–
Coffee and Check-In
–
LLM Reasoning
Denny Zhou (Google DeepMind)
–
Break
–
The Key Ingredients of Optimizing Test-Time Compute and What's Still Missing
Aviral Kumar (Carnegie Mellon University)
–
Lunch (on your own)
–
Openthinker: curating a reasoning post-training dataset and training open data reasoning models
Alex Dimakis (UC Berkeley)
–
Break
–
LLM skills and meta-cognition: scaffolding for new forms of learning?
Sanjeev Arora (Princeton University)
–
Reception
–
Coffee and Check-In
–
What will Transformers look like in 2027?
Yoon Kim (Massachusetts Institute of Technology)
–
Break
–
Reducing the Dimension of Language: A Spectral Perspective on Transformers
Elad Hazan (Princeton University)
–
Lunch (on your own)
–
Mixed-modal Language Modeling: Chameleon, Transfusion, and Mixture of Transformers
Luke Zettlemoyer (University of Washington)
–
Break
–
The Frontier between Retrieval-augmented and Long-context Language Models
Danqi Chen (Princeton University)
–
Attention to Detail: Fine-Grained Vision-Language Alignment
Kai-Wei Chang (UCLA)
–
Coffee and Check-In
–
Inference Scaling: A New Frontier for AI Capability
Azalia Mirhoseini (Stanford / DeepMind)
–
Break
–
DeepSeek-R1 Thoughtology: <Thinking> about LLM Reasoning
Siva Reddy (IVADO - Mila - McGill University)
–
Lunch (on your own)
–
On cognitive maps, LLMs, world models, and understanding
Dileep George (Google DeepMind)
–
Break
–
Talk by
Zaid Harchaoui (University of Washington)
–
Break
–
Light Refreshments
–
The Move Toward AGI: Why Large Language Models Surprised Almost Everyone, and What’s Coming Next | Theoretically Speaking
Anil Ananthaswamy (Simons Institute), Dileep George (Google DeepMind), Azalia Mirhoseini (Stanford), Luke Zettlemoyer (University of Washington and Meta)
–
Coffee and Check-In
–
On Knowledge Separation and Latent Diffusion for Text
Kilian Weinberger (Cornell University)
–
Break
–
Controllable and Creative Natural Language Generation
Nanyun (Violet) Peng (UCLA)
–
Lunch (on your own)
–
Transformers can learn compositional function
Jason Lee (Princeton University)
–
Break
–
Predicting and optimizing the behavior of large ML models
Andrew Ilyas (Stanford University)
–
Panel Discussion
–
Coffee and Check-In
–
Towards sequence-to-sequence models without activation functions
Grigorios Chrysos (University of Wisconsin-Madison)
–
Break
–
The Power of Resets! Learning better, one reset at a time
Kianté Brantley (Harvard University)
–
Lunch (on your own)
–
SILO Open LM: Training LMs on Siloed Datasets
Sewon Min (UC Berkeley)
–
Break
–
The Future of Language Models: A Perspective on Evaluation
Swabha Swayamdipta (University of Southern California)
–
Learning Generative Models from Corrupted Data
Gianis Daras (MIT)