Embedded Files
Program at a Glance
Program at a Glance
Wednesday, February 05
13.00 - 18.00
Wednesday, February 05
13.00 - 18.00
13.00 - 13.30: Opening Session
13.00 - 13.30: Opening Session
13.30 - 14.30: Invited Talk (Meelis Kull)
13.30 - 14.30: Invited Talk (Meelis Kull)
14.30 - 15.00: Coffee Break
14.30 - 15.00: Coffee Break
15.00 - 18.00: Talks: Uncertainty in ML
15.00 - 18.00: Talks: Uncertainty in ML
Thursday, February 06
09.00 - 18.00
Thursday, February 06
09.00 - 18.00
09.00 - 10.00: Invited Talk (Chris Cornelis)
09.00 - 10.00: Invited Talk (Chris Cornelis)
10.00 - 10.30: Coffee Break
10.00 - 10.30: Coffee Break
10.30 - 11.30: Poster Session 1
10.30 - 11.30: Poster Session 1
11.30 - 13.00: Talks: Set-based & Robust Methods
11.30 - 13.00: Talks: Set-based & Robust Methods
13.00 - 14.30: Lunch
13.00 - 14.30: Lunch
14.30 - 16.30: Talks: Causality, Humans & Knowledge
14.30 - 16.30: Talks: Causality, Humans & Knowledge
16.30 - 17.00: Coffee Break
16.30 - 17.00: Coffee Break
17.00 - 18.00: Poster Session 2
17.00 - 18.00: Poster Session 2
Friday, February 07
09.00 - 17.00
Friday, February 07
09.00 - 17.00
09.00 - 10.30: Talks: Uncertainty in Deep Learning
09.00 - 10.30: Talks: Uncertainty in Deep Learning
10.30 - 11.00: Coffee Break
10.30 - 11.00: Coffee Break
11.00 - 13.00: Talks: Uncertainty in Optimization
11.00 - 13.00: Talks: Uncertainty in Optimization
13.00 - 14.30: Lunch
13.00 - 14.30: Lunch
14.30 - 16.30: Talks: Uncertainty in Applications
14.30 - 16.30: Talks: Uncertainty in Applications
16.30 - 17.00: Closing Session (WUML 2026 announcement)
16.30 - 17.00: Closing Session (WUML 2026 announcement)
Coffee breaks will be held in the same building (U2-QUANTUM) where the workshop takes place.
Lunch breaks are self-organized, for dining recommendations, you can see the Venue page or ask the organizers
Invited Talks
Invited Talks
Meelis Kull
Meelis Kull
Professor, Institute of Computer Science, University of Tartu
Head, Estonian Centre of Excellence in Artificial Intelligence
The many faces of epistemic uncertainty
Epistemic uncertainty in machine learning systems reflects the limits of knowledge, influencing generalization, reliability and decision-making. Multiple definitions and estimation methods exist, often leading to conflicting views on how epistemic uncertainty should be interpreted, quantified and applied. This talk examines key approaches to modelling epistemic uncertainty, their theoretical foundations, and the implications of differing definitions and estimates for downstream decisions that rely on information about uncertainty.
Chris Cornelis
Chris Cornelis
Associate Professor, Department of Applied Mathematics, Computer Science and Statistics, Ghent University
Vice-president, International Rough Set Society
Machine learning with fuzzy rough sets
Fuzzy set theory (Zadeh, 1965) is a popular AI tool designed to model and process vague information. On the other hand, rough set theory (Pawlak, 1982) was proposed as a way to handle potentially inconsistent data inside information systems. In this lecture, I will explain how both theories can be combined, and how the resulting fuzzy rough sets may be applied in a machine learning context to handle gradual inconsistencies. In particular, I will cover fuzzy-rough feature selection and classification, and discuss a case study involving natural language processing.
Wednesday, February 05
Wednesday, February 05
Session "Uncertainty in Machine Learning" (15.00 - 18.00)
Epistemic Uncertainty in AI: A Trend - Keivan Shariatmadar
Explorations of the Softmax Space: Knowing When the Neural Network Doesn’t Know. - Daniel Sikar
Embedding of Uncertainty using Bayesian Reasoning - Goeran Kauermann
Baseline Aleatoric Uncertainty Estimates for Inverse Problems - Nina Maria Gottschling
Thursday, February 06
Thursday, February 06
Poster Session 1 (10.30 - 11.30)
Novin Shahroudi
Yusuf Sale
Mohammad Hossein Shaker Ardakani
Sam Goring
Marek Leibl
Jakob Gawlikowski
Paul Hofman
Kaizheng Wang
Valentin Margraf
Mira Juergens
Karim Belaid
Timo Löhr
Session "Set-Based and Robust Methods" (11.30 - 13.00)
Conformalized Credal Set Predictors - Alireza Javanmardi
Imprecise Probabilistic Machine Learning: Being Precise about Imprecision - Michele Caprio
Robust Classification in Bayesian Neural Networks - Kim-Dung Tran
Session "Causality, Humans and Knowledge" (14.30 - 16.30)
VariErr NLI: Separating Annotation Error from Human Label Variation - Siyao Peng
Noisy judgements and the pushing for "the optimal model" - Fabio Stella
Active Learning with Feature Acquisition - Arthur Hoarau
Rule induction with Granular Approximations - Henri Bollaert
Uncertainty in Causal Discovery - Alessio Zanga
Poster Session 2 (17.00 - 18.00)
Jonas Hanselle
Théo Dupuy
Constanze Schwan
Clemens Damke
Davide Cazzorla
Jelle Hüntelmann
Salvador Madrigal Castillo
Viacheslav Komisarenko
Federico Di Marco
Andrea Spinelli
Marco Piazza
Ludovica Di Marco
Friday, February 07
Friday, February 07
Session "Uncertainty in Deep Learning" (09.00 - 10.30)
Depth based epistemic uncertainty - Mihkel Lepson
Efficient Deep Ensemble Learning for Epistemic Uncertainty Modelling - Mehmet Ozgur Turkoglu
Uncertainty Disentanglement in Deep Learning Models - Sebastian Jimenez
Energy-based Epistemic Uncertainty for Graph Neural Networks - Dominik Fuchsgruber
Session "Uncertainty and Optimization" (11.00 - 13.00)
Advances in Multi-Agent in Stochastic Optimization - Larkin Liu
Simulation-based inference in diffusion MRI: From uncertainty mapping to probabilistic tractography - Jose Pedro Manzano-Patron
Robust and Distributionally Robust Optimization Models for Classification Problems - Francesca Maggioni
Session "Uncertainty in Applications" (14.30 - 16.30)
Information management using Generative AI in SME Health Care Facilities - Pawan Lingras
An Explainable Evidential Clustering Method : Application to Post-Stroke Rehabilitation - Victor Souza
Uncertainty estimation in color constancy - Marco Buzzelli
Reinforcement Learning and Fuzzy Logic Modelling for Personalized Dynamic Treatment - Marco Locatelli
Instructions for Talks and Posters
All talks are assigned 25 minutes: 20 minutes for the talk itself and 5 minutes for the Q&A
A laptop will be provided for the talks presentation
During the poster sessions, poster boards will be provided for attaching the posters
Note that, even though there are several copy shops near the Workshop site, printing a poster may take up to 2-4 working days: thus, we advise to bring your own posters
Posters should be in A1 or A0 format: note, however, that there are no strict requirement on the size or organization of the posters, as long as they fit on the supports.
Page updated
Report abuse