Yandex at ICML 2023

At the International Conference on Machine Learning (ICML 2023), Yandex Research will present papers on modelling tabular data with diffusion models, a model-parallel training algorithm designed for poorly connected devices and other topics. We also co-author a tutorial ​​on reinforcement learning from human feedback. 
For a brief summary of each paper, please see our previous ICML post; in this announcement, you will find a schedule of our presentations at the conference with links to pages at the ICML website.

Tutorial: Reinforcement Learning from Human Feedback The link has been copied to clipboard

This 2-hour tutorial, presented by Nathan Lambert from Hugging Face and Dmitry Ustalov from Toloka, covers both key aspects of RLHF: core machine learning techniques and human-in-the-loop data collection. You can read its ​​agenda on the tutorial webpage.

Papers by Yandex Research and our collaborators The link has been copied to clipboard

TabDDPM: Modelling Tabular Data with Diffusion Models by Akim Kotelnikov, Dmitry Baranchuk, Ivan Rubachev, Artem Babenko.
SWARM Parallelism: Training Large Models Can Be Surprisingly Communication-Efficient by Max Ryabinin, Tim Dettmers, Michael Diskin, Alexander Borzunov.
Poster presentation: Jul 25 11:00 AM HST, Exhibit Hall 1 #217
Which Tricks Are Important for Learning to Rank? by Ivan Lyzhin, Aleksei Ustimenko, Andrey Gulin, Liudmila Prokhorenkova.
FlexGen: High-throughput Generative Inference of Large Language Models with a Single GPU by Ying Sheng, Lianmin Zheng, Binhang Yuan, Zhuohan Li, Max Ryabinin, Beidi Chen, Percy Liang, Christopher Re, Ion Stoica, Ce Zhang.
Oral presentation: Jul 26, 2023, 4:40 PM HST, Meeting Room 313
Poster presentation: Jul 27, 2023, 10:30 AM HST, Exhibit Hall 1 #216