Development of crystal structure prediction methods using machine learning

• Project/Area Number: 18K13474
• Research Category: Grant-in-Aid for Early-Career Scientists
• Allocation Type: Multi-year Fund
• Review Section: Basic Section 13010:Mathematical physics and fundamental theory of condensed matter physics-related
• Research Institution: Nagaoka University of Technology (2019); National Institute for Materials Science (2018)
• Principal Investigator: Yamashita Tomoki 長岡技術科学大学, 産学融合トップランナー養成センター, 特任准教授 (60793099)
• Project Period (FY): 2018-04-01 – 2020-03-31
• Project Status: Completed (Fiscal Year 2019)
• Budget Amount: ¥4,160,000 (Direct Cost: ¥3,200,000、Indirect Cost: ¥960,000); Fiscal Year 2019: ¥1,300,000 (Direct Cost: ¥1,000,000、Indirect Cost: ¥300,000); Fiscal Year 2018: ¥2,860,000 (Direct Cost: ¥2,200,000、Indirect Cost: ¥660,000)
• Keywords: 結晶構造探索 / 機械学習 / 第一原理計算 / ベイズ最適化 / 進化的アルゴリズム / ランダムサーチ / マテリアルズインフォマティクス

Outline of Final Research Achievements

We have developed crystal structure prediction methods. The code development of Bayesian optimization and evolutionary algorithm has done and implemented in the open source software, CrySPY(https://github.com/Tomoki-YAMASHITA/CrySPY). CrySPY is written in Python to be used widely. Moreover the document and web site are published.

Academic Significance and Societal Importance of the Research Achievements

本研究により、オープンソースの結晶構造探索ツールであるCrySPYを公開することができた。CrySPYは機械学習を用いた高効率な安定構造探索が可能であり、新材料設計の基盤となるツールとして誰もが利用可能である。半年に一度の頻度でCrySPYのチュートリアルセミナーも開催しており、大学に所属する研究者や学生および企業の研究者などの間で利用されるようになった。

Research Products

• [Journal Article] Adjusting the descriptor for a crystal structure search using Bayesian optimization (2020)
  • Author(s): Sato Nobuya、Yamashita Tomoki、Oguchi Tamio、Hukushima Koji、Miyake Takashi
  • Journal Title: Physical Review Materials Volume: 4 Issue: 3 Pages: 033801-033801
  • DOI: 10.1103/physrevmaterials.4.033801
  • Peer Reviewed
• [Presentation] Hybrid Algorithm of Bayesian Optimization and Evolutionary Algorithm in Crystal Structure Prediction (2019)
  • Author(s): Tomoki Yamashita, Shinichi Kanehira, Nobuya Sato, Hiori Kino, Koji Tsuda, Takashi Miyake, and Tamio Oguchi
  • Organizer: MATERIALS RESEARCH MEETING 2019
  • Int'l Joint Research
• [Presentation] Descriptor for Efficient Crystal Prediction Using the Bayesian Optimization (2019)
  • Author(s): Nobuya Sato, Tomoki Yamashita, Tamio Oguchi, Koji Hukushima, and Takashi Miyake
  • Organizer: MATERIALS RESEARCH MEETING 2019
  • Int'l Joint Research
• [Presentation] Parameter in a Descriptor for Efficient Crystal Structure Search Using the Bayesian Optimization (2019)
  • Author(s): Nobuya Sato, Tomoki Yamashita, Tamio Oguchi, Koji Hukushima, and Takashi Miyake
  • Organizer: The 22nd Asian Workshop on First-Principles Electronic Structure Calculations
  • Int'l Joint Research
• [Presentation] 結晶構造探索におけるベイズ最適化と進化的アルゴリズムのハイブリッドアルゴリズム (2019)
  • Author(s): 山下智樹，兼平慎一，佐藤暢哉，木野日織，津田宏治，三宅隆，小口多美夫
  • Organizer: 日本物理学会2019年秋季大会
• [Presentation] Searching Efficiency of Bayesian Optimization and Evolutionary Algorithm in Crystal Structure Prediction (2019)
  • Author(s): Tomoki Yamashita, Shinichi Kanehira, Nobuya Sato, Hiori Kino, Koji Tsuda, Takashi Miyake, and Tamio Oguchi
  • Organizer: 10th International Conference on Materials for Advanced Technologies
  • Int'l Joint Research
• [Presentation] Crystal Structure Prediction by Bayesian Optimization and Evolutionary Algorithm (2019)
  • Author(s): Tomoki Yamashita, Shinichi Kanehira, N. Sato, Hiori Kino, Koji Tsuda, Takashi Miyake, and Tamio Oguchi
  • Organizer: APS March Meeting 2019
  • Int'l Joint Research
• [Presentation] Bayes最適化を用いた結晶構造探索の効率と記述子のパラメータ (2019)
  • Author(s): 佐藤暢哉，山下智樹，小口多美夫，福島孝治，三宅隆
  • Organizer: 日本物理学会第74回年次大会
• [Presentation] Development of crystal structure prediction tool (2019)
  • Author(s): Tomoki Yamashita, Kei Terayama, Shinichi Kanehira, N. Sato, Hiori Kino, Koji Tsuda, Takashi Miyake, and Tamio Oguchi
  • Organizer: PRESTO International Symposium on Materials Informatics
  • Int'l Joint Research
• [Presentation] Crystal structure prediction by machine learning (2018)
  • Author(s): Tomoki Yamashita
  • Organizer: CSRN-OSAKA Annual Workshop 2018
  • Int'l Joint Research / Invited
• [Presentation] Bayes最適化を用いた結晶構造探索における記述子 (2018)
  • Author(s): 佐藤暢哉，山下智樹，小口多美夫，福島孝治，三宅隆
  • Organizer: 日本物理学会2018年秋季大会
• [Remarks] CrySPY
  • URL: https://tomoki-yamashita.github.io/CrySPY/
• [Remarks] CrySPY
  • URL: https://github.com/Tomoki-YAMASHITA/CrySPY
• [Remarks] チュートリアルと解説
  • URL: https://tomoki-yamashita.github.io/cryspy/tutorial/outline.html