Towards Causally Interpretable Wi-Fi CSI-Based Human Activity Recognition with Discrete Latent Compression and LTL Rule Extraction

Researchers from multiple institutions, including Luca Cotti, Marco Cominelli, and Mani B. Srivastava, have introduced a novel pipeline called CHARL-TRE (CHAR Latent Temporal Rule Extraction) for Human Activity Recognition (HAR) using Wi-Fi Channel State Information (CSI). The work, accepted at FUSION 2026 and published on arXiv, addresses the challenge of making deep neural models for CSI-based HAR causally interpretable and symbolically controllable. Current deep models achieve strong predictive performance but rely on opaque continuous latent representations, while purely symbolic approaches cannot process raw CSI streams directly.

The CHARL-TRE pipeline operates in a fully automatic and strictly decoupled manner. First, CSI magnitude windows are compressed by a categorical variational autoencoder with Gumbel-Softmax latent variables under a capacity-controlled objective, yielding a compact discrete representation. The encoder is then frozen and used as a deterministic mapping to one-hot latent trajectories. Causal discovery is performed on these trajectories to estimate class-conditional temporal dependency graphs. Statistically supported lagged dependencies are translated into Linear Temporal Logic (LTL) rules, producing a fully symbolic and deterministic classifier based solely on rule evaluation and aggregation, without any learned discriminative head. Because rules are defined over discrete latent variables, antenna-specific rule sets can be combined at the symbolic level, enabling structured multi-antenna fusion without retraining the encoder. Results indicate competitive performance while preserving explicit temporal and causal structure.