Class aware replay improves hazard diagnosis in industrial AI systems
Class-aware Latent Attention Replay (CLAR): Class Incremental Hazard Diagnosis for Resilient AI in the Manufacturing Industrial Internet
Machine Learning
Summary
In manufacturing systems, AI helps detect problems early to keep things running smoothly. But new types of hazards keep appearing, often with only a few examples, making it hard for AI to learn them without forgetting old hazards. The authors present a method called CLAR that stores a small, smart summary of old hazard data and focuses on important features to better remember past and new hazards. Their method performs better than others on a real industrial test dataset.
What this means in practice
- •For industrial ai engineers: Deploy incremental AI hazard diagnosis models that efficiently learn new hazard classes with limited memory on factory-edge devices.
- •For manufacturing operations teams: Improve real-time detection and diagnosis of emerging equipment or process hazards across multiple data sources by using AI models resilient to new fault types.
Authors
Avranshu Mahanty, Ran Jin, Yingyan Zeng
Abstract
Artificial Intelligence (AI) systems provide critical computation services in the Manufacturing Industrial Internet (MII), and their resilience depends on timely diagnosis of hazards across the data, AI pipeline, and cyber-physical layers. During operation, new hazard types keep emerging with only a few samples each, hazards co-occur, and the diagnosis model runs on memory-constrained computation nodes. Replay-based class-incremental learning (CIL) can learn new hazard classes while revisiting a small memory of past samples, but existing methods assume single-label, balanced streams, store high-dimensional raw samples, and select them independently of the classifier. We propose Class-aware Latent Attention Replay (CLAR), which combines a dual buffer that stores mostly compact latent representations with a few raw samples, an Attention Leverage Score (ALS) that upper-bounds the classifier logit and guides class-balanced reservoir sampling, and an exponential-moving-average teacher that slows latent-space drift. For an idealized buffer model, we prove that the buffer is class-balanced and follows an ALS-weighted class distribution, and that below a critical per-class capacity, ALS-guided replay has a lower replay-gradient error than class-balanced uniform replay. In the case study, On hazard diagnosis data from an MII testbed, CLAR achieves the highest mean average weighted F1 score among the evaluated replay methods with the same buffer size