Fast Re-Trainable Attention Autoencoder for Liquid Sensor Anomaly Detection at the Edge

TL;DR

This paper presents a fast, lightweight autoencoder-based system for detecting anomalies in liquid sensors at the edge, capable of quick retraining with minimal data and suitable for real-time applications.

Contribution

The authors introduce a re-trainable attention-based autoencoder pipeline that achieves rapid deployment and high accuracy on small datasets for liquid sensor anomaly detection.

Findings

F1 score of 0.72 on synthetic anomalies

Inference latency below two seconds on edge device

Training and deployment completed within one hour

Abstract

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.