Andrea Agazzi
University of Bern
Imperial College London, South Kensington Campus
–
2026 / General information
The workshop will take place at 170 Queen’s Gate, an elegant Victorian venue in South Kensington.
Running for two and a half days, the programme begins on the morning of 23 November and continues through 25 November. A dedicated poster session is scheduled for the afternoon of 24 November.
2026 / Confirmed
Meet the researchers joining us in London this November. More speakers will be added as invitations are accepted.
University of Bern
TU Munich
University of Turin
Imperial College London
ENSAE Paris / CREST
Université Gustave Eiffel
University of Potsdam
Freie Universität Berlin
University of Ulm
University of Amsterdam
New York University
University of Warwick
Inria / Université Grenoble Alpes
2026 / Registration
Book your place at the 3rd Workshop on Gradient Flows for Sampling, Inference, and Learning.
Registration open
Registration is now available online, with reduced pricing for students.
Poster abstract submissions open
Final submission deadline
Submission contents Title + short abstract
File format PDF · maximum 1 page
Please submit as soon as possible. Abstracts will be reviewed as they arrive, and suitable submissions will be accepted on a rolling basis. Poster capacity is limited, so submissions may close before the final deadline once all available places have been filled.
Workshop registration is separate. A poster abstract does not reserve a place. Please check availability and register before submitting.
To be considered for the poster session, email your abstract to deniz.akyildiz@imperial.ac.uk using exactly the subject line below.
Email subject Poster Submission to Gradient Flows Workshop
Past events / 01—02
Two editions have already brought together the community around continuous dynamics, probabilistic inference and machine learning.
A one-day exploration of gradient-flow approaches for sampling, inference and learning—from Langevin and Wasserstein dynamics to diffusion models, robustness and sequential Bayesian methods.
Speakers
The inaugural one-day workshop connected continuous dynamics with Bayesian posterior sampling, variational inference, generative modelling and deep-learning optimisation.
Speakers