Thoth seminars
What? The Thoth seminar series is a weekly seminar of the Thoth team at Inria Grenoble Rhones-Alpes. We combine internal speakers from the group with invited speakers, most commonly presenting their recent research but also a broader survey of a topic. The talks cover a variety of topics in machine learning. This group is open to everyone. Please contact the organizers if you would like to join our mailing list, to which we send the link for each talk.
When? Talks are (usually) at 11am on Tuesday.
Where? They are (usually) in F107.
Who? We are always looking for speakers! If you would be happy to come and give a talk, please contact us at scott.pesme@inria.fr or giacomo.meanti@inria.fr.
Upcoming seminars
Dorian Baudry (Inria Ghost Team)
Tuesday, September 23, 2025 – room F107 – 11am
Does stochastic gradient really succeed for bandits?
Kenta Vert, Amogh Tiwari, Juliette Bertrand (Inria Thoth Team)
Tuesday, September 30, 2025 – room F107 – 11am
Presentations of their ongoing works
Past seminars
Manvi Agarwal (Télécom Paris)
Tuesday, September 16, 2025 – room F107 – 11am
Understanding positional encoding through the lens of kernels, with applications to music generation
Yifei Lou (University of North Carolina at Chapel Hill)
Tuesday, July 1, 2025 – room F107 – 11am
Scale-invariant regularizations for sparse signal recovery
Mathieu Besançon (Inria Polaris Team)
Tuesday, June 17, 2025 – room F107 – 11am
Frank-Wolfe Methods and Applications
Pierre-Cyril Aubin-Frankowski (CERMICS, ENPC)
Wednesday (unusual date!), June 11, 2025 – room F107 – 11am
Gradient descent and flows with c(x,y) cost
Emmanuel De Bézenac (Inria Paris)
Tuesday, June 3, 2025 – F107 – 11am
Some insights into neural operators or solving PDEs
Kimmo Riihiaho & Ilkka Pölönen (University of Jyväskylä, Finland)
Tuesday, May 27, 2025 – room F107 – 11am
Spectral Forest Simulations with HyperBlend
Adrien Taylor (Inria Sierra Team)
Tuesday, May 20, 2025 – room F107 – 11am
Principled approaches and tools for the analysis and design of first-order optimization algorithms
Diane Larlus (Naver labs)
Tuesday, May 13, 2025 – Grand Amphi (unusual place!) – 11am
Learning Versatile Visual Representations
Alexei Efros (UC Berkeley)
Tuesday, May 6, 2025 – Grand Amphi – 11am
The joys of Generative Models as Data Stores
Ieva Petrulionyte (Inria Thoth Team)
Tuesday, April 29, 2025 – room F107 – 11am
Practical tutorial on model-based reinforcement learning
Romain Ménégaux (Inria Thoth Team)
Tuesday, April 22, 2025 – room F107 – 11am
Tutorial on Flow Matching
Hadrien Hendrikx (Inria Thoth Team)
Tuesday, April 15, 2025 – room F107 – 11am
Constructing an ML pipeline from scratch: the case of ecology
Derek Hoiem (University of Illinois)
Joint seminar with Naver-labs
Tuesday, April 8, 2025 – PLACE TBD – 11am
Three Elements to AI that Get the Job Done
Kouroche Bouchiat
Tuesday, March 25, 2025 – room A103 (different room than usual!) – 11am
Advances in Laplace Inference for Reliable and Interpretable Bayesian Deep Learning
Mathilde Caron (Google Grenoble)
Tuesday, March 18, 2025 – room F107 – 11am
Wikipedia-Scale Visual Entity Recognition
Etienne Boursier (Inria Celeste Team)
Tuesday, March 11, 2025 – room F107 – 11am
Training overparametrized neural networks: early alignment phenomenon and simplicity bias
Pia Bideau (Inria Thoth Team)
Tuesday, March 4, 2025 – room F107 – 11am
Learning with Motion: Combining Perception and Interaction through Movement-Driven Computation
Vicky Kouni (Cambridge University)
Friday, February 28, 2025 – room F107 – 11am
Model-based (unfolding) networks and where to find them: the saga
Antoine Chatalic (GIPSA lab)
Tuesday, February 25, 2025 – room F107 – 11am
Approximate Kernel Statistical Distances for Efficient Numerical Integration and Two-Sample Testing
Stéphane Lathuilière (Inria RobotLearn Team)
Tuesday, February 18, 2025 – room F107 – 11am
Expanding Vision-Language Models: Adding Modalities and Tasks
Zhiqi Kang (Inria Thoth Team)
Tuesday, February 11, 2025 – room F107 – 11am
Advancing Prompt-Based Methods for Replay-Independent General Continual Learning
Giacomo Meanti and Scott Pesme (Inria Thoth team)
Tuesday, February 4, 2025 – room F107 – 11am
Ongoing work on diffusion models
In this talk, we’ll cover some of our recent progress on our respective projets on diffusion models for inverse problems.
