Reproduce “Hybrid Quantum-Classical AI for Industrial Defect Classification in Welding Images” on OpenResearch
Turn your favorite coding agents into research agents that review literature, develop hypotheses, and run experiments.
Universal app for Apple silicon and Intel. Requires macOS 11 or later.
About this paper
Akshaya Srinivasan, Xiaoyin Cheng, Jianming Yi, Alexander Geng, Desislava Ivanova, Andreas Weinmann, Ali Moghiseh
Hybrid quantum-classical machine learning models were developed to classify defects in aluminum Tungsten Inert Gas (TIG) welding images, demonstrating competitive performance against classical deep learning. A Variational Quantum Circuit (VQC)-based classifier achieved 99.7% accuracy for binary and 98.9% for multiclass defect classification on an industrial dataset.
Key insights
- Utilizing classical CNNs for feature extraction effectively addresses the limitations of NISQ quantum hardware, enabling quantum models to process complex industrial image data.
- The VQC-based classifier demonstrated a strong balance of accuracy and efficiency, proving more suitable for near-term industrial deployment compared to the VQLS-enhanced QSVM.
- Hyperparameter tuning, including feature size and qubit count, is crucial for optimizing the performance and resource efficiency of hybrid quantum models in practical applications.
Summary by alphaXiv.
Tip: change arxiv.org to autoarxiv.org in any arXiv abstract link to open the paper here.