Workshop Schedule 

Fri Dec 2, 2022


8:00 - 8:40  - coffee/breakfast

8:50 - 9:10 -  Irina Rish (Mila/UdeM): Overview  video/slides 

9:10 - 9:30 -  Felix Stollenwerk (AISweden) : Scaling GPTSw3

9:30 - 9:50  - Evi Gogoulou (RISE/KTH ):  Continual Learning in the presence of language shift video/slides 

9:50 - 10:20 coffee  (during the talks :)

10:20- 10:4- Ankesh Anand (DeepMind): Large-scale Pretrained Models and RL    video/slides

10:40 - 11:10 - David Krueger (U of Cambridge) : A discussion on AI Safety and Alignment     video/slides          

11:10- 11:30  - Leo Gao (OpenAI) : Scaling Laws for Reward Model Overoptimization     slides 

11:30 - 11:40 - coffee break/posters setup

11:50 - 12:10  -  Eric Michaud  (MIT): Neural Scaling Exponents Beyond the Manifold Dimension video/slides 

12:10 - 12:40 -  Ziming Liu (MIT) : Omnigrok: Grokking Beyond Algorithmic Data video/slides 

12:40  - 1:00  - Ethan Caballero (Mila/McGill): Broken Neural Scaling Laws   video/slides 

1:00 -2:00  -  lunch / poster session

2:35 - 3:00  - Jenia Jitsev (Juelich Supercomputing Center, ELLIS, LAION)  Supercomputers, supercommunities --> superintelligence? Kind of. video/slides 

3:00 - 3:30 - Christoph Schuhmann (LAION):   Together we are strong: Web scale data sets and foundation models through community organizing video/slides 

3:30 - 3:45 - Robert Kaczmarczyk  (TU  Munich, LAION ):   LAION-5B and beyond - datasets, models, and...? video/slides

3:45 - 4:10  - Julien Launay (lighton.ai): HQ data need not apply: training LLMs with with web data only video/slides

4:10 - 4:30 - Thomas Wolf (Hugging Face):  Collaboratively training a large multilingual language model video/slides

4:30 - 5:00  -  coffee break/posters

5:00 - 5:30  -  Jonas Andrulis and Robert Baldock (Aleph Alpha): Don't be dense: Sparse results and manifesto    video/slides

5:30 -  5:50 -  Joel Hestness (Cerebras): Toward Sparsity Scaling Laws: Needs and Opportunities​    video/slides

5:50  - 6:00  - Michael Trazzi:  Are AI Researchers Aligned About AGI Alignment? video/slides

6:00 - 6:30 -   panel: 

                               joining remotely: Jared Kaplan (Anthropic/JHU),  Surya Ganguli (Stanford),   David Krueger (U of Cambridge)

     in person:  Ethan Caballero   (Mila/McGill), Jenia Jitsev  (Juelich/LAION), Jonas Andrulis  (Aleph Alpha) ,   Joel Hestness (Cerebras)

7:00 - 10:00 - party

Posters:

Robert Baldock  (AlephAlpha): Pruned Models Win The New Hardware Lottery: Efficient Inference Of Pruned Large Language Models With Graphcore's IPU 

Ethan Caballero (Mila/McGill/UdeM): Broken Neural Scaling Laws

Muawiz Chaudhary (Mila)

Wei Deng (ML Research, Morgan Stanley): A Conditional Schrödinger Bridge Method forProbabilistic Time Series Imputation

António Góis (Mila)

Timothee Lesort (Mila)

TBA