CMR-Mamba uses causal mechanism monitoring for industrial fault detection

CMR-Mamba employs Mamba state-space encoders and a causal cross-modal predictor to detect coupling faults in industrial systems. This approach identifies anomalies by measuring the distance between observed and causally predicted effect embeddings on a manifold, outperforming marginal distribution monitoring.