Research3h ago
Wi-Fi Signal Research Adds Causal Interpretability
A new pipeline uses a categorical variational autoencoder to compress high-dimensional Wi-Fi signals into discrete representations. This method enables Linear Temporal Logic rule extraction for human activity recognition. It bridges the gap between raw signal processing and symbolic controllability. Practitioners gain a way to audit and modify how AI interprets physical movements via Wi-Fi.