Project Goals

The goal of ALGOLINC is to study fundamental algorithmic problems that arise from the ongoing synergies between machine learning and algorithmic game theory. The field of algorithmic game theory lies at the intersection of computer science, game theory and economics, and has evolved over the last two decades, out of the need for efficient algorithms in game-theoretic models. Our project intends to focus on two important—and to an extent complementary—research components. The first one is centered around optimization problems that form the backbone of building efficient learning (training) algorithms with an emphasis on min-max optimization problems. The second one moves along a reverse direction and concerns contributions of machine learning to mechanism design. Mechanism design forms one of the main pillars of algorithmic game theory, and our goal is to employ learning-oriented approaches (such as the paradigm of learning-augmented mechanisms) in order to design improved mechanisms, with an emphasis on auctions for allocating resources but also on other strategic settings.

Resulting publications in peer-reviewed conference proceedings

1. Michail Fasoulakis, Evangelos Markakis, Georgios Roussakis, Christodoulos Santorinaios. A Descent-based Method on the Duality Gap for Solving Zero-sum Games. In Proceedings of the 34th International Joint Conference on Artificial Intelligence, IJCAI 2025, Montreal, Canada, August 16-22, pages 3839-3847, 2025. Available here.
2. Evripidis Bampis, Bruno Escoffier, Dimitris Fotakis, Panagiotis Patsilinakos, Michalis Xefteris. Polynomial Time Learning Augmented Algorithms for NP-hard Permutation Problems. In Proceedings of the 42nd International Conference on Machine Learning, ICML 2025, Vancouver, BC, Canada, July 13-19, 2025. Available here.
3. Georgios Amanatidis, Evangelos Markakis, Christodoulos Santorinaios, Guido Schaefer, Panagiotis Tsamopoulos, Artem Tsikiridis. Online Budget-Feasible Mechanism Design with Predictions. In Proceedings of the 18th International Symposium on Algorithmic Game Theory, SAGT 2025, pages 402-421, Bath, UK, September 2-5, 2025. Available here.
4. Alviona Mantso, Evangelos Markakis, Nikos Protopapas. Fairness Under Equal-Sized Bundles: Impossibility Results and Approximation Guarantees. In Proceedings of the 18th International Symposium on Algorithmic Game Theory, SAGT 2025, pages 191-208, Bath, UK, September 2-5, 2025. Available here.

Articles under submission

1. Michail Fasoulakis, Evangelos Markakis, Georgios Roussakis, Christodoulos Santorinaios. Improved Last Iterate Convergence Properties for the FLBR Dynamics. Under submission to ICML 2026 (version available in Deliverable D.4.3), 2025.
2. Georgios Amanatidis, Alexandros Lolos, Evangelos Markakis, Victor Turmel. Online Fair Division for Personalized 2-Value Instances. Submitted to the Journal of Artificial Intelligence Research, JAIR, 2025. Available here.

Other technical reports and selected deliverables

• Deliverable D.3.1: Our survey on the relevant literature to ALGOLINC and current challenges and open problems can be found here.
• Deliverable D.4.3: This is a technical report with our research progress on WP4, namely on algorithms for converging to Nash equilibria in games. It can be found here.
• Deliverable D.5.2: This is a technical report with our research progress on WP5, on learning-augmented algorithms in a variety of problems. It can be found here.
• Deliverable D.6.2: An overview of our experimental evaluation of our algorithms for finding equilibria, which can be found here.

Dissemination Activities

• Participation of our research team with 2 poster presentations at the Workshop WALE 2024: Workshop on Algorithms, Learning and Economics, Kefalonia, Greece, June 17-21, 2024.
• Organization of the 19th Athens Colloquium on Algorithms and Complexity (ACAC 2024), Athens, Greece, August 27-29, 2024. This was accompanied by a presentation given by our team regarding some of our initial research results of WP4.
• Participation of the PI at the Workshop on Voting, Matching and Preference Aggregation, organized by the Institute for Mathematical Sciences, National University of Singapore, Singapore, December 9-13, 2024.
• Participation of the PI as keynote speaker at the workshop ALGA 2025: Workshop on Algorithms, Learning and Games, Sardinia, Italy, June 8-13, 2025.
• Participation of our postdoc Michail Xefteris, with oral and poster presentation at the conference ICML 2025: 42nd International Conference on Machine Learning, Vancouver, Canada, July 13-19, 2025.
• Co-organization of a workshop on Algorithmic Game Theory, Athens Univ. of Economics and Business, July 15-16, 2025.
• Participation of the PI and of our postdoc Michail Fasoulakis with oral and poster presentation at the conference IJCAI 2025: 34th International Joint Conference on Artificial Intelligence, Montreal, Canada, August 16-22, 2025.
• Participation of the PI as invited speaker and of the members Georgios Amanatidis, Panagiotis Tsamopoulos, Alviona Mantso with paper presentations, at the conference SAGT 2025: 18th International Symposium on Algorithmic Game Theory, SAGT 2025, Bath, UK, September 2-5, 2025.
• Participation of our PhD student Panagiotis Tsamopoulos with oral presentation at the workshop “Learning Augmented Algorithms”, CWI, Amsterdam, October 7-10, 2025.

Overview of our project's research