Past News
Conferences and lectures
Online2022.6.3 (JPN)
Scheduled to give a speech at the Kinki Chemical Association Computer Chemistry Division Public Lecture
Hiroki Iriguchi, Manager of Preferred Computational Chemistry (PFCC), will be giving a talk at the "Kinki Chemical Association Computer Chemistry Division (113th Regular Meeting) Public Lecture" to be held on Friday, June 3, 2022.
Click here for more details.
https://kinka.or.jp/compchem/prog/20220603_113prog.pdf
Please apply to participate here.
https://kinka.or.jp/form/view.php?id=99359
Details of the talk
| Date and Time | Friday, June 3, 2022 16:10-17:10 |
| 【Speakers】 | PFCC Manager: Hiroki Iriguchi |
| Venue | online |
| Abstract | In recent years, the importance of first-principles calculations has increased in material design for catalysts, batteries, semiconductors, etc., but many studies have imposed significant constraints on modeling due to the high calculation costs. There are various approaches to perform material calculations quickly with an accuracy close to that of first-principles calculations, but in most cases the aim is to reproduce a specific system with high accuracy, and the strong parameter dependency and low versatility have been an issue. Neural Network Potential (NNP) is a machine learning model that uses the results of first-principles calculations as training data to infer the energy and force of molecules and crystal structures, making it possible to remove the parameter dependency mentioned above, but previous NNPs were limited in the systems they could target. In this study, we aimed to use NNP to conduct material exploration in various fields, and developed PFP as a new general-purpose NNP that supports 55 elements. In this presentation, we will introduce examples of phenomenon analysis and material design using PFP, and discuss how the possibility of simulation-driven material development has expanded by demonstrating the success of achieving both high speed and accuracy. |
公開日：2022.05.19
-
URL
Copied
New Events and Seminars
NEW
Past News
Webinars
onlineThursday, October 1, 2026
Crystal Structure Prediction in Practice: Setting Up Searches and Reading Results with Matlantis CSP
Video Streaming
Webinars
OnlineOn-demand
[On-Demand Webinar] Simulating Semiconductor Unit Processes with a Universal Machine-Learning Potential
Video Streaming
Webinars
OnlineOn-demand
[On-Demand Webinar] Accelerating Computational Materials Design with CALPHAD, DFT, MLIPs, and AI Agents
Ootemachi9:30
Hands-on Seminar | computational chemistry with Matlantis: From Basics to Practical Applications
Video Streaming
Webinars
OnlineOn-demand
【On-demand】Expanding Matlantis Applications with Coding Agents: A New Workflow for computational chemistry using natural language
