances cross-domain performance via disentangled representation learning

Innovation

• Early-forking two-branch framework for feature separation
• Independence-based criterion for disentangling causal features
• Category-aware domain perturbation to prevent collapse