Seminar Talk Announcement

• Jan Flusser (Institute of Information Theory and Automation, Prague):

  Decomposition of 2D and 3D shapes into rectangular blocks.

  13.10.2026 13:30 CESTPlenary room (room. 318, second floor) Institute of Computer Science (ZOOM 914 0834 4018, 668534)Hora Informaticae

  The lecture presents an overview of the most important methods that decompose an arbitrary binary object into a union of rectilinear rectangles. We describe a run-length encoding and its generalization, decomposition based on quadtrees, on mathematical morphology, on the distance transform, and a theoretically optimal decomposition based on a maximal matching in bipartite graphs. While in 2D the optimal algorithm has a polynomial complexity, its 3D analog is NP- complete. Polynomial approximation of 3D optimal algorithm will be demonstrated. The lecture is self-contained, accessible also for non-specialists outside computer graphics and vision areas, and does not require any prior knowledge of the subject.

• Storm Visser:

  Artificial Immune Systems: Multiclass Classification and Hybrid Methods

  15.10.2026 14:00Lecture room S8, Faculty of Mathematics and Physics, Malostranské náměstí 2Machine Learning and Modelling Seminar

  Storm will explain the basics of Artificial Immune Systems, discuss some of the recent advances in the field, including multiclass and fine-grained classification, as well as potential future work

Past Talks

• Reinforcement Learning Built for Acceleration: Sparse Topology Control, Stability, and Hardware-Aware Training
  Isam Vrce
  01.10.2026 14:00 @ Machine Learning and Modelling Seminar
• Multi-Objective Optimization for Expensive Engineering Problems.
  Jana Burkotová (Faculty of Science, Palacký University Olomouc, Czech Republic)
  26.05.2026 @ Hora Informaticae
• Multi-objective Neural Architecture Design Using Monte Carlo Search
  Noe Lallouet
  14.05.2026 14:00 @ Machine Learning and Modelling Seminar
• Energy Portfolio Optimisation Project - Lessons Learned.
  Petra Vidnerová (Institute of Computer Science, Czech Academy of Sciences, Prague)
  12.05.2026 @ Hora Informaticae
• Scaling LLM Alignment Beyond Supervised Fine-Tuning: Solving the Data Reliance Bottleneck
  Dario Vajda
  30.04.2026 14:00 @ Machine Learning and Modelling Seminar
• A Sociological Application of Topic Modeling in Science Mapping.
  Radim Hladík (Institute of Philosophy, Czech Academy of Sciences)
  28.04.2026 @ Hora Informaticae
• Automated and Robust Shape Optimisation of Industrial Wastewater Systems Using Multi-Objective Bayesian Methods
  Benjamin Moore
  16.04.2026 14:00 @ Machine Learning and Modelling Seminar
• Generalization in Attention-Based Neural Networks: Solvable High-Dimensional Models.
  Lenka Zdeborová (École Polytechnique Fédérale de Lausanne, Swiss)
  14.04.2026 @ Hora Informaticae
• Evolving Prompt Embeddings of Diffusion Models for Optimization and Exploration of Synthetic Images
  Marcel Salvenmoser
  02.04.2026 14:00 @ Machine Learning and Modelling Seminar
• On Viable Computation and Trustworthy AI.
  Jiří Wiedermann (Institute of Computer Science, Czech Academy of Sciences, Prague)
  31.03.2026 @ Hora Informaticae
• Conformal Predictors meet Active Learning: Adaptive Systems for Efficient Industrial Monitoring in Non-Stationary Environments
  David Moe
  19.03.2026 14:00 @ Machine Learning and Modelling Seminar
• How to turn a Transformer model into a modern LLM? Instruct, RLHF, RAG, and others.
  Rudolf Rosa (Institute of Formal and Applied Linguistics, Charles Univ., Fac. Math. and Phys., Prague)
  17.03.2026 @ Hora Informaticae
• From artificial neurons to artificial intelligence.
  Věra Kůrková (Institute of Computer Science, Czech Academy of Sciences, Prague)
  03.03.2026 @ Hora Informaticae
• Machine-learned interatomic potentials for molecular dynamics: committees and foundation models
  Ondřej Maršálek and Pavol Šimko
  19.02.2026 14:00 @ Machine Learning and Modelling Seminar
• Riesz basis of ReLU neural networks and its use in function recovery.
  Jan Vybíral (Department of Mathematics, FNSPE, CTU Prague)
  10.02.2026 @ Hora Informaticae
• Robustness of Deep ReLU Networks to Misclassification of High-Dimensional Data.
  Věra Kůrková (Institute of Computer Science, Czech Academy of Sciences, Prague)
  18.11.2025 @ Hora Informaticae
• A General Constructive Upper Bound on Shallow Neural Nets Complexity.
  František Hakl, Vít Fojtík (Institute of Computer Science, Czech Academy of Sciences, Prague, Ludwig-Maximilians-Universität München)
  04.11.2025 @ Hora Informaticae
• Composite indicators in social sciences and clinical biostatistics
  Marco Marozzi
  22.07.2025 13:00 @ International Society for Clinical Biostatistics
• From Reasoning Agents to Language Agents: How LLMs are Reshaping the AI Agent Paradigm.
  Roman Neruda (Institute of Computer Science, Czech Academy of Sciences, Prague)
  17.06.2025 @ Hora Informaticae
• Understanding the neural networks through rule extraction.
  Tomáš Pevný (Czech Technical University, Prague)
  20.05.2025 @ Hora Informaticae
• Turning psychometrics on its head: What if our constructs were networks?
  Denny Borsboom (University of Amsterdam)
  12.05.2025 16:00 @ Computer Psychometrics Group
• Abstraction-Based Machine-Code Program Verification.
  Jan Onderka (Institute of Computer Science, Prague)
  22.04.2025 @ Hora Informaticae
• Proof By Abduction in Isabelle/HOL.
  Yutaka Nagashima (Institute of Computer Science, Prague)
  04.02.2025 @ Hora Informaticae
• New Functions Characterizing Continuous Distributions.
  Zdeňek Fabián (Institute of Computer Science, Prague)
  21.01.2025 @ Hora Informaticae
• Robots as Moral Agents: Human Intuition, Programmed Ethics, and Real-World Implications.
  Robin Kopecký (Institute of Philosophy, Prague)
  03.12.2024 @ Hora Informaticae
• Non-standard situations in regression modeling: From method selection to neural networks.
  Jan Kalina (Institute of Computer Science, Prague)
  05.11.2024 @ Hora Informaticae
• Approximation and classification by ReLU networks of increasing depth.
  Věra Kůrková (Institute of Computer Science, Prague)
  22.10.2024 @ Hora Informaticae
• Large Languages Models and Automata Theory.
  Jiří Wiedermann (Institute of Computer Science, Prague)
  08.10.2024 @ Hora Informaticae
• Language models and Artificial Inteligence.
  David Herel (Department of Cybernetics, Faculty of Electrical Engineering, CTU Prague)
  11.06.2024 @ Hora Informaticae
• Evolutionary Algorithms for Expansive Optimization.
  Martin Pilát (Faculty of Mathematics and Physics, Charles University, Prague)
  28.05.2024 @ Hora Informaticae
• Equivariance and Invariance in Neural Networks.
  Filip Šroubek and Tomáš Karella (Institute of Information Theory and Automation, Prague)
  14.05.2024 @ Hora Informaticae
• Czech Speech Synthesis in the AI Era.
  Jindřich Matoušek (Department of Cybernetics, Faculty of Applied Sciences, UWB Pilsen)
  16.04.2024 @ Hora Informaticae
• Performance Prediction for Neural Architecture Search.
  G. Kadlecová, P. Vidnerová (Department of Artificial Intelligence, ICS Prague)
  19.03.2024 @ Hora Informaticae
• Learning from few examples - nonlinearity and dimensionality.
  Oliver Sutton (Department of Mathematics, King's College London)
  12.03.2024 @ Hora Informaticae
• Dispersion of a point set - lower and upper bounds.
  Jan Vybiral (Department of Mathematics, FNSPE, CTU)
  20.02.2024 @ Hora Informaticae
• GPU-acceleration of program synthesis.
  Martin Berger (Montanarius Ltd, University of Sussex)
  30.01.2024 @ Hora Informaticae
• Getting Structure in Dialogue with Large Language Models.
  Ondřej Dušek (Institute of Formal and Applied Linguistics, Charles University Prague)
  23.01.2024 @ Hora Informaticae
• Principlism as the ethics of artificial intelligence.
  David Černý (Institute of Computer Science CAS, Institute of State and Law, CAS)
  09.01.2024 @ Hora Informaticae
• The Voynich Manuscript: the secret book of an unknown language.
  Ivan Zelinka (Technical University of Ostrava)
  05.12.2023 @ Hora Informaticae
• Chomsky-Like Neural Network Hierarchy.
  Jiří Šíma (Department of Theoretical Computer Science, CAS)
  07.11.2023 @ Hora Informaticae
• Approximation of classifiers of large data sets by deep ReLU networks.
  Věra Kůrková (Department of Artificial Inteligence, Institute of Computer Science, CAS)
  24.10.2023 @ Hora Informaticae
• Artificial Wisdom: On the Power of Generative AI.
  Jiří Wiedermann, Jan van Leeuwen (ICS CAS, Utrecht University)
  10.10.2023 @ Hora Informaticae
• HW Accelerated AI Inference and Fast, Recursive QR System Identification.
  Jiří Kadlec (Department of Signal processing, UTIA, CAS)
  13.06.2023 @ Hora Informaticae
• Causality and Machine Learning.
  Aditi Kathpalia (Department of Complex Systems, Institute of Computer Science, CAS)
  30.05.2023 @ Hora Informaticae
• TNL: Numerical library for modern parallel architectures.
  Tomáš Oberhuberk (Department of Mathematics, FNSPE, CTU)
  16.05.2023 @ Hora Informaticae
• Computability of Neural Network Problems.
  Vít Fojtík (Bavarian AI Chair for Mathematical Foundations of AI, LMU Munich)
  02.05.2023 @ Hora Informaticae
• Zero-cost proxies for neural network performance estimation.
  Gabriela Kadlecová (Institute of Computer Science; Charles Univ., Fac Math and Phys)
  18.04.2023 @ Hora Informaticae
• GPT et al.: Generating Texts with Transformer-Based Large Language Models.
  Rudolf Rosal (Institute of Formal and Applied Linguistics, Charles Univ., Fac Math and Phys)
  04.04.2023 @ Hora Informaticae
• From ridge functions to neural networks II.
  Jan Vybíral (Department of Mathematics, FNSPE, CTU)
  21.03.2023 @ Hora Informaticae
• Learning Logic Programs with negation, Predicate Invention, and Higher-Order Definitions Through the learning from Failures Paradigm.
  David M. Cerna (Department of Artificial Intelligence, ICS Prague)
  07.03.2023 @ Hora Informaticae
• From discrete to continuous variational inequalities with convex and non-convex constraints.
  Pavel Krejčí (Institute of Mathematics, CAS; Faculty of Civil Engineering, CTU)
  21.02.2023 @ Hora Informaticae
• From ridge functions to neural networks I.
  Jan Vybíral (Department of Mathematics, FNSPE, CTU)
  24.01.2023 @ Hora Informaticae
• How radar interferometry could reconcile fuzzy sets with probability
  Lubomír Soukup (Institute of Information Theory and Automation)
  10.01.2023 @ Hora Informaticae
• Probabilistic Forecasting with International Large-Scale Assessments: Applications to the UN Sustainable Development Goals
  David Kaplan (University of Wisconsin)
  20.12.2022 @ Hora Informaticae
• Model M - an agent-based epidemiological model
  RNDr. Petra Vidnerová, Ph.D.
  06.12.2022 @ Hora Informaticae
• Some implications of high-dimensional geometry for classification by neural networks
  RNDr. Věra Kůrková, DrSc.
  22.11.2022 @ Hora Informaticae
• Autonomous machines and machine minds
  prof. RNDr. Jiří Wiedermann, DrSc.
  08.11.2022 @ Hora Informaticae
• How to deal with data contamination in hypothesis testing
  RNDr. Jan Kalina, Ph.D.
  16.12.2019 @ Hora Informaticae
• Probabilistic Bounds on Complexity of Networks Classifying Large Data Sets
  RNDr. Věra Kůrková, DrSc.
  18.11.2019 @ Hora Informaticae
• Interpolant generation algorithms based on quantifier elimination
  Deepak Kapur (University of Mexico)
  01.07.2019 @ Hora Informaticae
• Automata and Turing Computation with Synfire Ring-Based Neural Networks
  Jérémie Cabessa (Université Paris 2)
  04.03.2019 @ Hora Informaticae
• Deep Neural Networks in Natural Language Processing
  Rudolf Rosa (UFAL MFF UK)
  14.01.2019 @ Hora Informaticae