November 22: an afternoon of presentations and discussions on large language models!
(announcement)
As part of the scientific activities of the Trustworthy AI Chair, we are pleased to invite you to an afternoon of presentations and discussions on large language models. LLMs (thanks in particular to Transformers) are at the origin of the upheaval of “Generative” AIs and the leap in performance of chatbots like ChatGPT, Claude, Gemini, Copilot.
While it is difficult today to do without them and not adopt them, do you understand the underlying technologies and their limitations?
We offer you, through 3 scientific presentations by specialists in the field, accessible to the general public, to understand their functioning, how they can be adapted to different tasks, and what strong limitations they still unfortunately present.
Two hours of exchanges with specialists to give you a complete overview!
Program
Wednesday, November 20, 2024
- 2:00 PM: Large language models: Innovation or progress? (Laurent Simon, LaBRI / Bordeaux INP)
- 2:15 PM: Understanding Transformers from the inside (Nathanaël Fijalkow, LaBRI / CNRS)
- 2:50 PM: Pre-training, post-training, and fine-tuning of LLMs: a short practical introduction (Erwan Simon, CATIE)
- 3:25 PM: How to raise awareness of the limitations of AI assistants (Sonia Tatbi, AI Consultant)
- 4:00 PM: End of the AI afternoon
Where?
At ENSEIRB-MATMECA, Amphi F
www.enseirb-matmeca.fr
(1 Avenue du Dr Albert Schweitzer, 33400 Talence, France)
Note: Registration is free but mandatory:
Register here