Fast Re-Trainable Attention Autoencoder for Liquid Sensor Anomaly Detection at the Edge
This work addresses the need for quick adaptation in liquid sensor anomaly detection at the edge, though it is incremental as it applies an existing method to a new domain-specific application.
The authors tackled the problem of detecting sensor anomalies in chemistry and biology laboratories by proposing a lightweight, edge-deployable pipeline using an Attention-based One-Class Autoencoder, achieving an F1 score of 0.72, precision of 0.89, and recall of 0.61 with only thirty minutes of training data.
A lightweight, edge-deployable pipeline is proposed for detecting sensor anomalies in chemistry and biology laboratories. A custom PCB captures seven sensor channels and streams them over the local network. An Attention-based One-Class Autoencoder reaches a usable state after training on only thirty minutes of normal data. Despite the small data set, the model already attains an F1 score of 0.72, a precision of 0.89, and a recall of 0.61 when tested on synthetic micro-anomalies. The trained network is converted into a TensorFlow-Lite binary of about 31 kB and runs on an Advantech ARK-1221L, a fan-less x86 edge device without AVX instructions; end-to-end inference latency stays below two seconds. The entire collect-train-deploy workflow finishes within one hour, which demonstrates that the pipeline adapts quickly whenever a new liquid or sensor is introduced.