Scholarlink

<p>Over more than a century following the discovery of superconductors, extensive research has been conducted to leverage their properties, particularly in fundamental materials science applications. Nonetheless, significant challenges still exist that hinder the comprehensive production of superconductors, including the optimal critical temperature (Tc) and current density (Jc), concerns regarding the purity of superconductor phases, sintering temperature, complexities in crystal growth, and various high-cost fabrication-related issues. Recent significant breakthroughs in artificial intelligence (AI) approaches have provided disruptive solutions, such as machine learning (ML), to address these&#xa0;fundamental issues. Thus, ML approaches can be employed to address the issues associated with superconductivity and serve as a means to achieve optimal conditions for superconductors and their applications. ML methodologies can deliver rapid, efficient, and precise solutions for intricate and nonlinear technological, manufacturing, and economic challenges in the domain of superconductivity. This paper initially presents the notion of AI and the often employed ML techniques. A comprehensive conceptual overview is provided for studies employing ML methods aimed at properties&#xa0;enhancement, condition monitoring, and the structural analysis of existing superconductors, along with other pertinent applications. This subject overview is organized into three primary topics: fundamental application utilizing ML, databases used by ML models, and our main focus which is the optimization techniques used alongside with ML in superconductors. Furthermore, the difficulties associated with using ML methodologies in superconductivity and their applications are presented. Ultimately, prospective developments regarding the integration of ML approaches with superconducting for various applications are examined.</p>

错误：搜索内容不能为空，请输入英文关键词
错误：关键词超出字数限制，请精简
高级检索

Data-Driven Superconductivity: a Review of Machine Learning Methods for Material Discovery, Property Prediction, and Process Optimization

  • 期刊论文

  • 发表日期：

    2026年4月24日

查看全文
  • Muhammad Kashfi Shabdin,
  • Muralidhar Miryala,
  • Mohd Mustafa Awang Kechik,
  • Chen Soo Kien,
  • Lim Kean Pah,
  • Mohd Asyadi Azam

摘要

Over more than a century following the discovery of superconductors, extensive research has been conducted to leverage their properties, particularly in fundamental materials science applications. Nonetheless, significant challenges still exist that hinder the comprehensive production of superconductors, including the optimal critical temperature (Tc) and current density (Jc), concerns regarding the purity of superconductor phases, sintering temperature, complexities in crystal growth, and various high-cost fabrication-related issues. Recent significant breakthroughs in artificial intelligence (AI) approaches have provided disruptive solutions, such as machine learning (ML), to address these fundamental issues. Thus, ML approaches can be employed to address the issues associated with superconductivity and serve as a means to achieve optimal conditions for superconductors and their applications. ML methodologies can deliver rapid, efficient, and precise solutions for intricate and nonlinear technological, manufacturing, and economic challenges in the domain of superconductivity. This paper initially presents the notion of AI and the often employed ML techniques. A comprehensive conceptual overview is provided for studies employing ML methods aimed at properties enhancement, condition monitoring, and the structural analysis of existing superconductors, along with other pertinent applications. This subject overview is organized into three primary topics: fundamental application utilizing ML, databases used by ML models, and our main focus which is the optimization techniques used alongside with ML in superconductors. Furthermore, the difficulties associated with using ML methodologies in superconductivity and their applications are presented. Ultimately, prospective developments regarding the integration of ML approaches with superconducting for various applications are examined.