Wednesday, 22.04.2026
09:00 Welcome introduction
09:10 Reliable counterfactuals for machine learning models
Organized by: Marika Kaden, Benjamin Paassen, Barbara Hammer, Thomas Villmann
09:10 Reliable Counterfactuals for Machine Learning Models - Current Aspects and Perspectives
- Marika Kaden, University of Applied Sciences Mittweida, Saxon Institute for Computational Intelligence and Machine Learning (Germany)
- Benjamin Paassen, Bielefeld University (Germany)
- Barbara Hammer, CITEC - Bielefeld University (Germany)
- Ronny Schubert, University of Applied Sciences Mittweida, Saxon Institute for Computational Intelligence and Machine Learning (Deutschland)
- Thomas Villmann, Mittweida University of Applied Sciences, Saxon Institute for Computational Intelligence and Machine Learning (Germany)
09:30 Identifying counterfactual probabilities using bivariate distributions and uplift modeling
- Théo Verhelst, Advanced Concepts Team, European Space Agency (Netherlands)
- Gianluca Bontempi, Machine Learning Group, ULB, Belgium (B)
09:50 Graph Diffusion Counterfactual Explanation
- David Bechtoldt, Technische Universität Berlin (Germany)
- Sidney Bender, Technische Universität Berlin (Germany)
10:10 Reliable counterfactuals for machine learning models - poster spotlights
Organized by: Marika Kaden, Benjamin Paassen, Barbara Hammer, Thomas Villmann
10:10 Geometric-analytical Generation of Counterfactuals for Prototype-based Classifiers
- Marika Kaden, University of Applied Sciences Mittweida, Saxon Institute for Computational Intelligence and Machine Learning (Germany)
- Lynn Reuss, University of Applied Sciences (Germany)
- Thomas Villmann, Mittweida University of Applied Sciences, Saxon Institute for Computational Intelligence and Machine Learning (Germany)
10:11 Interpretability, privacy, fairness
10:11 Learning Counterfactual Densities via Marginal Contrastive Discrimination
- Katia Meziani, Université Paris Dauphine-PSL (France)
- Aminata Ndiaye, CEREMADE, UMR 7534 Université Paris Dauphine PSL - France (France)
- Madalina Olteanu, CEREMADE, UMR 7534 Université Paris Dauphine PSL - France (/)
10:31 Low-Rank Lens for Scalable LLMs Interpretability
- Giuseppe Trimigno, University of Parma (Italy)
- gianfranco lombardo, University of Parma (Italy)
- Stefano Cagnoni, University of Parma (Italy)
10:51 Interpretability, privacy, fairness - poster spotlights
10:51 Breaking the Trade-off between Performance and Cost with an Improved Untargeted APGD
- Ammar Al-Najjar, University of Szeged (Hungary)
- Mark Jelasity, University of Szeged (Hungary)
10:52 Fairness in machine learning: A Compact Survey
- Jeremy de Bodt, UCLouvain (Belgium)
- Dounia Mulders, UCLouvain (Belgium)
- Cyril de Bodt, UNamur - naXys (Belgium)
- John Lee, UCLouvain (Belgium)
- Marco Saerens, UCLouvain (Belgium)
10:53 DeepFedNAS: Pareto Optimal Supernet Training for Improved and Predictor-Free Federated Neural Architecture Search
- Bostan Khan, Mälardalen University (Sweden)
- Masoud Daneshtalab, Mälardalen University (Sweden)
10:54 FedHENet: A Frugal Federated Learning Framework for Heterogeneous Environments
- Alejandro Dopico-Castro, Universidade da Coruña, CITIC (Spain)
- Oscar Fontenla-Romero, Universidade da Coruña (Spain)
- Amparo Alonso-Betanzos, CITIC, Universidade da Coruña (Spain)
- Bertha Guijarro-Berdiñas, Universidade da Coruña (Spain)
- Ivan Perez Digon, Universidade da Coruña (Spain)
10:55 RKLU: Redistributive KL Distillation for Efficient Retain-Free Machine Unlearning
- Varun Sampath Kumar, University of Southern Denmark (Denmark)
- Esmaeil S Nadimi, University of Southern Denmark (Denmark)
- Vinay Chakravarthi Gogineni, University of Southern Denmark (Denmark)
10:56 Coffee break
11:15 Neuro Symbolic AI and Complex Data
Organized by: Luca Pasa, Luca Oneto, Davide Anguita, Nicolò Navarin, Davide Rigoni
11:15 Neuro Symbolic AI and Complex Data
- Luca Oneto, University of Genoa (Italy)
- Nicolò Navarin, University of Padua (Italy)
- Luca Pasa, University of Padova (Italy)
- Davide Rigoni, University of Padova (Italy)
- Davide Anguita, DIBRIS - University of Genova (Italy)
11:35 Multi-label Complementary Labels Learning under Hard Logical Constraints
- Luca Oneto, University of Genoa (Italy)
- Yi Gao, Southeast University (China)
- Davide Anguita, DIBRIS - University of Genova (Italy)
- Fabio Roli, University of Genoa (Italy)
- Min-Ling Zhang, Southeast University (China)
- Fulvio Mastrogiovanni, University of Genoa (Italy)
11:55 Ring-constrained Molecular Graph Generation with Diffusion Models
- Davide Rigoni, University of Padova (Italy)
- Rana İşlek, University of Padova (Italy)
- Nicolò Navarin, University of Padua (Italy)
12:15 Code-Guided Reasoning in Vision-Language Models for Complex Diagram Understanding
- Daniel Steinigen, Fraunhofer IAIS (Germany)
- Lucie Flek, Fraunhofer IAIS (Germany)
- Sebastian Houben, Fraunhofer IAIS (Germany)
12:35 Neuro Symbolic AI and Complex Data - poster spotlights
Organized by: Luca Pasa, Luca Oneto, Davide Anguita, Nicolò Navarin, Davide Rigoni
12:35 NSA: Neuro-symbolic ARC Challenge
- Pawel Batorski, Heinrich Heine Universität Düsseldorf (Germany)
- Jannik Brinkmann, University of Mannheim (Germany)
- Paul Swoboda, Heinrich Heine Universität Düsseldorf (Germany)
12:36 Lunch
14:10 Beyond Performance: Comprehensive Evaluation Strategies for Impactful Machine Learning
Organized by: Valerie Vaquet, Ulrike Kuhl, Saša Brdnik, Benjamin Paassen
14:10 Beyond Performance: Comprehensive Evaluation Strategies for Impactful Machine Learning
- Valerie Vaquet, CITEC, Bielefeld University (Germany)
- Ulrike Kuhl, Bielefeld University (Germany)
- Saša Brdnik, University of Maribor, Faculty of Electrical Engineering and Computer Science (Slovenia)
- Benjamin Paassen, Bielefeld University (Germany)
14:30 High Performance, Low Reliability: Uncertainty Benchmarking for Tabular Foundation Models
- José Lucas De Melo Costa, Université Paris-Saclay (France)
- Fabrice Popineau, CentraleSupélec, Université Paris-Saclay (France)
- Arpad Rimmel, CentraleSupélec, Université Paris-Saclay (France)
- Bich-Liên Doan, CentraleSupélec, Université Paris-Saclay (France)
14:50 Adversarial Robustness by Combining Prototype Models with Lipschitz Training
- Benjamin Paassen, Bielefeld University (Germany)
- Adia Khalid, Bielefeld University (Germany)
15:10 Beyond Performance: Comprehensive Evaluation Strategies for Impactful Machine Learning - poster spotlights
Organized by: Valerie Vaquet, Ulrike Kuhl, Saša Brdnik, Benjamin Paassen
15:10 Drift-Aware Evaluation of Fair Stream Learning
- Kathrin Lammers, Bielefeld University (DE)
- Fabian Hinder, Bielefeld University (Germany)
- Barbara Hammer, Cognitive Interaction Technology (CITEC), Bielefeld University (Germany)
- Valerie Vaquet, CITEC, Bielefeld University (Germany)
15:11 Real vs. Virtual Drift: Creating Realistic Stream Learning Benchmarks
- Fabian Hinder, Bielefeld University (Germany)
- Johannes Brinkrolf, CITEC - Cognitive Interaction Technology Bielefeld University (Germany)
- Kathrin Lammers, Bielefeld University (DE)
- Barbara Hammer, CITEC - Bielefeld University (Germany)
15:12 Linearity of Sensitive Concepts in Language Models
- Sarah Schröder, Bielefeld University (Germany)
- Valerie Vaquet, CITEC, Bielefeld University (Germany)
- Barbara Hammer, CITEC - Bielefeld University (Germany)
15:13 Evaluation of Rashomon Sets for the Determination of Stable and Plausible Model Explanations
- Marika Kaden, University of Applied Sciences Mittweida, Saxon Institute for Computational Intelligence and Machine Learning (Germany)
- Mahrokh Karimi, University of Applied Sciences Mittweida, Saxon Institute for Computational Intelligence and Machine Learning (Germany)
- Subhashree Panda, Hochschule Mittweida (Germany)
- Thomas Pfaff, University of Applied Sciences Mittweida, Saxon Institute for Computational Intelligence and Machine Learning (Germany)
- Thomas Villmann, Mittweida University of Applied Sciences, Saxon Institute for Computational Intelligence and Machine Learning (Germany)
15:14 Towards meaningful evaluation of uncertainty-aware segmentation workflows for medical applications
- Dany Rimez, UCLouvain (Belgium)
- John Lee, UCLouvain (Belgium)
- Ana Maria Barragan Montero, UCLouvain (Belgium)
15:15 Time series, online learning and domain adaptation
15:15 A Deep Learning Diagnostic Observer for Time Series Anomaly Detection
- Assmaa ALSAMADI, Litis Lab, Université de Rouen (France)
- Fannia Pacheco, LITIS Lab, Université Rouen Normandie (France)
- Paul Honeine, LITIS Lab, Université de Rouen Normandie (France)
15:35 Time Series Forecasting in the Presence of Explosive Bubbles
- Julien Peignon, CEREMADE, CNRS, UMR 7534, Université Paris-Dauphine, PSL University (France)
- Fabrice Rossi, CEREMADE - Université Paris Dauphine PSL (France)
- Arthur Thomas, LEDa, CNRS, UMR 8007, Université Paris-Dauphine, PSL University, Paris, France (France)
15:55 On the Components That Enable Robust Generalization in HAR Models
- Otávio Napoli, UNICAMP (Brazil)
- Edson Borin, UNICAMP - Institute of Computing (Brazil)
16:15 Drift Localization using Conformal Predictions
- Fabian Hinder, Bielefeld University (Germany)
- Valerie Vaquet, CITEC, Bielefeld University (Germany)
- Johannes Brinkrolf, CITEC - Cognitive Interaction Technology Bielefeld University (Germany)
- Barbara Hammer, CITEC - Bielefeld University (Germany)
16:35 Time series, online learning and domain adaptation - poster spotlights
16:35 Stochastic Parroting in Temporal Attention - Regulating the Diagonal Sink
- Victoria Hankemeier, University of Münster - Faculty of Mathematics and Computer Science Autonomous Intelligent Systems Group (Germany)
- Malte Schilling, University of Münster, Faculty of Mathematics and Computer Science, Autonomous Intelligent Systems Group (Germany)
16:36 Exploring the Relationship Between Synaptic Dynamics Properties: Gain-Control and Temporal Filtering
- Ferney Beltran-Velandia, Leipzig University (Germany)
- Nico Scherf, Max-Planck Institute for Human Cognitive and Brain Sciences (Germany)
- Martin Bogdan, Leipzig University (Germany)
16:37 On the Importance of Time Constants in Spiking Neural Networks
- Filippa Brandt, Lund University (Sweden)
- Saeed Bastani, Ericsson (Sweden)
- Alexander Hunt, Ericsson (Sweden)
- Amir Aminifar, Lund University (Sweden)
- Baktash Behmanesh, Lund University (Sweden)
16:38 Enriching Graph Topology Representations with Line Graph Transformations
- Paolo Frazzetto, Universitat de València (SPAIN)
- Luca Pasa, University of Padova (Italy)
- Nicolò Navarin, University of Padua (Italy)
- Alessandro Sperduti, University of Padua (Italy)
16:39 Optimal Training of the Online Newton Algorithm based on Conformal CUSUM
- Thomas Grava, UCLouvain - LIDAM/Louvain Finance research center (LFIN) (Belgium)
- Frédéric Vrins, UCLouvain - LIDAM/Louvain Finance research center (LFIN) (Belgium)
16:40 Lightweight Personalisation for MEMS-Based Wearables: A Padel Stroke Recognition Case Study
- Alberto Gascón Roche, Aragon Institute of Engineering Research, University of Zaragoza (Spain)
- Fatemeh Akbarian, Lund University (Sweden)
- Amir Aminifar, Lund University (Sweden)
- Alvaro Marco, Aragon Institute of Engineering Research, University of Zaragoza (Spain)
- Roberto Casas, Aragon Institute of Engineering Research, University of Zaragoza (Spain)
16:41 General Continual Unsupervised Learning with Augmented Vector Quantization
- Hervé Frezza-Buet, CentraleSupélec (France)
- Nasr Allah Aghelias, Université de Lorraine (France)
- Bernard Girau, Université de Lorraine (France)
16:43 Natural Language Processing - poster spotlights
16:43 On the Rank Properties of the Renormalization Trick in GCNs
- Anna Bison, Department of Mathematics, University of Padova (Italy)
- Alessandro Sperduti, University of Padua (Italy)
16:44 Emotion Recognition in Multimodal Social Data
- Lucia Passaro, University of Pisa (Italy)
- Davide Amadei, University of Pisa (Italy)
- Davide Bacciu, Università di Pisa (Italy)
16:45 Alignment of Islamic Legal Texts
- HAZIM BAROUDI, CEREMADE, CNRS, UMR 7534, Université Paris-Dauphine - PSL (FRANCE)
- Wassim Ammar, IRHT-CNRS, Campus Condorcet Paris (France)
- Farid Bouchiba, IRHT-CNRS, Campus Condorcet Paris (France)
- Shadha Karoumi Marmardji, IRHT-CNRS, Campus Condorcet Paris (France)
- Christian Mueller, IRHT-CNRS, Campus Condorcet Paris (France)
- Fabrice Rossi, CEREMADE - Université Paris Dauphine PSL (France)
16:46 Network analysis of conferences: Mapping the backbone of ESANN topics
- Arya Nair, University of Birmingham (UK)
- Jens Christian Claussen, University of Birmingham (UK)
16:47 Four Quadrants of Difficulty: A Categorisation for Curriculum Learning in NLP
- Vanessa Toborek, Rheinische Friedrich-Wilhelms-Universität Bonn, Lamarr Institute (Germany)
- Sebastian Müller, Rheinische Friedrich-Wilhelms-Universität Bonn, Lamarr Institute (Germany)
- Christian Bauckhage, Fraunhofer IAIS & Fraunhofer Center for Machine Learning (Deutschland)
16:48 Lost in Modality: Evaluating the Effectiveness of Text-Based Membership Inference Attacks on Large Multimodal Models
- ZIYI TONG, Japan Advanced Institute of Science and Technology (Japan)
- FEIFEI SUN, Japan Advanced Institute Of Science And Technology (Japan)
- Le Minh Nguyen, Japan Advanced Institute Of Science And Technology (Japan)
16:49 Task-Specific Efficiency Analysis: When Small Language Models Outperform Large Language Models
- Jinghan Cao, San Francisco State University (USA)
- Yu Ma, Carnegie Mellon University (USA)
- Xinjin Li, Columbia University (USA)
- Qingyang Ren, Cornell University (USA)
- Xiangyun Chen, Pennsylvania State University (USA)
16:50 AsymPuzl: A minimal puzzle testbed for LLM-based two agent communication with information asymmetry
- Xavier Cadet, Dartmouth College (USA)
- Edward Koh, Dartmouth College (USA)
- Peter Chin, Dartmouth College (US)
16:51 Model Sees but Does Not Learn: Eliminating Error Propagation in Reasoning Distillation
- jaeeun jang, Hanwha Systems (Republic of Korea)
- Hansle Lee, Hanwha Systems (Republic of Korea)
- wonjun cho, Hanwha Systems (Republic of Korea)
- sangmin kim, Hanwha Systems (Republic of Korea)
16:52 Poster exhibition and coffee break
18:15 Walking tour of Bruges with guides
Thursday, 23.04.2026
09:00 Efficient and Resilient Machine Learning for Industrial Applications
Organized by: Philipp Wissmann, Marc Weber, Simon Leszek, Philip Naumann, Daniel Hein, Steffen Udluft, Thomas Runkler
09:00 Efficient and Resilient Machine Learning for Industrial Applications
- Philipp Wissmann, Siemens AG (Deutschland)
- Philip Naumann, TU Berlin (Germany)
- Daniel Hein, Siemens AG (Germany)
- Steffen Udluft, Siemens Technology (Germany)
- Marc Weber, Siemens AG (Germany)
- Simon Leszek, TU Berlin (Germany)
- Thomas Runkler, Siemens AG/ TU Munich (Germany)
09:20 Variational Deep Embedding for Unsupervised Clustering of Industrial Noise in Steelmaking Plants
- Muhammad Waseem Akram, Scuola Superiore Sant'Anna, TeCIP Institute (Italy)
- Marco Vannucci, TeCIP institute, Scuola Superiore Sant'Anna (Italy)
- Giorgio Carlo Buttazzo, Scuola Superiore Sant'Anna, TeCIP Institute (Italy)
- Valentina Colla, Scuola Superiore Sant'Anna, TeCIP Institute (Italy)
- Stefano Dettori, Scuola Superiore Sant'Anna (Italia)
- Donatella Salvatore, Scuola Superiore Sant'Anna, TeCIP Institute (Italia)
09:40 TSFM in-context learning for time-series classification of bearing-health status
- Michel Tokic, Siemens AG / LMU Munich (Germany)
- Slobodan Djukanovic, Siemens AG / University of Montenegro (Germany)
- Anja von Beuningen, Siemens AG / Corporate Technology (Germany)
- Cheng Feng, Siemens AG (Germany)
10:00 Label-Efficient and Adaptable Image Selection for Large-Scale E-Commerce Catalogs
- Maytal Messing, eBay (Israel)
- Guy Shani, eBay (Israel)
10:20 Context-Aware Graph Attention for Unsupervised Telco Anomaly Detection
- Sara Malacarne, Telenor ASA - Research and Innovation (Norway)
- Eirik Hoel-Høiseth, Telenor ASA - Research and Innovartion (Norway)
- David Zsolt Biro, Telenor Denmark (Denmark)
- Erlend Aune, Hance / NTNU (Norway)
- Massimiliano Ruocco, NTNU (Norway)
10:40 Efficient and Resilient Machine Learning for Industrial Applications - poster spotlights
Organized by: Philipp Wissmann, Marc Weber, Simon Leszek, Philip Naumann, Daniel Hein, Steffen Udluft, Thomas Runkler
10:40 A comparison of open time-series foundation models for industrial manufacturing applications
- Can Calisir, Technical University of Berlin (Germany)
- Simon Leszek, TU Berlin (Germany)
10:41 Coffee break
11:00 Reliability, Safety and Robustness of AI applications
Organized by: Caroline König, Cecilio Angulo, Pedro Jesús Copado, G. Kumar Venayagamoorthy
11:00 Techniques for Reliable, Safe and Robust AI Applications
- Caroline König, Universitat Politècnica de Catalunya (Spain)
- Cecilio Angulo, Universitat Politècnica de Catalunya (Spain)
- Pedro Jesús Copado, Universitat Politècnica de Catalunya (Spain)
- G. Kumar Venayagamoorthy, Clemson University (USA)
11:20 Out-of-Distribution Segmentation via Wasserstein-Based Evidential Uncertainty
- Arnold Brosch, Heinrich-Heine-Universität, HHU (Germany)
- Abdelrahman Eldesokey, King Abdullah University of Science and Technology (Saudi Arabia)
- Michael Felsberg, Linköping University (Sweden)
- Kira Maag, Heinrich-Heine-Universität (Germany)
11:40 Weakly Supervised Shortcut Learning Mitigation Using Sparse Autoencoders
- sari sadiya, goethe universit frankfurt (Germany)
- Muhammad Ahsan, Goethe University Frankfurt (Germany)
- Despina Tawadros, Goethe University Frankfurt (Germany)
- Phuong Quynh Le, University of Marburg (Germany)
- Jörg Schlötterer, Marburg University (Germany)
- Christin Seifert, Marburg University (Germany)
- Gemma Roig, Goethe university Frankfurt (Germany)
12:00 Neuromodulated Delta Adapters: Stabilizing Test-Time Adaptation via Gated Error Correction
- Jiarui Zhang, Institute of Information Engineering, Chinese Academy of Sciences，Bejing，China (China)
- Deng Yifan, Institute of Information Engineering, Chinese Academy of Sciences，Bejing，China (China)
12:20 Reliability, Safety and Robustness of AI applications - poster spotlights
Organized by: Caroline König, Cecilio Angulo, Pedro Jesús Copado, G. Kumar Venayagamoorthy
12:20 Revisiting Neural Activation Coverage for Uncertainty Estimation
- Benedikt Franke, German Aerospace Center (DLR) (Germany)
- Nils Förster, German Aerospace Center (DLR) (Germany)
- Frank Köster, Deutsches Zentrum für Luft- und Raumfahrt e.V. (DLR) - Institut für KI-Sicherheit (Germany)
- Markus Lange, Institute for AI-Safety and Security German Aerospace Center (DLR) (Germany)
- Arne Raulf, German Aerospace Center (DLR) (Germany)
- Asja Fischer, Ruhr University Bochum (Germany)
12:21 Model Selection Hijacking Adversarial Attack
- Luca Pajola, SpritzMatter Srl. (Italy)
- Riccardo Petrucci, University of Padova (Italy)
- Francesco Marchiori, University of Padova (Italy)
- Luca Pasa, University of Padova (Italy)
- Mauro Conti, University of Padova (Italy)
12:22 The Alignment Gate: Intent and Instruction Guardrails for Agentic AI
- Akash Borigi, Missouri University of Science and Technology (USA)
- Peggy Lindner, Missouri University of Science and Technology (USA)
- Alexander Schlager, AIceberg Inc (USA)
- Saifullah Shoaib, Missouri University of Science and Technology (USA)
- Rupendra Lekkala, Missouri University of Science and Technology (USA)
- Sai Sowjanya Bhamidipati, AIceberg Inc (USA)
- Amaury Lendasse, Missouri University of Science and Technology (USA)
12:23 Interpreting Logical Explanations of Classifying Neural Networks
- Fabrizio Leopardi, University of Lugano (Switzerland)
- Faezeh Labbaf, Università della Svizzera Italiana (University of Lugano) (Switzerland)
- Tomas Kolarik, University of Lugano (Switzerland)
- Michael Wand, SUPSI-IDSIA (Switzerland)
- Natasha Sharygina, University of Lugano (Switzerland)
12:24 Adversarial Confusion Attack: Disrupting Multimodal Large Language Models
- Jakub Hoscilowicz, Warsaw University of Technology (Poland)
- Artur Janicki, Warsaw University of Technology (Poland)
12:25 SPARC: Superpixel-based Black-Box Adversarial Attack with Regional Confidence
- Tram Ho, Faculty of Information Technology - University of Science, VNUHCM (Viet Nam)
- Ngoc-Thao Nguyen, Faculty of Information Technology, University of Science, Ho Chi Minh City, Vietnam Vietnam National University, Ho Chi Minh City, Vietnam (Vietnam)
- Bac Le, Faculty of Information Technology, University of Science, Ho Chi Minh City, Vietnam Vietnam National University, Ho Chi Minh City, Vietnam (Vietnam)
12:26 On the Impact of Differential Privacy on Federated Neuromorphic Learning Accuracy
- Luiz Pereira, Federal University of Campina Grande (Brazil)
- Dalton Valadares, Federal University of Campina Grande (Brazil)
- Mirko Perkusich, Federal University of Campina Grande (Brazil)
- Kyller Gorgônio, Federal University of Campina Grande (Brazil)
12:27 Classification and regression - poster spotlights
12:27 Implementation of Multi-Matrix Median Generalized Learning Vector Quantization
- Lukas Bader, August Rüggeberg GmbH / TH Köln (Germany)
- Ina Terwey-Scheulen, August Rüggeberg GmbH (Germany)
- Dietlind Zühlke, Technische Hochschule Köln (Germany)
12:28 Enhancing classification performance at the RAM-neuron level
- Antonio Sorgente, CNR (Italy)
- Gianluca Coda, CNR (Italy)
- Alessandro De Gregorio, Prisca Lab, DIETI, Unina (Italy)
- Massimo De Gregorio, CNR (Italy)
- Paolo Vanacore, AdR Napoli 1 (Italy)
12:29 Diminishing Returns - Data Integer Quantization and its Effects on Training Dynamics of Distance Based Classifiers
- Thomas Davies, University of Applied Sciences Mittweida, Saxon Institute for Computational Intelligence and Machine Learning (Germany)
- Alexander Engelsberger, University of Applied Sciences Mittweida, Saxon Institute for Computational Intelligence and Machine Learning (Deutschland)
- Magdalena Psenickova, University of Applied Sciences Mittweida, Saxon Institute for Computational Intelligence and Machine Learning (DE)
- Thomas Villmann, Mittweida University of Applied Sciences, Saxon Institute for Computational Intelligence and Machine Learning (Germany)
12:30 SMOTE k-out: Enhancing Class Separability through Outer Synthetic Sampling
- Verónica Bolón-Canedo, CITIC, Universidade da Coruña (Spain)
- José Luis Morillo-Salas, CITIC, Universidade da Coruña (Spain)
- Laura Morán-Fernández, CITIC, Universidade da Coruña (Spain)
- Amparo Alonso-Betanzos, CITIC, Universidade da Coruña (Spain)
12:31 Kernel Thinning for faster KSVM hyper-parametrization
- Blanca Cano-Camarero, Universidad Autónoma de Madrid (Spain)
- Ángela Fernández, Universidad Autónoma de Madrid (Spain)
- José R. Dorronsoro, Universidad Autónoma de Madrid (Spain)
12:32 Local Concept Embeddings in the Context of Self-Supervised Learning
- Kim Paulke, Helmut Schmidt University (Germany)
- Hans-Oliver Hansen, University of Lübeck, Institut for Neuro- and Bioinformatics (Germany)
- Thomas Martinetz, Institute or Neuro- and Bioinformatics, University of Lübeck (Germany)
- Gesina Schwalbe, University of Lübeck, Institute for Software Engineering and Programming Languages (Germany)
12:33 THDC: Training Hyperdimensional Computing Models with Backpropagation
- Hanne Dejonghe, Ghent University (Belgium)
- Sam Leroux, Ghent University (Belgium)
12:34 AdaCap: An Adaptive Contrastive Approach for Small-Data Neural Networks
- Bruno Belucci, Université Paris Dauphine-PSL (France)
- Karim Lounici, Ecole Polytechnique (France)
- Katia Meziani, Université Paris Dauphine-PSL (France)
12:35 Clagging: Generating and combining predictions using clustering
- Arnaud Germain, UCLouvain - LIDAM/Louvain Finance research center (LFIN) (Belgium)
- Frédéric Vrins, UCLouvain - LIDAM/Louvain Finance research center (LFIN) (Belgium)
12:36 mDAE: modified Denoising AutoEncoder for missing data imputation
- Mariette Dupuy, Inria Montpellier (FRANCE)
- Marie Chavent, INRIA Bordeaux (FRANCE)
- Rémi Dubois, IHU LIRYC
12:37 Lunch
14:10 Learning and Reasoning on Knowledge and Heterogeneous Graphs
Organized by: Matteo Zignani, Pasquale Minervini, Roberto Interdonato, Manuel Dileo
14:10 Learning and Reasoning on Knowledge and Heterogeneous Graphs in the era of Graph Foundation and Large Language Models
- Matteo Zignani, University of Milan (Italy)
- Pasquale Minervini, University of Edinburgh; Miniml.AI (United Kingdom)
- Roberto Interdonato, CIRAD (France)
- Manuel Dileo, Human Technopole Foundation (Italy)
14:30 A Possible Human-Centered Embedding Space Search in Degenerate Clifford Algebras
- Isaac Roberts, Bielefeld University - Faculty of Technology (Germany)
- Louis Mozart Kamdem teyou, Heinz Nixdorf Institute-Paderborn University (Germany)
- Alexander Schulz, Bielefeld University (Germany)
- N'Dah Jean Kouagou, Paderborn University (Germany)
- Axel Ngonga Ngomo, University of Paderborn (Germany)
- Barbara Hammer, CITEC - Bielefeld University (Germany)
14:50 Graph Representation Learning for Software Architecture Recovery
- Rakhshanda Jabeen, Linnaeus University (Sweden)
- Morgan Ericsson, Linnaeus University (Sweden)
- Jonas Nordqvist, Linnaeus University (Sweden)
- Anna Wingkvist, Linnaeus University (Sweden)
15:10 Learning and Reasoning on Knowledge and Heterogeneous Graphs - poster spotlights
Organized by: Matteo Zignani, Pasquale Minervini, Roberto Interdonato, Manuel Dileo
15:10 A Multi-Agent LLM System for Natural Language Querying of Operational Knowledge Graphs in Satellite Ground Stations
- Fosco Eugenio Quadri, Telespazio (Italy)
- Filippo Bianchini, Telespazio (Italy)
15:11 GraphTreeBoost: Soft Decision Tree-Based Graph Learning With Spectral Aggregation
- László Fetter, Budapest University of Technology and Economics (Hungary)
- András Gézsi, Budapest University of Technology and Economics (Hungary)
15:12 Vision, image processing and healthcare AI
15:12 See Without Decoding: Motion-Vector-Based Tracking in Compressed Video
- Axel Duché, LITIS : (Computer Science, Information Processing and Systems Laboratory) (France)
- Clement Chatelain , LITIS Lab, INSA Rouen (France)
- Gilles Gasso, LITIS Lab, INSA Rouen (France)
15:32 But Are These Images Conceptually Similar?
- Isaac Roberts, Bielefeld University - Faculty of Technology (Germany)
- Riza Velioglu, Bielefeld University - Faculty of Technology (Germany)
- Inaam Ashraf, Bielefeld University - Faculty of Technology (Germany)
- Luca Hermes, Bielefeld University CoR-Lab (CITEC) (Germany)
- Barbara Hammer, CITEC - Bielefeld University (Germany)
15:52 Predictive Coding inspired convolutional networks can capture the neural dynamics of recurrent processing in human image recognition
- Manshan Guo, Goethe universität Frankfurt, Freie universität Berlin (Germany)
- Bhavin Choksi, Goëthe Universität Frankfurt (Germany)
- sari sadiya, goethe universit frankfurt (Germany)
- Pablo Oyarzo, Freie universität Berlin (Germany)
- Radoslaw Cichy, Freie universität Berlin (Germany)
- Gemma Roig, Goethe university Frankfurt (Germany)
16:12 Hierarchical Multi-Scale Deep Neural Network for Schizophrenia Detection in Neuroimaging
- Carlos Dias Maia, PUC Minas (Brazil)
- Gabriel Fonseca, Pontifical Catholic University of Minas Gerais (PUC-MG) (Brasil)
- Silvio Jamil, Puc Minas (Brazil)
- Luis Zárate, Pontifical Catholic University of Minas Gerais (PUC-MG) (Brazil)
16:32 Vision, image processing and healthcare AI - poster spotlights
16:32 Ensembling Post-Hoc Image Explanations: When It Works, When It Fails, and How to Tell the Difference
- Luca Oneto, University of Genoa (Italy)
- jinhua xu, University of Genoa (Italy)
- Davide Anguita, DIBRIS - University of Genova (Italy)
- Fabio Roli, University of Genoa (Italy)
- Jing Yuan, Zhejiang Normal University (China)
16:33 When Curvature Counts: Hyperbolic Geometry in Prototype-Based Image Classification
- Silvia Grosso, INSA Lyon - LIRIS, Lyon, France (France)
- Samuele Fonio, Università degli studi di Torino (Italy)
- Mirko Polato, University of Torino (Italy)
- Roberto Esposito, Università degli studi di Torino (Italy)
- Sara Bouchenak, INSA Lyon - LIRIS, Lyon, France (France)
16:34 Movements as Images: CNNs are Good Feature Extractors in Sign Language Recognition
- Pierre Poitier, University of Namur (Belgium)
- Loïc Brangier, University of Namur (Belgium)
- Ariel Basso Madjoukeng, University of Namur (Belgium)
- Benoit Frénay, Université de Namur (Belgium)
16:35 Fast denoising of low-count Monte Carlo proton therapy dose distributions with ResUNet
- Pierre Merveille, UCLouvain (Belgique)
- Ana Maria Barragan Montero, UCLouvain (Belgium)
- Kevin Souris, Ion Beam Applications (Belgium)
- John Lee, UCLouvain (Belgium)
16:36 Non-Linear Activation Functions for Deep Riemannian Neural Networks
- Lucas H. dos Santos, Federal University of ABC (Brazil)
- Joao Barbon, State University of Campinas, Department of Computer Engineering and Industrial Automation. (Brazil)
- Sylvain Chevallier, LISN - Université Paris-Saclay (France)
- Denis G. Fantinato, Universidade Estadual de Campinas (Brazil)
16:37 XAI-Enabled Custom CNN for Cross-Modal Generalization in Breast Cancer Detection
- Maram Issaoui, University of Ibn Khaldoun (UIK), (Tunisia)
- Amal Jlassi, University of Ibn Khaldoun (UIK) (Tunisia)
- Abir Baâzaoui, Higher Institute of computer science and management of Kairouan (Tunisia)
- Walid Barhoumi, Université de Carthage, Ecole Nationale d’Ingénieurs de Carthage, (Tunisia)
16:38 Towards Learning a Generalizable 3D Scene Representation from 2D Observations
- Martin Gromniak, Universität Hamburg Department of Informatics (Germany)
- Jan-Gerrit Habekost, Universität Hamburg Department of Informatics (Germany)
- Sebastian Kamp, Universität Hamburg Department of Informatics (Germany)
- Sven Magg, Hamburger Informatik Technologie-Center e.V. (Germany)
- Stefan Wermter, University of Hamburg, Dept. of Informatics (Germany)
16:39 Poster exhibition and coffee break
18:15 Visit of the "Halve maan" brewery
19:15 Conference dinner at the "Halve Maan" brewery
Friday, 24.04.2026
09:00 Recurrent and reinforcement learning
09:00 Random Unicycle Network (RUN!): supercharging harmonic oscillator networks via non-holonomic constraints
- Mariano Ramirez, Delft University of Technology (The Netherlands)
- Andrea Ceni, University of Pisa (Italy)
- Andrea Cossu, University of Pisa (Italy)
- Davide Bacciu, Università di Pisa (Italy)
- Claudio Gallicchio, University of Pisa (Italy)
- Cosimo Della Santina, Delft University of Technology (The Netherlands)
09:20 Memristive-Friendly Hadamard Reservoirs
- Andrea Ceni, University of Pisa (Italy)
- Gianluca Milano, Istituto nazionale di ricerca metrologica (INRiM) (Italy)
- Carlo Ricciardi, Politecnico di Torino (Italy)
- Claudio Gallicchio, University of Pisa (Italy)
09:40 Unpacking the Role of Intrinsic Motivation in Elastic Decision Transformers: A Post-Hoc Analysis of Embedding Geometry and Performance
- Leonardo Guiducci, DISPOC, Unisi (Italy)
- Antonio Rizzo, DISPOC, Unisi (Italy)
- Giovanna Maria Dimitri, University of Milan (Statale) (italy)
10:00 Recurrent and reinforcement learning - poster spotlight
10:00 Physics-Informed Recurrent Architecture with Embedded Thermodynamic Dynamics for Robust Sequence Modeling
- Zafer Yigit, Mälardalen University, Volvo Construction Equipment (Sweden)
- Håkan Forsberg, Mälardalen University (Sweden)
- Masoud Daneshtalab, Mälardalen University (Sweden)
10:01 Constraint Guided Recurrent Convolutional AutoEncoders for Condition Indicator Estimation
- Maarten Meire, KULeuven (Belgium)
- Quinten Van Baelen, KU Leuven (Belgium)
- Ted Ooijevaar, Flanders Make (Belgium)
- Peter Karsmakers, KU Leuven Campus Geel (Belgium)
10:02 Mixed-Density Diffuser: Efficient Planning with Non-Uniform Temporal Resolution
- Crimson Stambaugh, University of Washington (United States of America)
- Rajesh Rao, University of Washington (United States)
10:03 Integrating Potential-Based Reward Shaping into AlphaZero
- Koen Boeckx, Royal Military Academy (Belgian Defense) (Belgium)
- Xavier Neyt, Royal Military Academy (Belgian Defense) (Belgium)
10:04 Hybrid-AIRL: Enhancing Inverse Reinforcement Learning with Supervised Expert Guidance
- Bram Silue, Vrije Universiteit Brussel (Belgium)
- Santiago Amaya-Corredor, Vrije Universiteit Brussel (Belgium)
- Patrick Mannion, University of Galway (Ireland)
- Lander Willem, University of Antwerp (Belgium)
- Pieter Libin, Vrije Universiteit Brussel (Belgium)
10:05 Deobfuscation as a GNN-Based Graph-Edit Problem by Reinforcement Learning
- Roxane Cohen, Quarkslab & LAMSADE, Université Paris-Dauphine - PSL (France)
- Robin David, Quarkslab (France)
- Samuel Hangouët, Quarkslab (France)
- Florian Yger, LITIS - INSA Rouen Normandie (France)
- Fabrice Rossi, CEREMADE - Université Paris Dauphine PSL (France)
10:06 Point-wise Q-value maximization for converging Q-learning in continuous state-spaces
- Philipp Wissmann, Siemens AG (Deutschland)
- Daniel Hein, Siemens AG (Germany)
- Steffen Udluft, Siemens Technology (Germany)
- Thomas Runkler, Siemens AG/ TU Munich (Germany)
10:07 Coffee break
10:30 DImension reduction, feature selection and unsupervised learning
10:30 LMAP: Local PCA Models with Global MDS Embeddings
- Oliver Kramer, University of Oldenburg (Germany)
10:50 SPDNet-AE: a Compact SPD Representation through Riemannian Autoencoding
- Thibault de Surrel, LAMSADE, Univ. Paris Dauphine, CNRS (France)
- Charlotte Boucherie, LITIS, Insa Rouen (France)
- Florian Yger, LITIS - INSA Rouen Normandie (France)
11:10 Multi-Scale Stochastic Neighbor Embedding with Twice Adaptive Bandwidths
- John Lee, UCLouvain (Belgium)
- Pierre Lambert, UCLouvain (Belgium)
- Edouard Couplet, UCLouvain (Belgique)
- Pierre Merveille, UCLouvain (Belgique)
- Dounia Mulders, UCLouvain (Belgium)
- Cyril de Bodt, UNamur - naXys (Belgium)
- Michel Verleysen, UCLouvain - ICTEAM institute (Belgium)
11:30 Interpretable Parametric Neighbour Embedding
- Edouard Couplet, UCLouvain (Belgique)
- Pierre Lambert, UCLouvain (Belgium)
- Michel Verleysen, UCLouvain - ICTEAM institute (Belgium)
- John Lee, UCLouvain (Belgium)
- Cyril de Bodt, UNamur - naXys (Belgium)
11:50 Enforcing Feature Sparseness for Reliable Classification by Prototype-Based Models
- Marika Kaden, University of Applied Sciences Mittweida, Saxon Institute for Computational Intelligence and Machine Learning (Germany)
- Julius Voigt, University of Applied Sciences Mittweida, Saxon Institute for Computational Intelligence and Machine Learning (SICIM) (Germany)
- Sascha Saralajew, NEC Laboratories Europe Heidelberg (Germany)
- Thomas Villmann, Mittweida University of Applied Sciences, Saxon Institute for Computational Intelligence and Machine Learning (Germany)
12:10 Autoencoders versus PCA for feature extraction in FDG PET scans in neurodegenerative diseases
- Roland Veen, University of Groningen, Bernoulli Institute for Mathematics, Computer Science and Artificial Intelligence (The Netherlands)
- Sofie Lövdal, University Medical Center Groningen (UMCG), Department of Nuclear Medicine and Molecular Imaging Hanzeplein 1, 9713 GZ Groningen -The Netherlands (The Netherlands)
- Kaitlin Vos, University of Groningen (Netherlands)
- Ciro Setolino, University of Salerno (Italia)
- Sanne Meles, University Medical Center Groningen (Netherlands)
- Michael Biehl, University of Groningen, Bernoulli Institute for Mathematics, Computer Science and Artificial Intelligence (The Netherlands)
12:30 DImension reduction, feature selection and unsupervised learning - poster spotlights
12:30 Information-Theoretic Unsupervised Feature Selection for High-Dimensional Spatial Data
- Samuel Suárez-Marcote, CITIC, Universidade da Coruña (Spain)
- Abhijeet Vishwasrao, University of Michigan (USA)
- Ricardo Vinuesa, University of Michigan (USA)
- Laura Morán-Fernández, CITIC, Universidade da Coruña (Spain)
- Verónica Bolón-Canedo, CITIC, Universidade da Coruña (Spain)
12:31 Topology-Preserving Prototype Learning on Riemannian Manifolds
- Lucas Schwarz, University of Technology Chemnitz (Germany)
- Magdalena Psenickova, University of Applied Sciences Mittweida, Saxon Institute for Computational Intelligence and Machine Learning (DE)
- Thomas Villmann, Mittweida University of Applied Sciences, Saxon Institute for Computational Intelligence and Machine Learning (Germany)
- Florian Röhrbein, Chemnitz University of Technology (Germany)
12:32 Polarizing Kernels: A Definite Approach to Clustering with Indefinite Similarities
- Frank-Michael Schleif, Technical University of Applied Sciences Würzburg-Schweinfurt (Germany)
- Manuel Röder, Technical University of Applied Sciences Würzburg-Schweinfurt (Germany)
- Maximilian Münch, Technische Hochschule Würzburg-Schweinfurt, Center for Artificial Intelligence and Robotics (Germany)
- Peter Preinesberger, Technische Hochschule Wuerzburg-Schweinfurt, TTZ-WUE Cyber Security (Germany)
12:33 Effects of a Parametrized Neighborhood Family on the Quality of Self-Organizing Maps
- Cesar Cardenas, Rice University (United States)
- Erzsébet Merényi, Rice University (USA)
12:34 Lunch
14:05 Graph learning
14:05 Assessing Graph Neural Networks for latency and power consumption prediction in application mappings on multicore architectures
- Oscar Roussel, LS2N / UMR CNRS 6004 and IETR / UMR CNRS 6164 (France)
- Zainab Ghrayeb, IETR / UMR CNRS 6164 (France)
- Sébastien Le Nours, IETR / UMR CNRS 6164 (France)
- Christine Sinoquet, Nantes Digital Science Laboratory (LS2N / UMR CNRS 6004) Nantes University (France)
14:25 GNNs Don't Need Backprop
- Pascal Welke, Lancaster University Leipzig (Germany)
- Benoit Goupil, LIX (École Polytechnique, IP Paris, CNRS) (France)
- Fabian Jogl, TU Wien (Austria)
14:45 Graph learning - poster spotlights
14:45 Boundary-Constrained Diffusion Models for Floorplan Generation: Balancing Realism and Diversity
- Leonardo Stoppani, University of Pisa (Italy)
- Davide Bacciu, Università di Pisa (Italy)
- Shahab Mokarizadeh, H&M Hennes & Mauritz AB (Sweden)
14:46 Domination Reliability Analysis Based on Graph Features Using Generalized Matrix LVQ
- Mandy Lange-Geisler , University of Applied Sciences Mittweida (germany )
- Klaus Dohmen, University of Applied Sciences Mittweida (Germany)
- Thomas Villmann, Mittweida University of Applied Sciences, Saxon Institute for Computational Intelligence and Machine Learning (Germany)
14:47 Scaling up graph-based classifiers with a divide and conquer approach
- Caius Souza, Universidade Federal de Minas Gerais (Brazil)
- Rafael Lopes Almeida, University of Eastern Finland (Finland)
- Frederico Gualberto Ferreira Coelho, Universidade Federal de Minas Gerais (Brazil)
- Antonio Padua Braga, Universidade Federal de Minas Gerais (Brasil)
14:48 A Collaborative Distillation Framework for Graph Neural Networks
- Paul Agbaje, University of Texas at Arlington (United States)
- Arkajyoti Mitra, University of Texas at Arlington (United States)
- Afia Anjum, University of Texas at Arlington (United States)
- Pranali Khose, University of Texas at Arlington (United States)
- Ebelechukwu Nwafor, Villanova University (United States)
- Habeeb Olufowobi, University of Texas at Arlington (United States)
14:49 Deep models and learning principles
14:49 Where to grow: a surprisingly straightforward criterion to detect under-expressive layers
- Yifan WANG, INSA Centre Val de Loire (France)
- Julien MILLE, INSA Centre Val de Loire (France)
- Moncef HIDANE, INSA Centre Val de Loire (France)
15:09 Scalable Linearized Laplace Approximation via Surrogate Neural Kernel
- Luis A. Ortega, Aalborg University (Spain)
- Simón Rodríguez-Santana, Universidad Pontificia de Comillas (Spain)
- Daniel Hernández-Lobato, Unviersidad Autónoma de Madrid (Spain)
15:29 Deep models and learning principles - poster spotlights
15:29 Improving the Linearized Laplace Approximation via Quadratic Approximations
- Pedro Jiménez García-Ligero, Universidad Autónoma de Madrid (Spain)
- Luis A. Ortega, Aalborg University (Spain)
- Pablo Morales-Álvarez, University of Granada (Spain)
- Daniel Hernández-Lobato, Unviersidad Autónoma de Madrid (Spain)
15:30 Towards Understanding The Winner-Take-Most Behavior Of Neural Network Representations
- Gilles Peiffer, Université catholique de Louvain ICTEAM (Belgium)
- Christophe De Vleeschouwer, UCLouvain (Belgium)
- Simon Carbonnelle, UCLouvain (Belgium)
15:31 Energy-Based Dropout with Patch-Level Regularization
- Tom Devynck, Paris-Saclay University, UVSQ (France)
- Bilal FAYE, LIPN UMR CNRS 7030, Sorbonne Paris Nord University (USPN) (France)
- Djamel Djamel BOUCHAFFRA, Center for Development of Advanced Technologies (Algeria)
- Nadjib Lazaar, Université Paris Saclay (France)
- Hanane AZZAG, LIPN, University Paris 13 (France)
- Mustapha Lebbah, Université Paris-Saclay - UVSQ Versailles Campus (France)
15:32 Distillation of a tractable model from the VQ-VAE
- Armin Hadzic, Artificial Intelligence Center at Faculty of Electrical Engineering, Czech Technical University in Prague (Czech Republic)
- Milan Papež, Artificial Intelligence Center at Faculty of Electrical Engineering, Czech Technical University in Prague (Czech Republic)
- Tomáš Pevný, Artificial Intelligence Center at Faculty of Electrical Engineering, Czech Technical University in Prague (Czech Republic)
15:33 Cross-tested Aggregated Hold-out
- Joseph Rynkiewicz, SAMM, Université de Paris 1 (France)
15:34 LLM-based Vulnerable Code Augmentation: Generate or Refactor?
- Dyna Soumhane Ouchebara, University of Mons (UMONS) (Belgium)
- Stéphane Dupont, MAIA Artificial Intelligence Lab, Faculty of Science, University of Mons (Belgium)
15:35 Boosting the Lottery Ticket Hypothesis with Knowledge Distillation: Finding Sparser Winning Tickets
- Daan Luyckx, KU Leuven Campus Geel (Belgium)
- Peter Karsmakers, KU Leuven Campus Geel (Belgium)
15:36 Hub-Aware Hybrid Search: Accelerating the Locally Aligned Ant Technique
- Simone Vilardi, University of Groningen (The Netherlands)
- Reynier Peletier, University of Groningen (The Netherlands)
- Felipe Contreras, University of Groningen - Kapteyn Astronomical Institute (Colombia)
- Kerstin Bunte, University of Groningen (The Netherlands)
15:37 Linear Evaluation Complexity of Surrogate-Assisted (1+1)-EA on OneMax
- Oliver Kramer, University of Oldenburg (Germany)
15:38 Self-Certified Deep Metric Learning with N-Tuple Losses
- Oritsemisan Meggison, University of Birmingham (United Kingdom)
- Sijia Zhou, University of Birmingham (United Kingdom)
- Ata Kaban, University of Birmingham (United Kingdom)
15:39 Poster exhibition and coffee break
17:00 End of conference