Mila Tea Talks

A cornerstone of our community for many years, Tea Talks are scientific keynote-like talks held every week during the academic year. These presentations cover a range of technical subjects related to machine learning. The talks are also open to the public.

Yoshua Bengio with other professors and students talking together

How do Tea Talks Work?

Typically, an in-house or guest speaker gives a talk lasting around an hour, with around 15 minutes devoted to questions throughout. As most of the participants are Mila students and professors, the content is geared towards this level of understanding. 

Speakers are also encouraged to stay after the lecture to talk with students in individual or group meetings to discuss research interests and ideas. 

When Will the Next Tea Talks Take Place? 

Tea Talks are typically held in Mila's auditorium on Friday mornings at 10:30 a.m. EDT. The talks are also live-streamed on Google Meet via the following link: https://meet.google.com/nqu-embv-juy 

The Tea Talks series is taking a break for the summer and return at the beginning of September 2025.

Get Schedule Notifications

To join the Tea Talks external mailing list, visit this link, click on the 'Ask to join group' button, and provide a brief description of yourself (e.g., name and affiliation/organization).

Get Tea Talks schedule notifications

Past Sessions

Watch the recordings of the latest sessions.

Tea talk page de couverture de la présentation

Autonomous UI Agents: Open-Source Tools, Best Practices, and Safety

May 15, 2025

Tea talks cover presentation "Democratizing Open LLMs for Global Sovereign AI Applications"

Democratizing Open LLMs for Global Sovereign AI Applications

May 2, 2025

Cover presentation

Towards universal AI models for protein-RNA interaction prediction and RNA therapeutic design

April 25, 2025

Page cover of Autonomous Agents in the Era of Foundation Models

Autonomous Agents in the Era of Foundation Models

April 18, 2025

Advances and Challenges in Foundation Agents.

Advances and Challenges in Foundation Agents

April 4, 2025

 

Page cover of presentation

Learning from Imperfect Supervision

March 28, 2025

Cover presentation

Identifying Latent Structures in Data

March 14, 2025

Cover presentation of "Advancing Foundation Models Through Interaction"

Enhancing Foundation Models Through Interaction

March 7, 2025

Page cover of the presentation

Multi-Fidelity Active Learning for Scientific Discoveries

February 28, 2025

Have questions or interested in becoming our next speaker?