ホーム > 広報活動 > イベント・セミナー > 2015年 > 情報統合型物質・材料開発イニシアティブ (MI2I) 主催の第7回情報統合型研究交流会・一般公開シリーズ (4) (旧 第4回MI2Iセミナー) を開催

情報統合型物質・材料開発イニシアティブ (MI2I) 主催の第7回情報統合型研究交流会・一般公開シリーズ (4) (旧 第4回MI2Iセミナー) を開催

開催日: 2015.11.13 終了


・日 時      : 2015年11月13日(金) 10時00分~12時00分
 Date and Time  10:00-12:00/Fri. 13th, Nov., 2015

・場 所      : 千現地区 8階 中セミナー室
 Venue       Sengen site, Main Bldg. 8F Seminar Room (811・812)

・講演者      : Dr. Matthias Rupp (Fritz-Haber-Institute, Germany)
 Speaker     Dr. Kristof Schutt (Technical University of Berlin, Germany)
           ※PH.D Student of Klaus Muller's Lab.

<Dr. Mattius Rupp>
Title :
Machine Learning Models for Materials Science at the Atomic Scale

Abstract :
Systematic computational study and design of materials requires rigorous, unbiased, and accurate treatment at the atomic scale. Existing numerical approximations to the many-electron problem have prohibitive computational cost, severely limiting their applicability in practice.
Based on the reasoning that electronic structure calculations of similar materials contain redundant information, machine learning models have been developed that interpolate between a computationally feasible number of ab initio reference calculations to predict properties of new similar materials. [1] This Ansatz of mapping the problem of solving the electronic Schrödinger equation onto a non-linear statistical regression problem has led to computational savings of up to several orders of magnitude for molecules, with accuracy on par with the reference method. [2] I will provide a conceptual introduction to our approach, [3] discuss selected studies showcasing its broad applicability, and present some of our recent results. [4]

[1] M Rupp, A Tkatchenko, K-R Mueller, OA von Lilienfeld: Fast and Accurate Modeling of Molecular Atomization Energies with Machine Learning, Physical Review Letters 108(5): 058301, 2012. DOI 10.1103/PhysRevLett.108.058301
[2] R Ramakrishnan, PO Dral, M Rupp, OA von Lilienfeld: Big Data Meets Quantum Chemistry Approximations: The Delta-Machine Learning Approach, Journal of Chemical Theory and Computation 11(5): 2087, 2015. DOI 10.1021/acs.jctc.5b00099
[3] M Rupp: Machine Learning for Quantum Mechanics in a Nutshell, International Journal of Quantum Chemistry 115(16): 1058, 2015. DOI 10.1002/qua.24954
[4] M Rupp, R Ramakrishnan, OA von Lilienfeld: Machine Learning for Quantum Mechanical Properties of Atoms in Molecules, Journal of Physical Chemistry Letters 6(16): 3309, 2015. DOI 10.1021/acs.jpclett.5b01456

Website: http://www.mrupp.info/

<Dr. Kristof Schütt>
Title :
Representing atoms for machine learning predictions

Abstract :
While Quantum Chemistry calculations are connected with a trade-off between accuracy and computational cost, machine learning is able to predict atomistic properties, given a suitable set of reference calculations. In order to achieve efficient and accurate predictions, the features used to represent a molecule or solid are critical. In my talk, I will give an overview about the progress of descriptors approaching chemical accuracy and introduce a method to evaluate features and data sets.

Klaus Mueller's Lab's website:
https://www.ml.tu-berlin.de/menue/members/klaus-robert_mueller/

第7回情報統合型研究交流会・一般公開シリーズ (4) (旧 第4回MI2Iセミナー)

日時/Date and Time

2015年11月13日 (金) 10:00 – 12:00
10:00-12:00/Fri. 13th, Nov., 2015

場所/Venue

物質・材料研究機構 千現地区 8階 中セミナー室
National Institute for Materials Science
Sengen site, Main Bldg. 8F Seminar Room (811・812)

講演者/Speaker

Dr. Mattius Rupp(Fritz-Haber-Institute, Germany)
Dr. Kristof Schutt(Technical University of Berlin, Germany)
※PH.D Student of Klaus Muller's Lab.

参加方法及びお問合わせ/Contact

以下までご連絡ください。
 
国立研究開発法人物質・材料研究機構
情報統合型物質・材料研究拠点 運営統括室
National Institute for Materials Science
Center for Materials research by Information Integration Secretariat


イベント・セミナーデータ

イベント・セミナー名
情報統合型物質・材料開発イニシアティブ (MI2I) 主催の第7回情報統合型研究交流会・一般公開シリーズ (4) (旧 第4回MI2Iセミナー) を開催
会場
物質・材料研究機構 千現地区 8階 中セミナー室
開催日: 時間
2015.11.13
10:00-12:00
参加料
無料
2019.10
Sun Mon Tue Wed Thu Fri Sat
1

2

3

4

5

6

7

8

9

10

11

12

13

14

15

16

17

18

19

20

21

22

23

24

25

26

27

28

29

30

31