• Search Research Projects
  • Search Researchers
  • How to Use
  1. Back to previous page

Transfer learning to predict the properties of compounds: pre-trained model library

Research Project

Project/Area Number 20K19866
Research Category

Grant-in-Aid for Early-Career Scientists

Allocation TypeMulti-year Fund
Review Section Basic Section 61030:Intelligent informatics-related
Research InstitutionThe Institute of Statistical Mathematics

Principal Investigator

Liu Chang  統計数理研究所, ものづくりデータ科学研究センター, 特任助教 (30814149)

Project Period (FY) 2020-04-01 – 2023-03-31
Project Status Completed (Fiscal Year 2022)
Budget Amount *help
¥4,160,000 (Direct Cost: ¥3,200,000、Indirect Cost: ¥960,000)
Fiscal Year 2022: ¥780,000 (Direct Cost: ¥600,000、Indirect Cost: ¥180,000)
Fiscal Year 2021: ¥780,000 (Direct Cost: ¥600,000、Indirect Cost: ¥180,000)
Fiscal Year 2020: ¥2,600,000 (Direct Cost: ¥2,000,000、Indirect Cost: ¥600,000)
Keywords転移学習 / スモールデータ / 物性予測 / 結晶構造予測 / マテリアルズインフォマティクス / 新規準結晶探索 / モデルライブラリー / ハイエントロピー合金 / 準結晶 / データベース / 記述子データベース
Outline of Research at the Start

機械学習のモデルは,既存のデータとの類似性から未知物質の特性を予測するため,一般には,学習データが存在しない真に革新的な物質の特性が予測できない.しかし,材料研究の究極の目標は,外挿的予測と発見の実現である.
転移学習は、あるタスクで学習されたモデルを他のタスクに流用するための解析技術である.人間で例えにすれば,英語が上手い人(英語を学習した)は英語に近いドイツ語も勉強しやすことと同じである.本研究は転移学習の特徴を利用し,物性の関連性を転移させることによって,外挿的予測を実現する研究である.

Outline of Final Research Achievements

In this study, we have successfully employed transfer learning, a powerful machine learning technique, to address the challenge of limited material data and enable predictive extrapolation in the field of materials informatics (MI).

Our research has yielded significant outcomes. Firstly, we have developed XenonPy.MDL, an extensive model library containing a multitude of trained models. This library serves as a valuable resource for further advancements in the field. Secondly, we have applied transfer learning on thermodynamic stability prediction of high-entropy alloys and lattice thermal conductivity prediction, leading to notable findings presented at prestigious international conferences. Furthermore, our work has extended into crystal structure prediction, where we have introduced prediction algorithms surpassing the performance of conventional methods. We have achieved remarkable results by proposing an innovative prediction algorithm, particularly in crystal structure prediction.

Academic Significance and Societal Importance of the Research Achievements

学術的意義としては,転移学習の技術はデータ収集に高いコストがかかる材料研究分野において必要不可欠である.本研究で挙げられたハイエントロピー合金の熱力学安定性予測と結晶構造予測の研究成果はその実例であり,小規模な第一原理計算の結果のみで高精度な予測を実現した.転移学習技術の導入は.研究の効率化と新たな技術の実現に向けた重要な一歩となった.
社会的意義としては,転移学習の導入により,材料の設計や特性予測の精度と外挿性能が向上し,材料開発のスピードが加速されることが期待できる.これにより,エネルギー効率の高い材料や環境負荷の低い製品の開発が促進され,持続可能な社会の実現に寄与することができる.

Report

(4 results)
  • 2022 Annual Research Report   Final Research Report ( PDF )
  • 2021 Research-status Report
  • 2020 Research-status Report
  • Research Products

    (10 results)

All 2022 2021 Other

All Journal Article (3 results) (of which Int'l Joint Research: 2 results,  Peer Reviewed: 3 results,  Open Access: 2 results) Presentation (1 results) (of which Int'l Joint Research: 1 results) Book (1 results) Remarks (5 results)

  • [Journal Article] Descriptors of intrinsic hydrodynamic thermal transport: screening a phonon database in a machine learning approach2022

    • Author(s)
      Torres Pol、Wu Stephen、Ju Shenghong、Liu Chang、Tadano Terumasa、Yoshida Ryo、Shiomi Junichiro
    • Journal Title

      Journal of Physics: Condensed Matter

      Volume: 34 Issue: 13 Pages: 135702-135702

    • DOI

      10.1088/1361-648x/ac49c9

    • Related Report
      2022 Annual Research Report 2021 Research-status Report
    • Peer Reviewed / Open Access / Int'l Joint Research
  • [Journal Article] Machine learning to predict quasicrystals from chemical compositions2021

    • Author(s)
      Liu, C., Fujita, E., Katsura, Y., Inada, Y., Ishikawa, A., Tamura, R., Kimura, K., Yoshida, R.
    • Journal Title

      Advanced Materials

      Volume: 33 Issue: 36 Pages: 2102507-2102507

    • DOI

      10.1002/adma.202102507

    • Related Report
      2021 Research-status Report
    • Peer Reviewed / Open Access
  • [Journal Article] Exploring diamondlike lattice thermal conductivity crystals via feature-based transfer learning2021

    • Author(s)
      Ju Shenghong、Yoshida Ryo、Liu Chang、Wu Stephen、Hongo Kenta、Tadano Terumasa、Shiomi Junichiro
    • Journal Title

      Physical Review Materials

      Volume: 5 Issue: 5 Pages: 053801-053801

    • DOI

      10.1103/physrevmaterials.5.053801

    • Related Report
      2020 Research-status Report
    • Peer Reviewed / Int'l Joint Research
  • [Presentation] First-principles study of Quaternary High Entropy Alloys consisting of Fe-Ni-Co-Cr-Mn/Pd2022

    • Author(s)
      Nguyen-Dung Tran
    • Organizer
      2022 TMS Annual Meeting & Exhibition
    • Related Report
      2021 Research-status Report
    • Int'l Joint Research
  • [Book] マテリアルズインフォマティクス2022

    • Author(s)
      伊藤 聡、吉田 亮、劉 暢、Stephen Wu、野口 瑶、山田 寛尚、赤木 和人、大林 一平、山下 智樹
    • Total Pages
      202
    • Publisher
      共立出版
    • ISBN
      9784320072022
    • Related Report
      2022 Annual Research Report
  • [Remarks] 結晶構造予測の論文

    • URL

      https://doi.org/10.48550/arXiv.2305.02158

    • Related Report
      2022 Annual Research Report
  • [Remarks] XenonPyプロジェクトHP

    • URL

      https://xenonpy.readthedocs.io/en/latest/

    • Related Report
      2022 Annual Research Report
  • [Remarks] XenonPyプロジェクトソースコード

    • URL

      https://github.com/yoshida-lab/XenonPy

    • Related Report
      2022 Annual Research Report 2020 Research-status Report
  • [Remarks] XenonPyプロジェクトHP

    • URL

      http://xenonpy.readthedocs.io/

    • Related Report
      2020 Research-status Report
  • [Remarks] XenonPy.MDL検索用ページ(開発中)

    • Related Report
      2020 Research-status Report

URL: 

Published: 2020-04-28   Modified: 2024-01-30  

Information User Guide FAQ News Terms of Use Attribution of KAKENHI

Powered by NII kakenhi