Application of Unsupervised Learning to AC Susceptibility Measurements of High-Temperature Superconductors

Explore the application of unsupervised learning techniques to AC susceptibility measurements of High-Temperature Superconductors in this 33-minute conference talk from the HyperComplex Seminar 2023. Delve into the innovative use of machine learning algorithms, specifically unsupervised learning, to analyze large datasets of superconductor measurements. Learn how Convolutional 1D Autoencoders and the Bag of Words model can represent complex AC measurements using just five numerical values. Discover the potential of this approach to reveal relationships between different types of High-Temperature Superconductors and their properties. Examine the results of cluster analysis and t-SNE visualization techniques applied to the dataset, and consider the implications for future research in this field. Gain insights into the challenges and opportunities of applying cutting-edge data analysis methods to advance our understanding of superconductor physics.