Gradient-free stochastic optimization
This talk will deal with optimization problems in a statistical learning setup where the learner has no access to unbiased estimators of the gradient of the objective function. It includes stochastic optimization with zero-order oracle, continuum bandit and contextual continuum bandit problems. I’ll give an overview of recent results on minimax optimal algorithms and fundamental limits for these problems.
Date:
6 June 2025, 15:00
Venue:
Mathematical Institute, Woodstock Road OX2 6GG
Venue Details:
L1, Mathematical Institute, Andrew Wiles Building
Speaker:
Professor Alexandre Tsybakov (CREST-ENSAE Paris)
Organising department:
Department of Statistics
Organisers:
Patrick Rebeschini (Department of Statistics, University of Oxford),
Beverley Lane (Department of Statistics, University of Oxford),
Professor Simon Myers (University of Oxford)
Organiser contact email address:
events@stats.ox.ac.uk
Hosts:
Professor Christl Donnelly (University of Oxford),
Patrick Rebeschini (Department of Statistics, University of Oxford),
Professor Simon Myers (University of Oxford)
Part of:
Annual Joint Maths/Stats Colloquium
Booking required?:
Required
Booking url:
https://www.stats.ox.ac.uk/events/maths-stats-colloquium-professor-alexandre-tsybakov
Booking email:
events@stats.ox.ac.uk
Cost:
Free of charge
Audience:
Members of the University only
Editor:
Beverley Lane
