Hexapawn is a tiny, beginner-friendly strategy game played on a 3×3 board using just pawns. Each player tries to reach the opponent’s back row, capture all opposing pawns, or leave the opponent with no legal moves. Because the game is so small, it’s perfect for exploring basic ideas in game strategy and AI, even if even the younger participants of the workshop managed to find perfect strategies for their AIs.

Thank you to everyone involved in the organization—especially my great assistants, who were fantastic at working with the kids! The Swedish material I made for the lab can (at least for a while) be found on my Google Drive. The idea for the lab was heavily influenced by a YouTube video by Vsauce2 and the AI Unplugged.org materials.

At the end of January I visited Singapore for the first time in order to attend AAAI 2026. I had two accepted papers, I was there to present a poster on our paper on detecting ordered objectives in MaxSAT and to support my co-author who gave an oral presentation on our paper on efficient and reliable hitting set computations. Additionally, I was invited to give a talk in a workshop on automated reasoning at Nanyang Technological University. There I talked about my more recent work on computing explanations for tree ensembles. I would like thank Yonk Kiam Tan for the invite.

It was my first time visiting both Singapore and AAAI. I am happy with the visit, even if coordinating a conference as large as AAAI has gotten is challening at best. Thanks to the organizers of AAAI and to all colleagues I met in Singapore for the discussions. See you all next time!

I am glad to report that my papers “Ordered Objectives in Maximum Satisfiability”, and “Efficient and Reliable Hitting-Set Computations for the Implicit Hitting Set Approach” were accepted at AAAI 2026 as two of the roughly 4000 papers. Both papers deal with fundamental approaches to combinatorial optimization and how to improve our understanding, efectivenesss and trustworthiness of fundamental AI methods.

I would like to thank all of my coauthors for their work. Once the papers are officially published, I will add links to them on these pages. Before that, you can come to Singapore and talk to us at the posters. See you there!

I would like to (once again) thank everyone involved in the organization of the conferences. Glasgow and the University of Glasgow were wonderful hosts and I recommend both Glasgow, and Beth Mays show wholeheartedly to anyone.

I am happy to let you know that our paper Engineering and Evaluating Multi-objective Pseudo-Boolean Optimizers has been accepted for publication at the 19th edition of the European Conference on Logics in Artificial Intelligence. In the paper, we report on a cross-community evaluation of different approaches for computing pareto-optimal solutions under 0-1 integer linear inequalities. We also build on out recent TACAS paper in order to show how such solvers can be proof-logged.

I want to thank all of my co-authors, especially my student Christoph Jabs, for their excellent work on the project!

I was very happy to learn, that the paper Symmetric Core Learning for Pseudo-Boolean Optimization by Implicit Hitting Sets by my student Hannes Ihalainen will be awarded the best student paper of the 31st International Conference on Principles and Practice of Constraint Programming, one of the foremost international conference in my research field.

The paper presents an approach for using the symmetries of an optimization problem in order to speed up search for an optimal solution. The paper summarizes work that was started briefly after my 2020 paper on abstract cores when Bart Bogaerts, one of the co-authors of this paper contacted me.

I would like to congratulate Hannes and thank all of the other people on this paper as well, Bart Boagerts and Matti Järvisalo. I look forward to travelling to Glasgow in August and meeting many colleagues and friends.

We are all amateurs, so the primary goal of the trip was to have fun and explore some nice courses. From that perspective, the trip was a huge success. We were delighted with the Hektor Design Hostel in Tartu, with the nearby restaurants, and the courses we played. The only thing that could have been–slightly–better was the weather, which was either a bit hot for achieving maximal performance on the course, or a bit rainy. But on the whole, a very lovely trip. Thanks to Lasse and Tobbe for organizing and being good company.

I am happy to let you know that our papers SLS-Enhanced Core-Boosted Linear Search for Anytime Maximum Satisfiability, From Scalable SAT to MaxSAT: Massively Parallel Solution Improving Search and Symmetric Core Learning for Pseudo-Boolean Optimization by Implicit Hitting Sets have been accepted at the The 18th International Symposium on Combinatorial Search and The 31st International Conference on Principles and Practice of Constraint Programming.

The papers describe a massively distributed search algorithm for Maximum Satisfiability, a method of incorporating symmetry reasoning in an implicit hitting set approach that does not decrease performance, and an empirical study of combining SLS methods with CDCL reasoning. I will link the papers to these pages after the conferences in August.

I want to thank all my co-authors for their outstanding work on the projects. I hope to meet many of you in Glasgow.

I was happy with how the presentation turned out, and I would like to thank the course organizers (mainly Stefan Taubert) for the opportunity, as well as Frida Hartman, who gave the second computer science talk on fairness in AI. My slides can be found below.

I am happy to let you know that our paper Certifying Pareto-Optimality in Multi-Objective Maximum Satisfiability has been accepted for publication at the 31st International Conference on Tools and Algorithms for the Construction and Analysis of Systems. We show how the pseudo-Boolean proof-logging system VeriPB can be used to log the Pareto-optimal solutions of multi-objective problems without any extensions to the proof system itself. This significantly increases the applicability of the system. The full paper can be found here in open access.

I want to thank all of my co-authors for their excellent work on the project!