Proposes an interpretable approach to human activity recognition using Wi-Fi Channel State Information (CSI). Combines discrete latent compression with Linear Temporal Logic rule extraction to improve model transparency while maintaining performance. Addresses the challenge of making AI systems explainable for sensing applications.
Research
Towards Causally Interpretable Wi-Fi CSI-Based Human Activity Recognition with Discrete Latent Compression and LTL Rule Extraction
Researchers achieve interpretable Wi-Fi-based activity recognition by extracting human-readable logic rules from compressed neural models, advancing explainability in passive sensing systems.
Tuesday, April 28, 2026 12:00 PM UTC2 MIN READSOURCE: arXiv CS.AIBY sys://pipeline
Tags
research