Prototypical networks improve sensor fault diagnosis with few examples

Robust Prototypical Networks for Few-Shot Sensor Fault Diagnosis

Artificial IntelligenceMachine Learning

Summary

Detecting faults in industrial sensors is hard when there are only a few examples of each fault type to learn from. The authors propose a way to make prototypical networks more reliable by combining information from multiple small groups of examples instead of just one. This method stabilizes the network’s understanding of each fault, making it better at identifying faults even with very limited data. They tested their approach on real industrial sensor data with various simulated problems and found it outperforms standard methods when examples are scarce.

What this means in practice

Authors

Mohammed Ayalew Belay, Amirshayan Haghipour, Pierluigi Salvo Rossi

Abstract

Industrial fault diagnosis often operates with only a handful of labeled fault examples, making few-shot learning attractive for sensor monitoring. Standard prototypical networks are simple and effective; however, their class prototypes may become unstable in the very-low-shot regime because each decision relies on a small support set. We propose \emph{Multi-Episode Prototypical Networks} (MEPN), which aggregate prototypes from multiple disjoint support episodes and use their mean as the final class representative, reducing prototype variance without changing the encoder architecture. We evaluate MEPN on the DeFACTO sensor dataset using five-way fault classification with synthetic bias, drift, spike, and noise faults injected into real industrial measurements. Over 100 independent runs, MEPN reaches \textbf{\SensorOneShotGcpn\%} in the per-episode one-shot setting ($K\!=\!1$ shot, aggregated over $N_{\text{agg}}\!=\!10$ support episodes), substantially above single-episode baselines. Under an equal 10-sample support budget, MEPN and ProtoNet at $K\!=\!10$ are statistically indistinguishable, confirming prototype accumulation as the mechanism rather than superior fixed-budget learning.