Generative Relational Learning (Czech Science Foundation Project)

This project is planned to have four main parts. The first part addresses theoretical foundations. The second part is about the framework of Neural Markov Logic Networks that we introduced recently. The third part is the framework of relational variational autoencoders. The fourth part of the project is further development of Lifted Relational Neural Networks and their applications for generative modelling applications. The project runs from 2020 to 2022.

Journal and conference papers

  1. 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. [preliminary version]
  2. Ondřej Kuželka. Complex Markov Logic Networks: Expressivity and Liftability. UAI 2020: 36th Conference on Uncertainty in Artificial Intelligence, 2020. [preliminary version]
  3. 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. [pdf coming soon]
  4. Ondřej Kuželka and Yuyi Wang. Domain-Liftability of Relational Marginal Polytopes. AISTATS 2020: 23rd International Conference on Artificial Intelligence and Statistics, 2020. [arxiv]
  5. 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]

Team Members

  • Ondřej Kuželka (PI, [homepage])
  • Jianhang Ai (PhD student co-supervised with Filip Železný).
  • Jáchym Barvínek (PhD student co-supervised with Filip Železný).
  • Martin Svatoš (PhD student co-supervised with Filip Železný).
  • Gustav Šourek (PhD student co-supervised with Filip Železný).

Preprints / Technical Reports

  • Ondřej Kuželka. Weighted First-Order Model Counting in the Two-Variable Fragment With Counting Quantifiers. [arxiv]
  • Gustav Šourek, Filip Železný and Ondřej Kuželka. Beyond Graph Neural Networks with Lifted Relational Neural Networks. [arxiv]
  • Ondřej Kuželka. Lifted Inference in 2-Variable Markov Logic Networks with Function and Cardinality Constraints. [arxiv]
  • Giuseppe Marra and Ondřej Kuželka. Neural Markov Logic Networks. [arxiv]