Ondřej Kuželka

I am an assistant professor working on ML and AI in CS Dept @FEE, Czech Technical University in Prague. In the past, I did two postdocs in the DTAI group at KU Leuven, with Luc De Raedt and Jan Ramon, and one at Cardiff University with Steven Schockaert.

ondrej.kuzelka at fel.cvut.cz

Journal and conference papers

  1. Yuanhong Wang, Timothy van Bremen, Yuyi Wang and Ondřej Kuželka. Domain-Lifted Sampling for Universal Two-Variable Logic and Extensions. AAAI 2022: 36th AAAI Conference on Artificial Intelligence, 2022. [online]
  2. Jianhang Ai, Ondřej Kuželka and Yuyi Wang. Hoeffding–Serfling Inequality for U-Statistics Without Replacement. Journal of Theoretical Probability, 2022. [online]
  3. Jianhang Ai, Ondřej Kuželka and Yuyi Wang. Hoeffding and Bernstein Inequalities for U-statistics Without Replacement. Statistics and Probability Letters, 2022. [online]
  4. Ondřej Kuželka. Weighted First-Order Model Counting in the Two-Variable Fragment With Counting Quantifiers. Journal of Artificial Intelligence Research, 2021. [online]
  5. Giuseppe Marra and Ondřej Kuželka. Neural Markov Logic Networks. UAI 2021: 37th Conference on Uncertainty in Artificial Intelligence, 2021. [pdf] [supplement]
  6. Timothy van Bremen and Ondřej Kuželka. Lifted Inference with Tree Axioms. KR 2021: 18th International Conference on Principles of Knowledge Representation and Reasoning, 2021. (Marco Cadoli Best Student Paper Award Runner-up) [online]
  7. Timothy van Bremen and Ondřej Kuželka. Faster Lifting for Two-Variable Logic Using Cell Graphs. UAI 2021: 37th Conference on Uncertainty in Artificial Intelligence, 2021. [pdf]
  8. Jáchym Barvínek, Timothy van Bremen, Yuyi Wang, Filip Železný and Ondřej Kuželka. Automatic Conjecturing of P-Recursions Using Lifted Inference. ILP 2020-21: 30th International Conference on Inductive Logic Programming, 2021. [preliminary version]
  9. Yuanhong Wang, Timothy van Bremen, Juhua Pu, Yuyi Wang and Ondřej Kuželka. Fast Algorithms for Relational Marginal Polytopes. IJCAI 2021: 30th International Joint Conference on Artificial Intelligence, 2021. [online]
  10. Nitesh Kumar, Ondřej Kuželka and Luc De Raedt. Learning Distributional Programs for Relational Autocompletion. Theory and Practice of Logic Programming (accepted). [preliminary version]
  11. Nitesh Kumar and Ondřej Kuželka. Context-Specific Likelihood Weighting. AISTATS 2021: 24th International Conference on Artificial Intelligence and Statistics, 2021. [online]
  12. Gustav Šourek, Filip Železný and Ondřej Kuželka. Lossless Compression of Structured Convolutional Models via Lifting. ICLR 2021: The Ninth International Conference on Learning Representations, 2021. [online]
  13. Gustav Šourek, Filip Železný and Ondřej Kuželka. Beyond Graph Neural Networks with Lifted Relational Neural Networks. Machine Learning, 2021. [online] [preliminary version]
  14. Ondřej Kuželka, Vyacheslav Kungurtsev and Yuyi Wang. Lifted Weight Learning of Markov Logic Networks (Revisited One More Time). PGM 2020: 10th International Conference on Probabilistic Graphical Models, 2020. [online]
  15. Ondřej Kuželka. Complex Markov Logic Networks: Expressivity and Liftability. UAI 2020: 36th Conference on Uncertainty in Artificial Intelligence, 2020. [arxiv]
  16. Timothy van Bremen and Ondřej Kuželka. Approximate Weighted First-Order Model Counting: Exploiting Fast Approximate Model Counters and Symmetry. IJCAI 2020: 29th International Joint Conference on Artificial Intelligence, 2020. [online]
  17. Ondřej Kuželka and Yuyi Wang. Domain-Liftability of Relational Marginal Polytopes. AISTATS 2020: 23rd International Conference on Artificial Intelligence and Statistics, 2020. [arxiv]
  18. Martin Svatoš, Steven Schockaert, Jesse Davis and Ondřej Kuželka. STRiKE: Rule-Driven Relational Learning Using Stratified k-Entailment. ECAI 2020: 24th European Conference on Artificial Intelligence, 2020. [pdf]
  19. Ondřej Kuželka and Jesse Davis. Markov Logic Networks for Knowledge Base Completion: A Theoretical Analysis Under the MCAR Assumption. UAI 2019: 35th Conference on Uncertainty in Artificial Intelligence, 2019. [online]
  20. Ondřej Kuželka and Vyacheslav Kungurtsev. Lifted Weight Learning of Markov Logic Networks Revisited. AISTATS 2019: 22nd International Conference on Artificial Intelligence and Statistics, 2019. [arxiv]
  21. Arcchit Jain, Tal Friedman, Ondřej Kuželka, Guy Van Den Broeck and Luc De Raedt. Scalable Rule Learning in Probabilistic Knowledge Bases. AKBC 2019: Automated Knowledge Base Construction, 2019. [online]
  22. Victor Gutierrez Basulto, Jean Christoph Jung and Ondřej Kuželka. Quantified Markov Logic Networks. KR 2018: 16th International Conference on Principles of Knowledge Representation and Reasoning, 2018. [arxiv]
  23. Ondřej Kuželka, Yuyi Wang, Jesse Davis and Steven Schockaert. PAC-Reasoning in Relational Domains. UAI 2018: 34th Conference on Uncertainty in Artificial Intelligence, 2018. [online]
  24. Ondřej Kuželka, Yuyi Wang and Steven Schockaert. VC-Dimension Based Generalization Bounds for Relational Learning. ECMLPKDD 2018: European Conference on Machine Learning and Knowledge Discovery in Databases, 2018. [arxiv]
  25. Ondřej Kuželka, Yuyi Wang, Jesse Davis and Steven Schockaert. Relational Marginal Problems: Theory and Estimation. AAAI 2018: 32nd AAAI Conference on Artificial Intelligence, 2018. [arxiv]
  26. Thomas Ager, Ondřej Kuželka and Steven Schockaert. Modelling Salient Features as Directions in Fine-Tuned Semantic Spaces CoNLL 2018: SIGNLL Conference on Computational Natural Language Learning, 2018. [online]
  27. Gustav Šourek, Vojtěch Aschenbrenner, Filip Železný, Steven Schockaert and Ondřej Kuželka. Lifted Relational Neural Networks: Efficient Learning of Latent Relational Structures. Journal of Artificial Intelligence Research. 2018. [online]
  28. Ondřej Kuželka, Jesse Davis and Steven Schockaert. Induction of Interpretable Possibilistic Logic Theories from Relational Data. IJCAI 2017: 26th International Joint Conference on Artificial Intelligence, 2017. [arxiv]
  29. Gustav Šourek, Martin Svatoš, Filip Železný, Steven Schockaert and Ondřej Kuželka. Stacked Structure Learning for Lifted Relational Neural Networks. ILP 2017: Post-proceedings of the 27th International Conference on Inductive Logic Programming, 2017. (Best Paper Award) [arxiv]
  30. Martin Svatoš, Gustav Šourek, Filip Železný, Steven Schockaert and Ondřej Kuželka. Pruning Hypothesis Spaces Using Learned Domain Theories. ILP 2017: Post-proceedings of the 27th International Conference on Inductive Logic Programming, 2017. [pdf]
  31. Ondřej Kuželka, Jesse Davis and Steven Schockaert. Learning Possibilistic Logic Theories from Default Rules. IJCAI 2016: 25th International Joint Conference on Artificial Intelligence, 2016. [pdf] [longer arxiv version]
  32. Ondřej Kuželka, Yuyi Wang and Jan Ramon. Bounds for Learning from Evolutionary-Related Data in the Realizable Case. IJCAI 2016: 25th International Joint Conference on Artificial Intelligence, 2016. [pdf]
  33. Ondřej Kuželka, Jesse Davis and Steven Schockaert. Interpretable Encoding of Densities using Possibilistic Logic. ECAI 2016: 22nd European Conference on Artificial Intelligence, 2016. [pdf]
  34. Gustav Šourek, Suresh Manandhar, Filip Železný, Steven Schockaert and Ondřej Kuželka. Learning Predictive Categories Using Lifted Relational Neural Networks. ILP 2016: Post-proceedings of the 26th International Conference on Inductive Logic Programming 2016. [pdf]
  35. Radomír Černoch, Ondřej Kuželka and Filip Železný. Polynomial and Extensible Solutions in Lock-Chart Solving. Applied Artificial Intelligence 30(10): 923-941, 2016. [online]
  36. Ondřej Kuželka, Jesse Davis and Steven Schockaert. Encoding Markov logic networks in Possibilistic Logic. UAI 2015: Uncertainty in Artificial Intelligence, 2015. [pdf] [online]
  37. Ondřej Kuželka and Jan Ramon. Mine ’Em All: A Note on Mining All Graphs. ILP 2015: Post-proceedings of the 25th International Conference on Inductive Logic Programming 2015. [pdf] [online]
  38. Ondřej Kuželka, Jesse Davis and Steven Schockaert. Constructing Markov Logic Networks from First-Order Default Rules. ILP 2015: Post-proceedings of the 25th International Conference on Inductive Logic Programming 2015. [pdf] [online]
  39. Matěj Holec, Ondřej Kuželka and Filip Železný. Novel Gene Sets Improve Set-Level Classification of Gene Expression Data. BMC Bioinformatics 16: 348, 2015. [online]
  40. Gustav Šourek, Ondřej Kuželka and Filip Železný. Learning to detect network intrusion from a few labeled events and background traffic. AIMS 2015: Autonomous Infrastructure, Management and Security, 2015. [pdf] [online]
  41. Ondřej Kuželka, Andrea Szabóová and Filip Železný, A Method for Reduction of Examples in Relational Learning. Journal of Intelligent Information Systems 42(2): 255-281, 2014. [pdf] [online]
  42. Roman Barták, Radomír Černoch, Ondřej Kuželka and Filip Železný. Formulating the Template ILP Consistency Problem as a Constraint Satisfaction Problem. Constraints 18(2): 144-165, 2013. [online]
  43. Andrea Szabóová, Ondřej Kuželka, Filip Železný and Jakub Tolar. Prediction of DNA-binding proteins from relational features. Proteome Science, 10(1), 66, 2012. [online]
  44. Andrea Szabóová, Ondřej Kuželka, Filip Železný and Jakub Tolar. Prediction of DNA-binding Propensity of Proteins by the Ball-Histogram Method using Automatic Template Search. BMC Bioinformatics, 13, Sup 10, 2012.
  45. Ondřej Kuželka, Andrea Szabóová and Filip Železný. Bounded Least General Generalization. ILP 2012: Inductive Logic Programming, 2012. [pdf] [online]
  46. Ondřej Kuželka, Andrea Szabóová and Filip Železný. Extending the Ball-Histogram Method with Continuous Distributions and an Application to Prediction of DNA-Binding Proteins. BIBM 2012: IEEE International Conference on Bioinformatics and Biomedicine, 2012. [pdf]
  47. Ondřej Kuželka, Andrea Szabóová and Filip Železný. Relational Learning with Polynomials. ICTAI 2012: IEEE International Conference on Tools with Artificial Intelligence, 2012. [pdf]
  48. Ondřej Kuželka and Filip Železný. Block-Wise Construction of Tree-like Relational Features with Monotone Reducibility and Redundancy. Machine Learning 83(2): 163-192, 2011. [pdf] [online]
  49. Ondřej Kuželka, Andrea Szabóová, Matěj Holec and Filip Železný. Gaussian Logic for Predictive Classification. ECML/PKDD 2011: European Conference on Machine Learning and Principles and Practice of Knowledge Discovery in Databases, 2011. [pdf] [online]
  50. Ondřej Kuželka, Andrea Szabóová and Filip Železný. Gaussian Logic and Its Applications in Bioinformatics. ACM-BCB 2011: ACM Conference on Bioinformatics, Computational Biology and Biomedicine, 2011. [pdf]
  51. Andrea Szabóová, Ondřej Kuželka, Sergio Morales E., Filip Železný and Jakub Tolar. Prediction of DNA-binding Propensity of Proteins by the Ball-Histogram Method. ISBRA 2011: The 7th International Symposium on Bioinformatics Research and Applications, 2011.[online]
  52. Ondřej Kuželka and Filip Železný. Seeing the World through Homomorphism: An Experimental Study on Reducibility of Examples. ILP 2010: Inductive Logic Programming, 2010. [pdf] [online]
  53. Roman Barták, Ondřej Kuželka and Filip Železný. Using Constraint Satisfaction for Learning Hypotheses in Inductive Logic Programming. FLAIRS 2010 - Florida Artificial Intelligence Research Society, The 23rd International Conference, 2010.
  54. Filip Železný and Ondřej Kuželka. Taming the Complexity of Inductive Logic Programming (invited talk). SOFSEM 2010: 36th International Conference on Current Trends in Theory and Practice of Computer Science, 2010.
  55. Ondřej Kuželka and Filip Železný. Block-Wise Construction of Acyclic Relational Features with Monotone Irreducibility and Relevancy Properties. ICML 2009: The 26th International Conference on Machine Learning, 2009. [online]
  56. Ondřej Kuželka and Filip Železný. A Restarted Strategy for Efficient Subsumption Testing. Fundamenta Informaticae, 89:95-109, 2008. [pdf]
  57. Ondřej Kuželka and Filip Železný. Fast Estimation of First-Order Clause Coverage through Randomization and Maximum Likelihood. ICML 2008: 25th International Conference on Machine Learning, 2008. [online]

Lightly refereed papers (workshop…)

  1. Gustav Šourek, Filip Železný and Ondřej Kuželka. Learning with Molecules beyond Graph Neural Networks. ML4Molecules workshop @ Neurips 2020. [arxiv]
  2. Timothy van Bremen and Ondřej Kuželka. Approximate Weighted First-Order Model Counting: Exploiting Fast Approximate Model Counters and Symmetry. StarAI 2020 Workshop, 2020. [arxiv]
  3. Giuseppe Marra and Ondřej Kuželka. Neural Markov Logic Networks. NeurIPS 2019 Workshop KR2ML, 2019. [arxiv]
  4. Ondřej Kuželka and Yuyi Wang. Generalization Bounds for Knowledge Graph Embedding (Trained by Maximum Likelihood). NeurIPS 2019 Workshop on Machine Learning with Guarantees, 2019. [pdf]
  5. Ondřej Kuželka, Jesse Davis and Steven Schockaert. Stratified Knowledge Bases as Interpretable Probabilistic Models (Extended Abstract). NIPS 2016 Workshop on Interpretable Machine Learning in Complex Systems, 2016. [arxiv]
  6. Thomas Ager, Ondřej Kuželka and Steven Schockaert. Inducing symbolic rules from entity embeddings using auto-encoders. NeSy 2016: Proceedings of the 11th International Workshop on Neural-Symbolic Learning and Reasoning. [online]
  7. Gustav Šourek, Vojtěch Aschenbrenner, Filip Železný and Ondřej Kuželka. Lifted Relational Neural Networks. CoCo 2015: Cognitive Computation: Integrating Neural and Symbolic Approaches, 2015. [pdf], [longer arxiv version], [video of Gustav's talk]
  8. Ondřej Kuželka and Jan Ramon. A Note on Restricted Forms of LGG. Late-breaking proceedings of ILP 2015. [pdf] [longer version]
  9. Vojtěch Aschenbrenner and Ondřej Kuželka: Horn-Clause Neural Networks (poster). Spring workshop on Mining and Learning (SML), 2014. [poster]
  10. Gustav Šourek, Ondřej Kuželka and Filip Železný: Predicting Top-k Trends on Twitter using Graphlets and Time Features. ILP 2013: Inductive Logic Programming - Late Breaking Papers, 2013. [pdf]
  11. Andrea Fuksová, Ondřej Kuželka and Andrea Szabóová: A Method for Mining Discriminative Graph Patterns. MLCB 2013: NIPS Machine Learning in Computational Biology Workshop, 2013. [online]
  12. Ondřej Kuželka, Andrea Szabóová and Filip Železný. Reducing Examples in Relational Learning with Bounded-Treewidth Hypotheses. Selected papers from the Workshop on New Frontiers in Mining Complex Patterns (NFMCP 2012), 2012. [pdf] [online]
  13. Andrea Szabóová, Ondřej Kuželka and Filip Železný. Prediction of Antimicrobial Activity of Peptides using Relational Machine Learning. IEEE International Conference on Bioinformatics and Biomedicine Workshops (BIBMW 2012), 2012.
  14. Ondřej Kuželka and Filip Železný. An Experimental Evaluation of Lifted Gene Sets. MIND 2011: Workshop on Mining Complex Entities from Network and Biomedical Data, 2011. [pdf]
  15. Ondřej Kuželka, Andrea Szabóová, Matěj Holec and Filip Železný. Gaussian Logic for Proteomics and Genomics. MLSB 2011: the 5th International Workshop on Machine Learning in Systems Biology, 2011. [pdf]
  16. Andrea Szabóová, Ondřej Kuželka, Filip Železný and Jakub Tolar. Prediction of DNA-Binding Proteins from Structural Features. MLSB 2010: Proceedings of the Fourth International Workshop on Machine Learning in Systems Biology, 2010.
  17. Ondřej Kuželka and Filip Železný. Shrinking Covariance Matrices using Biological Background Knowledge. MLSB 2010: Proceedings of the Fourth Workshop on Machine Learning in Systems Biology, 2010. [pdf]
  18. Roman Barták, Ondřej Kuželka and Filip Železný. Formulating Template Consistency in Inductive Logic Programming as a Constraint Satisfaction Problem. AAAI-10 Workshop on Abstraction, Reformulation, and Approximation (WARA-2010), 2010.
  19. Ondřej Kuželka and Filip Železný. Block-Wise Construction of Acyclic Relational Features with Monotone Relevancy. ILP 2009: 19th International Conference on Inductive Logic Programming, 2009
  20. Ondřej Kuželka and Filip Železný. HiFi: Tractable Propositionalization through Hierarchical Feature Construction. ILP 2008: Late Breaking Papers, the 18th International Conference on Inductive Logic Programming, 2008. [pdf]

Current and Former Students

PhD (Finished )

  • Timothy van Bremen, 2022 . Tim's PhD thesis was about lifted inference, he defended his PhD at KU Leuven and was co-supervised with Luc De Raedt. [homepage]
  • Gustav Šourek, 2021 . Gustav's PhD thesis was about lifted relational neural networks (LRNNs), a deep relational learning framework. Gustav was co-supervised with Filip Železný. Winner of Antonin Svoboda Award for the Best PhD Thesis awarded by the Czech Society for Cybernetics and Informatics. [thesis] [homepage]

PhD (Current)

  • Jianhang Ai (PhD student working on concentration inequalities for ML and AI).
  • Jáchym Barvínek (PhD student working on expressivity of probabilistic logic programming, co-supervised with Filip Železný).
  • Peter Jung (PhD student working on modelling large networks).
  • Nitesh Kumar (PhD student at KU Leuven, co-supervised with Luc De Raedt, working on probabilistic logic inference). [homepage]
  • Martin Svatoš (PhD student working on structure learning for lifted relational neural networks (LRNNs)). [homepage]
  • Jan Tóth (PhD student working on neural Markov logic networks).

Informal Collaboration

  • Yuanhong Wang (PhD student at Beihang University working on lifted inference and sampling and other topics). [homepage]

BSc and MSc

  • Jan Tóth, 2021 . Dean's Award. Jan Tóth studied learning of Markov logic networks with complex weights. [thesis]
  • Jan Kozák, 2021 . Jan Kozák studied the problem of relational marginal polytope construction. [thesis]
  • Vu Huy Hoang, 2020 . Vu Huy Hoang's thesis introduced a method for learning graphons through neural network learning. [thesis]
  • Peter Jung, 2020 . The results from Peter Jung's MSc thesis on product description matching got into production. [thesis]
  • Andrea Fuksová, 2014 . Winner of CISCO outstanding thesis award, recipient of Dean's Award and Czech and Slovak ACM-SPY thesis competition finalist. Andrea Fuksová studied so-called bounded least general generalization. [thesis]
  • Gustav Šourek, 2013 . Gustav Šourek's thesis was about predicting the spread of information in social networks). [thesis]
  • Vojtěch Aschenbrenner, 2013 . The first incarnation of lifted relational neural networks (LRNNs) appeared in Vojtěch Aschenbrenner's thesis. [thesis]

Projects

Misc

TreeLiker software for relational feature construction [link].

Reviewing

I have been serving on the PC of AAAI, AISTATS, UAI, KR, IJCAI. I am an editorial board member of Machine Learning Journal (MLJ) and an associate editor of Artificial Intelligence Communications (AIC). I have also been reviewing for journals such as JMLR, AIJ, MLJ, DAMI and others.