Collaborative Temporal Feature Generation via Critic-Free Reinforcement Learning for Cross-User Sensor-Based Activity Recognition

2026-03-16

AI Summary

This work addresses the challenge of cross-user generalization in human activity recognition using wearable sensors by proposing a collaborative temporal feature generation framework. The approach models universal feature extraction as an autoregressive generation process and introduces a critic-free Group-Relative Policy Optimization algorithm, which replaces value function estimation with intra-group normalization to eliminate distribution-dependent bias. A three-objective reward mechanism—encompassing class discriminability, user invariance, and temporal fidelity—is integrated to optimize a Transformer-based generator for producing robust feature sequences. Evaluated on the DSADS and PAMAP2 datasets, the method achieves cross-user accuracies of 88.53% and 75.22%, respectively, significantly reducing training variance, accelerating convergence, and demonstrating strong generalization across diverse action spaces.

Technology Category

• Humans and AI: Human-Aware Planning and Behavior Prediction
• Machine Learning: Deep Generative Models & Autoencoders
• Natural Language Processing: Generation

Application Category

• User Modeling, Personalization and Recommendation: Fairness-aware retrieval and ranking
• Search and Retrieval-Augmented AI: Personalized, context-aware and across-device search
• Economics, Online Markets and Human Computation: Trust and reliance of crowd workers and data experts on GenAI

Abstract

Human Activity Recognition using wearable inertial sensors is foundational to healthcare monitoring, fitness analytics, and context-aware computing, yet its deployment is hindered by cross-user variability arising from heterogeneous physiological traits, motor habits, and sensor placements. Existing domain generalization approaches either neglect temporal dependencies in sensor streams or depend on impractical target-domain annotations. We propose a different paradigm: modeling generalizable feature extraction as a collaborative sequential generation process governed by reinforcement learning. Our framework, CTFG (Collaborative Temporal Feature Generation), employs a Transformer-based autoregressive generator that incrementally constructs feature token sequences, each conditioned on prior context and the encoded sensor input. The generator is optimized via Group-Relative Policy Optimization, a critic-free algorithm that evaluates each generated sequence against a cohort of alternatives sampled from the same input, deriving advantages through intra-group normalization rather than learned value estimation. This design eliminates the distribution-dependent bias inherent in critic-based methods and provides self-calibrating optimization signals that remain stable across heterogeneous user distributions. A tri-objective reward comprising class discrimination, cross-user invariance, and temporal fidelity jointly shapes the feature space to separate activities, align user distributions, and preserve fine-grained temporal content. Evaluations on the DSADS and PAMAP2 benchmarks demonstrate state-of-the-art cross-user accuracy (88.53\% and 75.22\%), substantial reduction in inter-task training variance, accelerated convergence, and robust generalization under varying action-space dimensionalities.

Problem

• Human Activity Recognition
• Cross-User Variability
• Domain Generalization
• Temporal Dependencies
• Wearable Sensors

Innovation

• critic-free reinforcement learning
• collaborative feature generation
• cross-user generalization
• temporal feature modeling
• Group-Relative Policy Optimization