Chaque atelier produira une publication non technique destinée aux décideurs publics et privés ainsi qu'au public. Le premier atelier, sur l'équité, l'interprétabilité et la protection de la vie privée pour les systèmes algorithmiques, aura lieu à l'Institut Alan Turing les 3 et 4 juin.
Au programme
Mon, 3rd June
8h30-9h30: Breakfast, Registrations
9h30-9h40: AW welcome and intro
9h40-9h55: Minister intro (Margot James MP, Minister of State for Digital, Culture, Media and Sport) still tbc whether she will confirm on morning of 3 or 4 June
Interpretability sessions:
9h55: Nozha Boujemaa - Interpretability of AI in Personalized Medicine
10h10: Plamen Angelov - Transparent, Interpretable and Explainable Classifiers 1025: Chris Russell - Counterfactual explanations in practice
10h40: Arthur Gretton - Interpretable model comparison
11h00 Coffee Break
11h15 David Leslie on ICO (regulator) guidance on interpretability
11h30-12h15 Panel Discussion on Interpretability, What do we need and how can we get there? Call for collaborators?
12h15-13h15 Lunch
Fairness sessions:
13h15-13h45: Christine Balagué - Towards using fairness in product recommender systems in marketing
13h45: Matt Kusner - Making Decisions the Reduce Discriminatory Impact
14h00: Claire Nedellec - Making textual data FAIR and interpretable by ontology-based information extraction. Where FAIR means "Findable, Accessible, interpretable, Reusable." Serge Pajak - Cases of ethical issues with IoT
Gracia Cecere - STEM and Teens: Algorithms in social media Panel discussion on Fairness? + Roger Taylor (CDEI)
Call for collaborators?
Privacy sessions:
15h00 Coffee break
15h15: Catuscia Palamidessi (Fr) - Local Differential Privacy and trade-off with Utility 15h30: Pierre Zweigenbaum (Fr) - Clinical text mining and privacy
15h45: Jon Crowcroft - Secure data havens - where are we today?
16h00: Emiliano de Cristofaro - Privacy and Machine Learning: It's Complicated 1615: Adria Gascon
16h16-17h00: Panel discussion on Privacy (+ Sophia A-B)
18h00 - 20h00: Reception at Quebec House
Tue 4 June
8h30-9h30: Breakfast, Registrations for people who can only join on 4 June 9h30-9h40: AW intro to second day
Robustness sessions
9h40-9h55: Pushmeet Kohli: Towards Verified Learnt Intelligence
9h55-10h10: Alan Mackworth: Building Trustworthy Artificial Autonomous Agents
10h10-10h25: Frederic Pascal: On Robust Estimation for high-dimensional settings applied to Signal and Image Processing
10h25-10h40: Michele Sebag: Structural Agnostic Modelling: Adversarial Learning of Causal Graphs
10h40-10h55: Coffee break
10h55-11h10: Bertrand Thirion - Statistical inference in high-dimension & application to medical imaging
11h10-11h25: Gaël Varoquaux (Fr) – Dirty data: learning without curating 1125-1140: Laurence Devillers - Affective computing & nudging with a robot 1140-1155: Joel Martin - Examples of NLP for good
11h55-12h40 Panel discussion on robustness
Call for collaborators?
Policy sessions
12h40-13h40: Lunch
13h40-14h00: Michael Geist
14h00-14h20: Tim Clement-Jones - Online harms (+ Sarah C) 14h20-14h40: Ian Kerr
(note. Villani paper)
Alan Winfield - on transparency and the approach we are taking in drafting IEEE standard P7001 on Transparency in Autonomous Systems
15h30-16h30: Breakout sessions to scope out collaboration? (shared google doc)
16h30-17h00: Wrap up and presentation of future projects for collaboration, maybe Press Release
19h Dinner near King’s Cross (or reception in the French Embassy tbc)