Skip to content
New issue

Have a question about this project? Sign up for a free GitHub account to open an issue and contact its maintainers and the community.

By clicking “Sign up for GitHub”, you agree to our terms of service and privacy statement. We’ll occasionally send you account related emails.

Already on GitHub? Sign in to your account

interpretability of models trained on xtal2png #190

Open
kjappelbaum opened this issue Jul 27, 2022 · 3 comments
Open

interpretability of models trained on xtal2png #190

kjappelbaum opened this issue Jul 27, 2022 · 3 comments
Labels
manuscript-enhancements Interesting things to explore that can enhance the manuscript

Comments

@kjappelbaum
Copy link
Contributor

kjappelbaum commented Jul 27, 2022

one interesting application of this representation might be that it might be explainable in a useful form.

If we train a model on the image, we can then use one of the established interpretability techniques to obtain a mask that highlights the 'important' parts of the image. If we decode this, we have the relevant structural fragments (and could also mine them - and potentially use them to assemble new structures).

the advantage of doing this on the image representation and not with a GNN is that the image should have fewer issues with longer-range interactions

@sgbaird sgbaird added the manuscript-enhancements Interesting things to explore that can enhance the manuscript label Jul 29, 2022
@sgbaird
Copy link
Member

sgbaird commented Aug 5, 2022

Following up from our chat, maybe the following two could be combined without too much hassle:

@HarshaSatyavardhan
Copy link

Hey what is the progress of this project. I am interested to work on this

@sgbaird
Copy link
Member

sgbaird commented Nov 10, 2023

I don't think either @kjappelbaum or I have immediate plans to explore the interpretability piece. Feel free to give it a try and let us know how it goes! Happy to provide feedback or suggestions.

Sign up for free to join this conversation on GitHub. Already have an account? Sign in to comment
Labels
manuscript-enhancements Interesting things to explore that can enhance the manuscript
Projects
None yet
Development

No branches or pull requests

3 participants