Dr. Matti Hellström is SCM’s product developer and supervised the development of ParAMS, the parameter optimization toolkit of the Amsterdam Modeling Suite.
In this third session of the AMS & Science 2022 webinar series titled

Parametrize ReaxFF and DFTB with ParAMS

Matti will provide an overview on how to optimize empirical models such as ReaxFF and DFTB with the help of ParAMS.

Interested in parametrizing your own model?
Check out the ParAMS tutorial collection [ Ссылка ]
and the upcoming ParAMS ReaxFF parametrization challenge
[ Ссылка ]

Content
00:00 intro
1:06 What is ParAMS?
3:50 ReaxFF/DFTB applications
7:18 ReaxFF reparametrization problem from industry
10:24 AMS and ParAMS make it easy to to fix the issue
13:17 General ParAMS workflow
14:37 The loss function
16:57 Types of reference values
21:00 ParAMS demo: reference values
27:57 ParAMS settings
29:14 The ParAMS ReaxFF parametrization challenge 2022
30:23 ParAMS demo: output ReaxFF parametrization
32:48 ParAMS demo: parameter settings
35:16 12 bonus features
40:15 Future developments
40:40 Summary
41:46 Discussion

свернуть