23-24 September 2023
ELLIS RobustML Workshop
Dipoli, Espoo, Finland
Harness the power of reliable, resilient machine learning with RobustML! As ML systems become increasingly integral, the demand for robust, reliable models under real-world changes is more vital than ever. From principles to practical deployment, our workshop delves into the heart of robust ML, tackling questions like:
- How can we apply robustness principles to real-world scenarios?
- How can we ensure ML system reliability in real-world applications?
- What societal and legal challenges do robust ML systems face?
Join us as we merge cutting-edge RobustML research with real-world demands. We'll delve into practical, theoretical, and societal challenges, exploring topics like characterizing real-world changes, ensuring system reliability, and managing societal and legal considerations. With RobustML, let's bring machine learning into the future, today!
Important Dates
Deadline for Funding Application – August 23, 2023 (11:59 pm AoE)
Deadline for Registration – August 31, 2023 (11:59 pm AoE)
Abstract Submissions due – September 4, 2023 (11:59 pm AoE)
Workshop Date – September 23-24, 2023
Workshop Registration
You can register to the workshop through this form.
Apply for funding
For those interested in applying for travel or accommodation funding, please complete this form.
Bear in mind, our resources are limited, and the final decision will be contingent upon the number of requests received and the amount of funding requested.
Topics
We encourage paper submissions regarding robustness and reliability of ML systems in the real world, and highlight emerging directions, paradigms, and applications which include:
Different perspectives on robustness:
Robustness to adversarial manipulation of models and data.
Robustness to distribution shifts.
The notions of robustness in the unsupervised learning landscape.
Reliability of Real-world Systems:
Exploring uncertainty quantification for enhancing system reliability.
Reviewing the role of challenging datasets and simulations in gauging system reliability.
Evaluating the potential and limitations of system properties verification for reliability assurance.
Societal and Legal Considerations:
Discussing legal implications of ML robustness, including hacking liabilities and performance guarantees.
Investigating strategies for building robust systems that promote fairness and avoid bias.
Exploring robustness requirements in high-stake domains like manufacturing, autonomous driving, and healthcare
Programme Schedule
Time Zone: GMT +02
(Tentative Schedule)
Day 1 (23rd September 2023)
8:30-9:15 Registration + Coffee break
9:15-9:30 Opening remarks - Professors Chris Holmes, Yee Whye Teh, Sami Kaski
9:30-10:30 Keynote 1 - Aaditya Ramdas (Distribution-free uncertainty quantification for black-box predictors under distribution drifts/shifts)
10:30-11:00 Coffee break
11:00-11:45 Guest 1 - Chris Williams (Automating Data Science)
11:45-12:30 Guest 2 - Arthur Gretton (Causal effect estimation with hidden confounders using instruments and proxies)
12:30-13:30 Lunch
13:30-14:15 Guest 3 - Isabel Valera (Causality for robust and ethical machine learning: from theory to practice)
14:15-15:00 Guest 4 - Chris Russell (Monsters from the I.D.)
15:00-15:30 Coffee break
15:30-16:30 Poster Session
16:30-17:30 Keynote 2 - Emmanuel Candes (Model-free selective inference)
19:00-21:00 Dinner at 10 kerros (Take elevator to 10th floor, Kaivokatu 3, Helsinki; next to Helsinki central railway station). Google map link
Day 2 (24th September 2023)
9:00-9:30 Coffee break
9:30-10:30 Keynote 3 - Kevin Murphy (Adapt or die: online and unsupervised learning for test-time adaptation)
10:30-11:00 Coffee break
11:00-11:45 Guest 5 - Maneesh Sahani (Recognition-parametrised learning: identifying latent probabilistic structure without generation)
11:45-12:30 Guest 6 - Eric Nalisnick (Learning to defer to one, multiple, or a population of expert(s))
12:30-13:30 Lunch
13:30-13:40 Closing remarks - Professors Chris Holmes, Yee Whye Teh, Sami Kaski
Speakers
Aaditya Ramdas
Assistant Professor
Carnegie Mellon University
Arthur Gretton
Professor
Gatsby Computational Neuroscience Unit
Chris Russell
Research Scientist
Amazon
Chris Williams
Professor
University of Edinburgh
Emmanuel Candès
Professor
Stanford University
Eric Nalisnick
Assistant Professor
University of Amsterdam
Isabel Valera
Professor
Saarland University
Kevin P Murphy
Research Scientist
Maneesh Sahani
Professor
Gatsby Computational Neuroscience Unit
Venue
Aalto Dipoli
Otakaari 24, 02150, Espoo, Finland
The workshop will be held in Dipoli, located in Otaniemi. This parkland-style campus established in the 1950s forms the core of Otaniemi. The overall vision of the campus was that of Alvar Aalto, and its individual buildings were designed by Aalto and other celebrated Finnish architects such as Reima and Raili Pietilä and Heikki and Kaija Sirén.
Organizers
Professor Samuel Kaski
Aalto University | FCAI
Professor Yee Whye Teh
University of Oxford
Professor Chris Holmes
University of Oxford
Co-Organizers
Chris Williams
University of Oxford
David Blanco Mulero
Aalto University
Fabian Falck
University of Oxford
Guneet Singh Dhillon
University of Oxford
Kevin Sebastian Luck
Vrije Universiteit Amsterdam
Markus Heinonen
Aalto University
Muhammad Faaiz Taufiq
University of Oxford
Let us know if you'll be attending!
Contact
If you have any questions please contact ellis DOT robustmlworkshop AT gmail DOT com