Andrej Risteski (@risteski_a) 's Twitter Profile
Andrej Risteski

@risteski_a

Machine learning researcher.
Assistant Professor, ML department at CMU (@mldcmu).

ID: 110103071

linkhttp://www.andrew.cmu.edu/user/aristesk/ calendar_today31-01-2010 10:20:43

608 Tweet

3,3K Followers

1,1K Following

The Data Science Institute at Columbia University (@datascicolumbia) 's Twitter Profile Photo

Join us on Friday, May 2 for a Foundations of Data Science Center Tutorial! This day-long event features talks from leading experts Andrej Risteski and Ankur Moitra. Register here: datascience.columbia.edu/event/center-e…

Join us on Friday, May 2 for a Foundations of Data Science Center Tutorial!

This day-long event features talks from leading experts <a href="/risteski_a/">Andrej Risteski</a> and Ankur Moitra.

Register here: datascience.columbia.edu/event/center-e…
Andrej Risteski (@risteski_a) 's Twitter Profile Photo

Check out Abhishek Panigrahi 's talk and poster at #ICLR2025 on understanding the mechanism by which progressive distillation helps. Some theory in toy setup, and a variety of experiments on synthetic and real data. Joint w/ Bingbin Liu, Surbhi Goel, Sadhika Malladi Details below !

Max Simchowitz (@max_simchowitz) 's Twitter Profile Photo

Hey Everyone!! I will be giving a lecture at the RL Theory Virtual Seminar tomorrow, on my new paper about the “Pitfalls of Imitation Learning" in continuous action spaces. 🧵 below; please read because the time is somewhat TBD..... 🧐

Hey Everyone!! I will be giving a lecture at the RL Theory Virtual Seminar tomorrow, on my new paper about the “Pitfalls of Imitation Learning" in continuous action spaces.  🧵 below;  please read because the time is somewhat TBD..... 🧐
Andrej Risteski (@risteski_a) 's Twitter Profile Photo

Looking forward to speaking at the Columbia University Data Science Institute tomorrow, alongside Ankur Moitra. If you're in the area, stop by and say hi!

Andrej Risteski (@risteski_a) 's Twitter Profile Photo

Recently, we were trying to understand whether using code as an intermediary abstraction would be useful for ML-based PDE solvers. Turns out, a good LLM workflow can do quite well! Check out Shanda's thread below for more details.