Embedded Files
What to do if you're interested:
Sign up for our mailing list on MCommunity here by clicking "Join Group" in the bar at the top. This will let you receive weekly communications on reading group meetings.
We need volunteers each week to present -- consider signing up to present here. Presentations don't have to be perfect; everyone's busy, and no one wants a reading group presentation to be super stressful! Instead, focus on being able to hold a meaningful conversation about the paper.
For further questions, see this page.
Overview
What?
This is a reading group to discuss recent research papers in machine learning theory. "Machine learning theory" is a very broad category that includes but is not limited to optimization, generalization, interpetability, robustness, theory of deep learning...if the paper you want to present is on machine learning, optimization, or statistical learning, then it probably fits.
Note that computer vision, NLP, and reinforcement learning have their own reading groups, so you may get more presenting a relevant paper to those audiences than here.
Who?
All are welcome to join. This group is primarily intended for graduate students across departments who are researching topics in machine learning theory, but if the material is relevant to your work or is just interesting, feel free to drop by!
When?
Weekly 1-hour meetings on Tuesdays 12:00-1:00pm Eastern.
Where?
On Zoom at https://umich.zoom.us/j/96965198538. We will remain remote for Summer 2022, and will be meeting in-person options beginning Fall 22.
Why?
Our main goals are to 1) connect people in different departments who are conducting research on machine learning theory, and 2) allow graduate students to share and learn about new ideas in ML theory with each other by discussing papers. We hope there can be some other benefits for you too, be it practicing presentation skills, learning how to digest papers, showcasing your own work, etc.
Presentation schedule below. To sign up, you can access the spreadsheet here or click it below to open it in a new tab. Requires UM authentication.
Page updated
Report abuse