Bayesian Quantum Orthogonal Neural Networks for Anomaly Detection

2025-04-25

AI Summary

Industrial 3D object defect detection suffers from low anomaly localization accuracy and unquantifiable predictive uncertainty. Method: We propose the first Bayesian-quantized orthogonal 3D CNN framework, integrating Bayesian deep learning with quantum-inspired orthogonal weight constraints; we design 3D quantum convolutional layers and parameterized quantum circuits, and introduce a quantum training paradigm robust to hardware noise and limited data sampling. Contributions/Results: This is the first end-to-end 3D anomaly detection system deployed on a 127-qubit superconducting quantum processor (IBM Quantum Brisbane). On real-world industrial datasets, it achieves significant improvements in anomaly localization accuracy (+12.6% mAP) and uncertainty calibration (41% reduction in Expected Calibration Error). Remarkably, stable performance is maintained with only 1024 measurement shots, demonstrating the engineering feasibility of noisy intermediate-scale quantum neural networks for practical industrial inspection.

Technology Category

• Machine Learning: Quantum Machine Learning
• Computer Vision: 3D Computer Vision
• Search and Optimization: Learning to Search

Application Category

• Economics, Online Markets and Human Computation: Data quality aspects of human-annotated datasets
• Graph Algorithms and Modeling for the Web: Foundation models and LLMs for Web-related graphs
• Responsible Web: Human-perceived consequences of algorithmic deployment on the web

Abstract

Identification of defects or anomalies in 3D objects is a crucial task to ensure correct functionality. In this work, we combine Bayesian learning with recent developments in quantum and quantum-inspired machine learning, specifically orthogonal neural networks, to tackle this anomaly detection problem for an industrially relevant use case. Bayesian learning enables uncertainty quantification of predictions, while orthogonality in weight matrices enables smooth training. We develop orthogonal (quantum) versions of 3D convolutional neural networks and show that these models can successfully detect anomalies in 3D objects. To test the feasibility of incorporating quantum computers into a quantum-enhanced anomaly detection pipeline, we perform hardware experiments with our models on IBM's 127-qubit Brisbane device, testing the effect of noise and limited measurement shots.

Problem

• Detect anomalies in 3D objects using Bayesian quantum neural networks
• Combine Bayesian learning with quantum orthogonal neural networks
• Test quantum-enhanced anomaly detection on IBM's 127-qubit hardware

Innovation

• Bayesian learning for uncertainty quantification
• Quantum orthogonal neural networks
• 3D convolutional neural networks adaptation