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

2019 Fiscal Year Final Research Report

High-performance computing of statistical physics of spin systems using GPU

Research Project

  • PDF
Project/Area Number 16K05480
Research Category

Grant-in-Aid for Scientific Research (C)

Allocation TypeMulti-year Fund
Section一般
Research Field Mathematical physics/Fundamental condensed matter physics
Research InstitutionTokyo Metropolitan University

Principal Investigator

Okabe Yutaka  首都大学東京, 理学研究科, 客員教授 (60125515)

Project Period (FY) 2016-04-01 – 2020-03-31
Keywordsモンテカルロ法 / GPU / フラストレート系 / 相転移 / BKT転移 / イジングモデル / クロックモデル / 機械学習
Outline of Final Research Achievements

High-performance Monte Carlo studies were performed. First, I studied the phase transition of spin systems on the pyrochlore lattice. I observed multiple plateaus in the magnetization curve in the diluted spin-ice materials. I also found large peaks of residual entropy in the diluted spin-ice model. Second, I studied q-state clock models of regular and Villain types with q = 5, 6 and observed double transitions in each model. I calculated the correlation ratio and size-dependent correlation length as quantities for characterizing the existence of Berezinskii-Kosterlitz-Thouless (BKT) phase. Third, with the recent developments in machine learning, I proposed a new approach for the study of spin models, which is complementary to the conventional approach investigating the thermal average of macroscopic physical quantities. Paying attention to correlation configurations to use machine learning, I succeeded in the phase classification for both the second-order and the BKT transitions.

Free Research Field

数物系科学

Academic Significance and Societal Importance of the Research Achievements

効率のよいアルゴリズムを用いて、複雑な物理系のシミュレーションを実行する研究で、研究成果と共に、方法論としても意味がある。特に、機械学習の進展に注目して、新規に始めた機械学習の統計力学への応用は、BKT転移を示す6状態クロックモデルの訓練データを用いて、2次相転移を示す4状態クロックモデルのテストデータによる相分類を行い、2次相転移の臨界点とBKT転移の臨界線との関連を明らかにするなど、ユニークな研究結果も得た。スピン系の研究の新しいパラダイムを提示するもので、その方法は、一般的で多方面への応用範囲がある。特に量子系への応用は、量子情報・量子計算の研究への展開も視野に入れると、興味深い。

URL: 

Published: 2021-02-19  

Information User Guide FAQ News Terms of Use Attribution of KAKENHI

Powered by NII kakenhi