New Neuro-Symbolic Framework Enhances Explainable Biosignal Anomaly Detection

2026-09-25

A novel neuro-symbolic framework, Signal2Symbol, has been proposed to address the need for transparent decision-making in physiological time-series anomaly detection. It converts biosignals into symbolic sequences, enabling more interpretable explanations for detected anomalies.

VERA Brief

AI-generated. Grounded in the article and its cited sources.

A new neuro-symbolic framework called Signal2Symbol has been developed for explainable anomaly detection in biosignals. It converts biosignals into symbolic sequences to provide interpretable explanations for detected anomalies.

Key facts

Signal2Symbol is a neuro-symbolic framework for explainable anomaly detection in physiological time-series data.
The framework converts biosignals into symbolic sequences using a learned VQ-VAE codebook or SAX baseline.
Anomalies are scored using evidence from rare itemsets and merged into intervals for composite temporal explanations.
A rare temporal concept lattice groups anomalous intervals based on shared symbolic evidence and temporal relations.

Source: arXiv · cs.LG

Reported by VERA Newswire.