1st Oslo Invitational Workshop on Model-Agnostic Explainable AI
- Avdeling SAMBA
- Involverte fagområder Statistisk modellering, Maskinlæring og Kunstig intelligens (AI), Forklarbar kunstig intelligens
- Sted Norsk Regnesentral, Gaustadalleen 23A
Time and place: September 12, 2024 08:45 AM – 3:30 PM, Norsk Regnesentral
Welcome to the 1st Oslo Invitational Workshop on Model-Agnostic Explainable AI (XAI)!
Join us and our esteemed group of speakers as we discuss core XAI methodology and explore its practical applications.
The workshop is open to everyone, though background knowledge of XAI and interpretable machine learning is advisable. The first session is specifically aimed at researchers in relevant fields, while the second session is designed for data scientists and practitioners interested in practical applications of XAI.
The workshop will consist of two sessions and you may attend either one or both:
Please note that there are a limited number of available places and admission is granted on a first-come, first-served basis.
Tentative programme
| From 08.45 | Registration, coffee and welcome | ||
| METHODOLOGY (09.00-12.20) | 09.00 | Setting the stage and XAI-research at NR | Martin Jullum, Norwegian Computing Center |
| 09.15 | Global interpretation methods: Overview and recent developments | Giuseppe Casalicchio, LMU Munich | |
| 9.55 | Pitfalls of global interpretation methods | Julia Herbinger, Leibniz Institute for Agricultural Engineering and Bioeconomy | |
| 10.35 | Coffee break | ||
| 10.50 | Methods for estimating conditional Shapley values, and when to use them | Lars Henry Berge Olsen, University of Oslo | |
| 11.20 | MCCE: Monte Carlo sampling of realistic counterfactual explanations | Kjersti Aas, Norwegian Computing Center | |
| 11.50 | XAI for misalignment detection: Comparison of local and global methods | Inga Strümke, NTNU | |
| 12.20 | LUNCH (free of charge) | ||
| APPLICATION (13.05 – 15.30) | 13.05 | Shapley values and their uncertainties in complex regression models for epidemiological applications | Mark van De Wiel, Amsterdam UMC |
| 13.45 | Explego: An XAI-method selection tool | Martin Jullum, Norwegian Computing Center | |
| 14.00 | Transparent modelling in the insurance industry | TBD, Gjensidige | |
| 14.15 | Transparency and explainability for credit risk modelling in the crowdfunding business | Mirko Drazic, FundingPartner | |
| 14.30 | Coffee break | ||
| 14.45 | Panel and joint discussion: Experiences on practical usage of XAI | Lead by Anders Løland, Norwegian Computing Center | |
| 15.20 | Brief summary and closing remarks |
