The Workshop will take place in the SPS Department of the Università degli Studi di Milano, Sala Lauree, Via del Conservatorio 7, 20122, Milan (Italy). The department is extremely central,...
Read moreProgram and Social Event
ANNPR 2026 will include a three days program, with oral presentations, keynotes and social events. Check out below the complete program!
Check the page important dates for deadlines, and information on ANNPR 2026 conference dates
7 October
| Time | Session | Paper | Authors |
| 7/10 9:00 | Opening | ||
| Chair Session Main I: Stefano Montanelli | |||
| 7/10 9:30 | Main | Benchmarking LLMs for ASA Physical Status Classification from German Anesthesia Records | Daniel Nico Bigler, Andreas Vogt, Kaspar Riesen |
| 7/10 9:50 | Main | Effect of Multiple Cross-Validation Solutions on Generalization Abilities | Shigeo Abe |
| 7/10 10:10 | Main | Hyperbolic Graph Representation Learning for Differential Diagnosis on Biomedical Knowledge Graphs | Mauricio Soto-Gomez, Pietro Miotto, Lucia Mellini, Tommaso Marzi, Cesare Alippi, Elena Casiraghi, Alberto Paccanaro, Giorgio Valentini |
| 7/10 10:30 | Coffee Break | ||
| 7/10 11:00 | Keynote Talk | Diffusion models for image and video generation | Dr. Sander Dieleman, DeepMind (Google), London |
| Chair Session Main II: Sergio Picascia | |||
| 7/10 12:00 | Main | JusDef: Defeasible Message Passing for Exception-Aware Legal Document Classification | Awais Abdul Khaliq, Stefano Montanelli, Hercules Dalianis, Nuzhat Naqvi |
| 7/10 12:20 | Main | Estimating Shortcut Learning and Inductive Bias in Multi-View Behavioural Trace Classification | Marzio De Corato, Danilo Bruschi, Andrea Monzani |
| 7/10 12:40 | Main | Deep Inversion of Hybrid Dynamic Probabilistic Graphical Models | Edmondo Trentin, Raoof HojatJalali |
| 7/10 13:00 | Lunch | ||
| 7/10 14:00 | Keynote Talk | Hyperbolic and Hybrid GRL for ontology-based Biomedical KGs | Prof. Elena Casiraghi (University of Milan), Milan (Italy) |
| Chair Session Main III: Marzio De Corato | |||
| 7/10 15:00 | Main | Employing Discrete Fourier Transform in Representation Learning | Raoof HojatJalali, Edmondo Trentin |
| 7/10 15:20 | Main | Understanding What is Uncertain: Uncertainty Disentanglement in Stereo Depth Estimation | Sven van Loon, Ivo Pascal de Jong, Matias Valdenegro-Toro |
| 7/10 15:40 | Main | LEAP: Two-Phase Neural Architecture for Tabular Feature Prediction | Leonardo Guiducci, Antonio Rizzo |
| 16:00 | Coffee Break | ||
| Chair Session Main IV: Edmondo Trentin | |||
| 7/10 16:20 | Main | Counterfactual Explanations for Graph Neural Networks in Patient Outcome Prediction | Nikolaos Chaidos, Angeliki Dimitriou, Hari Calzi, Elena Casiraghi, Giorgos Stamou, Giorgio Valentini |
| 7/10 16:40 | Main | Evaluation of Vision-Language Models for Iris Verification Across Datasets and Representations | Ruggero Donida Labati, Alfio Ferrara, Sergio Picascia, Vincenzo Piuri, Elisabetta Rocchetti, Fabio Scotti |
| 7/10 17:00 | Main | Investigating Spurious Cue Sensitivity in Semantic Segmentation | Muhammad Ahsan, Sari Saba-Sadiya, Gemma Roig |
| 7/10 17:20 | Main | Lost in Translation? Large Language Models alter strategic behaviour when prompted in low-resource languages | Andrew Powell, Nikita Huber-Kralj, Grace Ibukunoluwa Ufeoshi, Daniele Proverbio, Alessio Buscemi, Chaimaa Tarzi, The Anh Han, Pietro Lio, Alessandro Di Stefano |
| 7/10 20:00 | Social Dinner at Ristorante il Cantinone (See below for details) |
8 October
| Time | Session | Paper | Authors |
| Chair Session Main V: Elisabetta Rocchetti | |||
| 8/10 9:00 | Main | Exploring Ethnic Bias in Facial Super Resolution via Uncertainty Quantification | Dhruvs Sharma, Marco Zullich, Matias Valdenegro-Toro |
| 8/10 9:20 | Main | Disentangling Model and Human Data Uncertainty in Apparent Facial Age Estimation | Andrei Foitos, Ivo Pascal de Jong, Matias Valdenegro-Toro |
| 8/10 9:40 | Main | LLM-Assisted Character Network Construction and Community Analysis in Late Ottoman Turkish Fiction | Beyza Altuner, Deniz Eren Terziler, Kaya Oğuz, Miriam Walsh |
| 8/10 10:00 | Main | Anisotropic Hyper-Grid Coresets for Resource-Constrained Retraining of Visual Defect Classifiers | Andrea Calvagna, Andrea Orazio Caruso, Francesco Rundo, Sebastiano Battiato, Emiliano Tramontana |
| 8/10 10:20 | Main | A Layer-Wise Mechanistic Analysis of Privacy in Vision Transformer Pose Estimators | Matteo Mugnai, Francesco Pistolesi |
| 8/10 10:40 | Coffee Break | ||
| 8/10 11:00 | Keynote Talk | Designing biological molecules with generative AI: from RNA structure to synthesisable drugs | Prof Pietro Lió, University of Cambridge, Cambridge (UK) |
| Chair Session Main VI: Mauricio Soto Gomez | |||
| 8/10 12:00 | Main | Automated component testing for agentic AI in medical decision support systems | Waqar Ali, Benjamin Meyer, Linus Stuhlmann, Schlomo Aschkenasy, Paola Daniore, Thilo Stadelmann, Ahmed Abdulkadir |
| 8/10 12:20 | Main | Benchmarking Human Pose Estimators on Anonymized Part-Based Body Representations: Accuracy and Adaptation | Francesco Pio Crispino, Francesco Pistolesi |
| 8/10 12:40 | Main | Language Model Architecture Shapes Brain Alignment and Neuroscientific Inference | Anthea Hohmann, Sari Sadiya, Gemma Roig |
| 8/10 13:00 | Lunch | ||
| Chair Session Main VII: Giovanna Maria Dimitri | |||
| 8/10 14:00 | Main | HexaWorld: An Open Framework for Comparing Grid Geometries in Constrained RL Exploration | Giulia Lombardi, Luigi Amedeo Bianchi, Rochelle Schneider |
| 8/10 14:20 | Main | Rerankers in the SHADOW: What Dense Retrieval Leaves Behind | Tobias Steiner, Merlin Streilein, Kaspar Riesen, Andreas Fischer |
| Special Session Archaeology: Chairs Nevio Dubbini and Marek Bundzel | |||
| 8/10 14:40 | Archaeology | LandTalk.AI: Democratizing Archaeological Remote Sensing through Vision-Enabled Large Language Models | Jürgen Landauer, Sara Kassehl |
| 8/10 15:00 | Archaeology | Similarity-Based Retrieval of Maya Glyph Signs from Epigraphic Catalogues Using DINOv2 Embeddings | Marek Bundzel |
| 8/10 15:20 | Archaeology | Explainable Multimodal AI for Adaptive Calibration of Archaeological Sensing Workflows | Nevio Dubbini, Daniël P. van Helden, Claudia Sciuto, Martina Naso, Arthur Leck, Clément Joubert, Heeli C. Schechter, Rémy Chapoulie, Gabriele Gattiglia |
| 8/10 15:40 | Archaeology | Multimodal fusion of visual and morphometric features for avian bone classification | Nevio Dubbini, Lisa Yeomans, Marco Pavia, Ramazan Parmaksız, Ayse Atas Hooglugt, Gabriele Gattiglia, Beatrice Demarchi |
| 8/10 16:00 | Archaeology | A Pipeline for Large-Scale Archaeological Site Detection: Integrating Remote Sensing Data with Historical Maps | Richard Albrecht |
| 8/10 16:20 | Archaeology | Restoring Severely Fragmented Roman Frescoes with Expert-Guided Diffusion Models | Andrea Munarin, Sinem Aslan |
| 8/10 16:40 | Archaeology | Classifying Decorated Pottery by Reshaping Embedding Geometry | Federica Mauro, Franco Cicirelli, Ettore Ritacco, Gabriele Gattiglia |
| 8/10 17:00 | Coffee Break | ||
| Position Papers Session: Chair Francesco Pistolesi | |||
| 8/10 17:20 | Position Paper | From Patch-Centric to Cell-Centric Vision-to-Omics: A Position Paper on Molecular Prediction from Histopathology | Arianna Castellini, Andrea Tangherloni, Francesca Buffa |
| 8/10 17:40 | Position Paper | How Streaming LDA Questions our Progress in Continual Learning with Pretrained Models | Giovanni Donghi, Luca Pasa, Alessandro Sperduti, Nicolò Navarin |
| 8/10 18:00 | Position Paper | The Missing Link Between Space Law and Artificial Intelligence: A Position Paper on Explainability for Future Intelligent Space Systems | Federico Sabbatini, Catia Grimani, Roberta Calegari |
9 October
| Time | Session | Paper | Authors |
| Session VI Main: Chair Luca Pasa | |||
| 9/10 9:00 | Main | SeeAct-VC: Improving Browser Automation Through Complementary Visual Clues | Zeynep Sila Kaya, Luca Bergamin, Davide Rigoni, Ahmed Shahhat, Marco De Toni, Alessandro Sperduti, Luca Pasa |
| 9/10 9:20 | Main | MARLON: MARine Litter Object detectioN | Francisco Cruz Macedo, Luis Macedo, João Duarte, Gil Gonçalves |
| 9/10 9:40 | Main | The Generalization Gap in Named Entity Recognition: Static Benchmarks Overestimate the Transferable Performance of Neural Pattern Recognizers | Varun Kotte |
| 9/10 10:00 | Coffee Break | ||
| Special Session Explainability: Chair Federico Sabbatini | |||
| 9/10 10:20 | Explainability | Pedagogical Knowledge Extraction from Black-Box Classifiers with DiViNE | Federico Sabbatini, Roberta Calegari |
| 9/10 10:40 | Explainability | Semantic Structure in Large Language Model Disagreement: An Analysis of Representation Spaces in Conspiracy Content Classification | Ayan-Yue Gupta, Giovanna Maria Dimitri, Raffaele Vacca |
| 9/10 11:00 | Explainability Invited Talk | Formal Explanations: From Classification to Ranking | Dr. Francesco Chiariello, Inria & ENS | Université PSL (France) |
| 9/10 11:50 | Explainability | How Instruction-Tuning Imparts Length Control: A Cross-Lingual Mechanistic Analysis | Elisabetta Rocchetti, Alfio Ferrara |
| 9/10 12:10 | Explainability | Explaining Country Image Classification through Network Science: Insights into DNN Interpretability | Niccolò Di Marco, Omran Alamayreh, Giovanna Maria Dimitri, Gabriele Etta, Benedetta Tondi, Walter Quattrociocchi |
| 9/10 12:30 | Closing Panel and Awards Announcement | ||
| 9/10 13:00 | Lunch and Farewell |
Program and Social Event
The social Dinner will be held at Cantinone Restaurant on the 7th of October at 8.30 pm!
The venue is walking distance (10 minutes) from the conference venue, and we will have a full 4 course menu tasting local food and delicacies and is located 1 minute away from the Duomo Square.
The address is: VIA DELL’UNIONE 8, 20122, Milan (Italy)
Following our Call for Special Sessions we are glad to announce the Special sessions accepted! The two special sessions are described below! To submit to this special session follow the...
Read moreThe social Dinner will be held at Cantinone Restaurant on the 7th of October at 8.30 pm! The venue is walking distance (10 minutes) from the conference venue, and we...
Read moreSander Dieleman, Google DeepMind, London (UK) 7th October 2026 Morning Keynote Sander Dieleman is a Research Scientist at Google DeepMind in London, UK, where he has worked on the development of...
Read more