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

Can lassonet works in transformer or S4(Structured State Spaces)? #38

Open
xueliu8617112 opened this issue Mar 16, 2023 · 6 comments
Open

Comments

@xueliu8617112
Copy link

i hope to use lassonet in S4, is it okay to do that?

@louisabraham
Copy link
Collaborator

Are you talking about Structured State Space for Sequence Modeling? I don't know very well the details nor how you want to apply LassoNet.

Do you have a concrete example in mind?

@xueliu8617112
Copy link
Author

yeah, it is S4 for Structured State Space for Sequence Modeling. I just want to do some time series prediction work in the industry. I see u use the NN or CNN or AE as the backbone, so I wonder whether s4 can be utilized as the backbone.

@louisabraham
Copy link
Collaborator

CNN as backbone is very experimental. The goal of LassoNet is to select variables in a meaningful space (eg features).

@943fansi
Copy link

@xueliu8617112 I am also interested in this topic. May I ask if you have made any substantial progress regarding this issue recently?

@ShailenSmith
Copy link

Hi @louisabraham, I'm a PhD student working on a project where we're hoping to use an extension of LassoNet for convolutional layers. The readme mentions that adding support for convolutional layers is in the works -- do you have any updates on this, or ideas for how the implementation would be done?

@louisabraham
Copy link
Collaborator

Hi, I should update the readme, we are not working on this anymore.

The difficulty is that you want to select features that are interpretable!

Sign up for free to join this conversation on GitHub. Already have an account? Sign in to comment
Labels
None yet
Projects
None yet
Development

No branches or pull requests

4 participants