Workshop · Gatsby Computational Neuroscience Unit · UCL · June 18-19, 2026

Advances in Adaptive
Experimentation Workshop

A two-day workshop at the Gatsby Computational Neuroscience Unit, UCL, on June 18-19, 2026, focused on adaptive experimentation, causal inference, online learning, and modern statistical methods for sequential decision-making.

We warmly encourage applications from researchers and students interested in these topics. Accepted applicants will be asked to pay a registration fee (TBC) for the workshop.

Hosted by the Gatsby Computational Neuroscience Unit at UCL, with support from ELLIS Unit London and CSML.

01 — About

Workshop Overview

This workshop brings together researchers working on adaptive experimentation, causal inference, online learning, and related methods for sequential decision-making. It is intended as a focused meeting point for theory, methodology, and discussion across neighboring communities.

Hosted at the Gatsby Computational Neuroscience Unit at UCL on June 18-19, 2026, the workshop will feature invited talks, informal discussion, and a poster session. The goal is to create an event with strong technical depth and enough room for interaction.

i

Adaptive Experimentation

Exploration-exploitation tradeoffs, stability conditions, adaptive data collection.

ii

Bandits and Online Learning

Regret minimization, best arm identification, structured feedbacks.

iii

Causal Inference

Identification, nuisance-robust methods, semiparametric efficiency theory.

iv

Open Problems

Novel and/or open questions at the interface of online experimentation, learning and inference.

02 — Invited Speakers

Confirmed Speakers

Confirmed speakers spanning bandits, online learning, inference, and modern adaptive methods.

Tor Lattimore

Tor Lattimore

DeepMind
Gergely Neu

Gergely Neu

ICREA & UPF
Koulik Khamaru

Koulik Khamaru

Rutgers University
Kelly Zhang

Kelly Zhang

Imperial College London
Aurélien Bibaut

Aurélien Bibaut

Netflix
Ian Waudby-Smith

Ian Waudby-Smith

UC Berkeley
03 — Schedule

Two-Day Programme

The programme spans two full days at Gatsby UCL on June 18-19, 2026, with invited talks, an open problems session, and posters across both days. Speaker order and exact timings can be refined once the programme is finalized.

2
Days at Gatsby UCL
Day 1 Thursday, June 18 Invited talks, coffee breaks, and an open problems session.
Day 2 Friday, June 19 Invited talks, a contributed talk, and the poster session.

Open Problems Session

The workshop will include a dedicated open problems session: a space for sharp questions, partial ideas, and technically interesting obstacles that deserve collective attention.

Pitch the question Present a problem, conjecture, limitation, or unresolved technical point in a concise format.
Open the discussion Use the room to test intuitions, connect methods, and surface approaches from neighboring areas.
Leave with leads The goal is not a polished result, but a better question, a sharper framing, or a new collaboration.
09:00

Coffee

09:30 - 09:45

Introduction

Organizers

09:45 - 10:45

Tor Lattimore

Talk: Continuous-time As a Tool for Understanding in RL, ML and Beyond

10:45 - 11:00

Coffee Break

11:00 - 12:00

Aurélien Bibaut

Talk: Fast Best-in-Class Regret for Contextual Bandits

12:00 - 13:30

Lunch

13:30 - 14:30

Gergely Neu

Invited Talk: TBA

14:30 - 14:45

Coffee Break

14:45 - 15:45

TBA

Invited Talk: TBA

15:45 - 16:15

Coffee Break

16:15 - 17:15

Open Problems Session

Short presentations and discussion of open questions

09:30 - 09:45

Coffee

09:45 - 10:45

Koulik Khamaru

Invited Talk: TBA

10:45 - 11:00

Coffee Break

11:00 - 12:00

Kelly Zhang

Talk: Impatient Bandits: Optimizing for the Long-Term Without Delay

12:00 - 13:30

Lunch

13:30 - 14:15

Ian Waudby-Smith

Talk: Multi-Armed Sequential Hypothesis Testing by Betting

14:15 - 17:15

Poster Session

Accepted posters

04 — Call for Posters

Poster Submission

We invite poster submissions on recent or ongoing theoretical work in adaptive experimentation and closely related foundations of sequential decision-making.

  • Submit a poster title and abstract
  • Suitable for ongoing theoretical work, recent results, or discussion-stage ideas
  • Submissions should foreground a theoretical contribution in adaptive experimentation
Submission Format Google Form submission with a title, abstract and URL link if publicly available. The application link is the same as the poster submission link.
Submit →

Important Dates

Form opensApril 20, 2026
Abstract deadlineMay 21, 2026
Acceptance noticeMay 22, 2026
Workshop daysJune 18-19, 2026
05 — Organizers

Organizing Committee

Houssam Zenati

Houssam Zenati

Gatsby Computational Neuroscience Unit · UCL

Bariscan Bozkurt

Bariscan Bozkurt

Gatsby Computational Neuroscience Unit · UCL

Arthur Gretton

Arthur Gretton

Gatsby Computational Neuroscience Unit · UCL

06 — Location

Venue Information

The workshop will take place at the Gatsby Computational Neuroscience Unit in central London.

Address

Gatsby Computational Neuroscience Unit
25 Howland St
London W1T 4JG
United Kingdom

Getting There

The venue is located in Fitzrovia, within easy reach of central London transport links and the wider UCL area.

  • Goodge Street — Northern line — about 5 to 10 minutes on foot
  • Warren Street — Northern and Victoria lines — about 8 minutes on foot
  • Great Portland Street — Circle, Hammersmith & City, and Metropolitan lines — about 12 minutes on foot
  • Tottenham Court Road — Northern and Central lines, plus the Elizabeth line — about 15 minutes on foot
Gatsby building exterior
Host Unit

Gatsby Computational Neuroscience Unit, hosting the workshop at 25 Howland Street.

University College London campus
University

University College London.

07 — Sponsors

Sponsors