ExplainabilityTransparencySafetycs.LG
CLAIRE: Compressed Latent Autoencoder for Industrial Representation and Evaluation -- A Deep Learning Framework for Smart Manufacturing
Mohammadhossein Ghahramani, Mengchu Zhou
March 6, 2026
IEEE Transactions on Systems, Man & Cybernetics. Systems
2 citations
Key Takeaways
-
CLAIRE improves fault detection by simplifying complex sensor data: The framework uses deep learning to compress high-dimensional industrial sensor data into a simpler "latent space," which helps identify equipment faults more accurately than traditional methods that work directly with raw data.
-
The system explains its own decisions using game-theory tools: After making predictions, CLAIRE uses a post hoc interpretability technique to identify which specific input features were most important for detecting faults, addressing a key concern about "black box" AI systems in manufacturing.
-
Explainability and robust design go hand-in-hand: The paper demonstrates that incorporating interpretability techniques alongside feature-aware regularization creates more reliable fault detection, suggesting that transparency and performance are complementary rather than competing goals in responsible AI.
-
The modular design enables responsible deployment across industries: Because CLAIRE's architecture separates data compression from classification and adds an interpretability layer, the authors note it can be adapted to other high-stakes domains like healthcare and finance—suggesting that responsible AI frameworks designed for one industry can transfer to others with similar data challenges.
Abstract
Accurate fault detection in high-dimensional industrial environments remains a major challenge due to the inherent complexity, noise, and redundancy in sensor data. This paper introduces CLAIRE, i.e., a hybrid end-to-end learning framework that integrates unsupervised deep representation learning with supervised classification for intelligent quality control in smart manufacturing systems. It employs an optimized deep autoencoder to transform raw input into a compact latent space, effectively capturing the intrinsic data structure while suppressing irrelevant or noisy features. The learned representations are then fed into a downstream classifier to perform binary fault prediction. Experimental results on a high-dimensional dataset demonstrate that CLAIRE significantly outperforms conventional classifiers trained directly on raw features. Moreover, the framework incorporates a post hoc phase, using a game-theory-based interpretability technique, to analyze the latent space and identify the most informative input features contributing to fault predictions. The proposed framework highlights the potential of integrating explainable AI with feature-aware regularization for robust fault detection. The modular and interpretable nature of the proposed framework makes it highly adaptable, offering promising applications in other domains characterized by complex, high-dimensional data, such as healthcare, finance, and environmental monitoring.