Predicting new superconductors and their critical temperatures using unsupervised machine learning

Abstract

We used the superconductors in the SuperCon database to construct element vectors and then perform unsupervised learning of their critical temperatures (T${}_c$). Only the chemical composition of superconductors was used in this procedure. No physical predictors (neither experimental nor computational) of any kind were used. We achieved the coefficient of determination R${}^2$$\simeq$0.93, which is comparable and in some cases higher then similar estimates using other artificial intelligence techniques. Based on this machine learning model, we predicted several new superconductors with high critical temperatures. We also discuss the factors that limit the learning process and suggest possible ways to overcome them.

References in corpus (11)

• Supplementary information for "Quantum supremacy using a programmable superconducting processor"
• Superconductivity at 250 K in lanthanum hydride under high pressures
• The doping dependence of T* - what is the real high-Tc phase diagram?
• Machine learning modeling of superconducting critical temperature
• Materials Cartography: Representing and Mining Material Space Using Structural and Electronic Fingerprints
• Universal scaling relation in high-temperature superconductors
• Atom2Vec: learning atoms for materials discovery
• Materials informatics based on evolutionary algorithms: Application to search for superconducting hydrogen compounds
• Recent advances in high-throughput superconductivity research
• Dynamic Superconductivity Responses in Photo-excited Optical Conductivity and Nernst Effect
• Universal Correlation between Critical Temperature of Superconductivity and band structure features

Cited by in corpus (16)

• Inverse Design of Next-generation Superconductors Using Data-driven Deep Generative Models
• Machine-Learning-Guided Prediction Models of Critical Temperature of Cuprates
• Predicting Superconducting Transition Temperature through Advanced Machine Learning and Innovative Feature Engineering
• InvDesFlow: An AI-driven materials inverse design workflow to explore possible high-temperature superconductors
• Data-driven Design of High Pressure Hydride Superconductors using DFT and Deep Learning
• Accelerating superconductor discovery through tempered deep learning of the electron-phonon spectral function
• Pressure-Induced Superconductivity in Layered Transition-metal Chalcogenides (Zr,Hf)GeTe${}_4$ Explored by Data-driven Approach
• Classification-based detection and quantification of cross-domain data bias in materials discovery
• Leveraging Composition-Based Material Descriptors for Machine Learning Optimization
• Semi-automatic staging area for high-quality structured data extraction from scientific literature
• Superconductor discovery in the emerging paradigm of Materials Informatics
• Data analysis of $ab$ $initio$ effective Hamiltonians in iron-based superconductors $\unicode{x2014}$ Construction of predictors for superconducting critical temperature
• Machine-learning Guided Search for Phonon-mediated Superconductivity in Boron and Carbon Compounds
• Developing a Complete AI-Accelerated Workflow for Superconductor Discovery
• Tree Models Machine Learning to Identify Liquid Metal based Alloy Superconductor
• Effects of Li Doping on Superconducting Properties of Citrate Gel Prepared Y1-xLixBa2Cu3O7-d Compound