Workshop: Game Theory in the Era of AI
University of Birmingham - COSIMO - IDAI - Network+

Workshop: Game Theory in the Era of AI
University of Birmingham
COSIMO - IDAI - Network+
Organisers: Leonardo Stella, Jens Christian Claussen

Date: 13th-15th May 2026.
Location: Elm House, Business Park, UoB.
351 Bristol Rd, B5 7SW.
Registration: closed. Confirmation has been sent to participants via email.
Accommodation: Edgbaston Park Hotel (suggested).

Workshop Description

This workshop aims to highlight the most recent research at the intersection of game theory, optimisation, artificial intelligence (AI), multi-agent learning and systems and control, with applications to a variety of domains, including biology, economics, cyber-physical systems. Second, it aims to bring colleagues with expertise in game theory, optimisation, AI/ML, and systems and control together to consider grand challenges in the current UK and EU research landscape for networking and planning for joint funding proposals. Hence, the format of this workshop designed as a hybrid between a tutorial series and research networking event as a hybrid 3-day in-person event organised at the University of Birmingham in May 13-15, 2026 with online streaming of all talks on Zoom and additional opportunities for online contributed talks.

This workshop has received support from the EPSRC Network+ funding as well as from COSIMO and the Institute for Data and AI (IDAI) at the University of Birmingham, Elm House. To maximise the impact of the event, all invited talks will be streamed and permanently shared on this webpage.

Streaming/Online Participation

We will be streaming and recording all invited/contributed talks via Zoom to allow a wider participation (link to be provided in due course). Recordings will be posted on this page after the end of the workshop. To access the online streaming, please click on the following links:

Zoom link for Day 1 (Wednesday)
Zoom link for Day 2 (Thursday)
Zoom link for Day 3 (Friday)

Programme

The final programme for the workshop is summarised below. Invited speakers and contributed talks can be found below.

Invited Talks and Biographies

Prof. Tamer Başar

Tamer Başar (Life Fellow, IEEE) received the B.S.E.E. degree from Robert College, Istanbul, Turkiye, in 1969, and the M.S., M.Phil., and Ph.D. degrees in engineering and applied science from Yale University, New Haven, CT, USA, in 1970, 1971, and 1972, respectively. He has been with the University of Illinois Urbana-Champaign since 1981, where he is currently Swanlund Endowed Chair Emeritus and Center for Advanced Study (CAS) Professor Emeritus of Electrical and Computer Engineering, with also affiliations with the Coordinated Science Laboratory, Information Trust Institute, and Mechanical Science and Engineering. At Illinois, he has also served as Director of CAS (2014–2020), Interim Dean of Engineering (2018), and Interim Director of the Beckman Institute (2008–2010). His current research interests include stochastic teams, games, and networks; risk-sensitive estimation and control; mean-field game theory; multiagent systems and learning; data-driven distributed optimization; epidemics modeling and control over networks; strategic information transmission, spread of disinformation, and deception; security and trust; energy systems; and cyber-physical systems. Dr. Başar received the Wilbur Cross Medal in 2021.

He is a Member of the US National Academy of Engineering and a Fellow of the American Academy of Arts and Sciences, as well as Fellow of IEEE, IFAC, and SIAM. He has served as the President of IEEE Control Systems Society (CSS), International Society of Dynamic Games (ISDG), and American Automatic Control Council (AACC). He has received several awards and recognitions over the years, including the highest awards of IEEE CSS, IFAC, AACC, and ISDG, the IEEE Control Systems Award, and a number of international honorary doctorates and professorships. He has over 1000 publications in systems, control, communications, optimization, networks, and dynamic games, including books on noncooperative dynamic game theory, robust control, network security, wireless and communication networks, and stochastic networked control. He was the Editor-in-Chief of Automatica between 2004 and 2014, and is currently the editor of several book series.

Title: RL and Equilibria for Multi-Agent Dynamical Systems in the High-Population Regime.

Abstract: Decision making in dynamic uncertain environments with multiple agents arises in many disciplines and application domains, including control, communications, distributed optimization, social networks, and economics. Here a natural framework, and a comprehensive one, for modeling, optimization, and analysis is the one provided by stochastic dynamic games (SDGs), which accommodates different solution concepts depending on how the interactions among the agents are modeled, particularly whether they are in a cooperative mode (with the same objective functions, as in teams) or in a noncooperative mode (with different objective functions) or a mix of the two, such as teams of agents interacting noncooperatively across different teams (and of course cooperatively within each team). What also affects (strategic) interactions among the agents is the asymmetric nature of the information different agents acquire (and do not share or only partially share (selectively) with others, even within teams). What makes such problems even more challenging in a dynamic environment with networked agents is the dependence of the information available to one agent at some point in time on the policies or decisions of other agents who have already acted at earlier instants of time. Such decision problems, initially studied in a team framework, are known as those with nonclassical information where optimal policies of team agents must be designed to balance a tradeoff between contribution to optimality of the team objective function and signaling through their actions useful information to other agents in their neighborhood who would be acting after them. Existence of such a tradeoff between signaling and optimization creates even more challenging issues in SDGs with mis-aligned objectives among at least a subset of agents, which however can be addressed effectively for a specially structured subclass of such games, namely mean-field games.

This talk will provide an overview of the landscape above, first for a general class of stochastic dynamic teams and games, and then for a subclass where the objective functions are quadratic, and the interaction relationships are linear. The talk will also cover reinforcement learning embedded into policy development when agents do not have precise information on the underlying models.

Prof. Dario Bauso

Dario Bauso has received the Laurea degree in Aeronautical Engineering in 2000 and the Ph.D. degree in Automatic Control and System Theory in 2004 from the University of Palermo, Italy. Since 2018 he has been with the Jan C. Willems Center for Systems and Control, ENTEG, Faculty of Science and Engineering, University of Groningen (The Netherlands), where he is currently Full Professor and Chair of Operations Research for Engineering Systems. Since 2005 he has also been with the Dipartimento di Ingegneria, University of Palermo (Italy). Since 2018 he has been a guest professor at Keio University, Japan. His research interests are in the field of Optimization, Optimal and Distributed Control, and Game Theory. Bauso was an Associate Editor of IEEE Transactions on Automatic Control from 2011 to 2016, of IFAC Automatica from 2015 to 2021, of IEEE Control Systems Letters from 2016 to 2021, of Dynamic Games and Applications from 2011 to 2022, and is Associate Editor of Journal of Dynamics and Games since 2019.

Title: Stability and Transparency in GANs: Synthesis of Neonatal Necrotizing Enterocolitis (NEC) Data.

Abstract: The deployment of Generative Adversarial Networks (GANs) in clinical settings—specifically for the synthesis of rare disease data such as Neonatal Necrotizing Enterocolitis (NEC)—presents unique challenges in privacy preservation, robustness, and performance. In the medical domain, strict legal and ethical requirements demand that synthetic data generation be not only high-fidelity but also numerically stable and theoretically grounded. These constraints serve as a critical testbed, stimulating new scientific questions regarding the fundamental convergence properties of adversarial learning. In this talk, we address the persistent issue of training instabilities—often manifesting as oscillations in parameter space—which can compromise the reliability of synthetic medical cohorts. We derive a theoretical framework to quantify these dynamics by linearizing the GAN Jacobian near equilibrium, identifying two distinct interaction regimes: a latent-mediated weight interaction and a direct bias-bias coupling. By reducing the high-dimensional game to a system of coupled "adversarial springs", we obtain closed-form expressions for the natural frequency of the system, ω. Our analysis reveals that this frequency scales with the geometric mean of the learning rates (ηD ηG)^(1/2) and is modulated by the generator activation sensitivity G(1-G). We validate these predictions through numerical experiments, demonstrating that bias-bias interactions dominate the observed fluctuations. Spectral analysis confirms our frequency scaling law with high precision, providing a robust metric for predicting limit-cycle behavior and ensuring the stability necessary for sensitive clinical applications.

Dr. Galit Ashkenazi-Golan

Galit Ashkenazi-Golan's research focus is in Game Theory, the mathematical modelling of strategic interactions. In particular, she is interested in dynamic games (such as repeated games, stochastic games, or Borel Games), and often in the effect that information has on the equilibrium of a game. She is also interested in related topics such as social learning, Markov Decision Processes and Computational Game Theory. She joined LSE in 2021 as an Assistant Professor. Before this, she was a research affiliate in the School of Mathematical Sciences of Tel-Aviv University. She completed her PhD under the supervision of Ehud Lehrer from Tel-Aviv University, though she spent several years of her PhD as a visiting scholar at the École Polytechnique in Paris.

Title: The Bounds for Algorithmic Collusion: Q-Learning, Gradient Learning and the Folk Theorem.

Abstract: We explore the behaviour emerging from learning agents repeatedly interacting strategically for a wide range of learning dynamics, including Q-learning, projected gradient, replicator and log-barrier dynamics. Going beyond the better understood classes of potential games and zero-sum games, we consider the setting of a general repeated game with finite recall under different forms of monitoring. We obtain a Folk Theorem-style result and characterise the set of payoff vectors that can be obtained by these dynamics, discovering a wide range of possibilities for the emergence of algorithmic collusion. Achieving this requires a novel technical approach, which, to the best of our knowledge, yields the first convergence result for multi-agent Q-learning algorithms in repeated games.

This work is joint work with Domenico Mergoni Cecchelli, Edward Plumb, and Clemens Possnig.

Prof. Manh Hong Duong

Hong Duong is a Professor of Mathematics. Hong Duong’s research interests span a wide range of topics in the intersections of analysis, applied probability, and computational mathematics including partial differential equations, interacting particle systems, non-equilibrium thermodynamics, and evolutionary game theory. Most of his research are inspired from applications in statistical physics and biological/social/material sciences. His research has been supported from the ITN Fronts and Interfaces in Science and Technology (EU), the NWO (Netherlands), the London Mathematical Society (UK) and the EPSRC (UK). Hong Duong's research interests span a wide range of topics in the intersections of analysis, applied probability and computational mathematics. Most of his research are inspired from applications in statistical physics and biological/social/material sciences.

Title: Evolutionary Game Theory, Evolution of Cooperation and Institutional Incentives.

Abstract: Darwin's theory of evolution by natural selection, also known as “survival of the fittest”, implies that evolution is based on a fierce competition between individuals and should therefore only reward selfish behavior. However, cooperation occurs at all levels of biological organizations, from cellular clusters to bees to humans. Cooperation is in fact needed for evolution to construct new levels of organization. The problem of promoting cooperative behaviour within populations of self-regarding individuals has been intensively investigated across diverse fields of behavioural, social and computational sciences. Evolutionary Game Theory provides a powerful mathematical framework for understanding this key challenge over the last 50 years. Under this framework, several mechanisms for promoting the evolution of cooperation have been identified, including kin selection, direct reciprocity, indirect reciprocity, network reciprocity, group selection and different forms of incentives.

In this talk, I will discuss these topics and present our recent research on promoting cooperation via institutional incentives. We will show how this practical problem can be formulated into mathematically constrained, multi-objective optimization problems, where one wishes to minimize the cost of providing incentives while ensuring a minimum level of cooperation, sustained over time.

Contributed Talks

We thank all participants for signing up and contribute a talk to the workshop. All contributed talks are now confirmed and listed below, including bio of the speaker and title/abstract. All talks will be streamed on Zoom and the recording will be shared on this page after the end of the workshop.

Dr. Pablo R. Baldivieso-Monasterios

Pablo R. Baldivieso-Monasterios is a Lecturer in the School of Electrical and Electronic Engineering, University of Sheffield, UK. He received a PhD in robust distributed model predictive control from the University of Sheffield, UK 2018. His research interests include robust and distributed model predictive and optimisation-based control and game-theoretic methods for control and smart grids.

Title: Coalition-Based Demand-Side Management in a Micro-Grid with Multiple Electricity Retailers.

Abstract: This talk presents a game-theoretic framework for demand-side management based on coalition formation. The approach combines cooperative and non-cooperative game models to capture both the stability and the emergence of consumer groups. A cooperative game is used to assess the stability of a given partition, providing conditions under which coalitions are beneficial and resilient to deviations. Coalition formation itself is modelled through a non-cooperative game with an underlying potential structure, so that coalitions evolve along finite improvement paths towards stable configurations. The talk highlights how this hybrid perspective offers a natural framework for dynamic pricing and coordinated demand response in future power systems.

Dr. Masoumeh Iran Mansouri

Masoumeh Iran Mansouri is an Associate Professor in the School of Computer Science at the University of Birmingham, UK. Previously, she was a researcher at the Centre for Applied Autonomous Sensor Systems at Örebro University. She has also been a visiting researcher at the Oxford Robotics Institute and Sven Koenig’s lab at the University of Southern California. Her research primarily focuses on AI planning methods for robotics, particularly on integrating robot task and motion planning to enable robots to make autonomous or semi-autonomous decisions in unstructured environments shared with humans. Her research also extends to cultural robotics, a field dedicated to studying the effects of integrating cultural models into robots at the intersection of cultural theory and robotics. She is also the co-founder of the Critical Cultural Robotics Network and the Contentious Politics of AI, both based at the University of Birmingham.

Title: Robust Decision Making in Multi-Robot Systems.

Abstract: This talk explores several principles of robust decision-making in multi-robot systems, focusing on how autonomous robots can operate reliably in complex, uncertain, and dynamic environments. In particular, I will cover my team's research on the planning and coordination of nonholonomic robots, both within and without formation, ranging from physically coupled transportation to non-physically coupled cooperation, with applications such as cooperative transportation, assembly robotics, and visual tracking, to name a few.

Prof. Per Kristian Lehre

Per Kristian Lehre a Professor in the School of Computer Science at the University of Birmingham. Prof Lehre's research interests are in theoretical aspects of nature-inspired search heuristics, in particular runtime analysis of population-based evolutionary algorithms. He is associate editor of ACM Transactions on Evolution and Learning, editorial board member of Evolutionary Computation Journal, and former associate editor of IEEE Transactions on Evolutionary Computation. Most recently, he was a Turing AI Acceleration Fellow funded by the UKRI with a project on Runtime Analysis of Co-Evolutionary Algorithms.

Title: Runtime Analysis of Co-evolutionary Algorithms.

Abstract: Co-evolutionary algorithms (CoEAs) model evolutionary arms races between populations of competing agents. They have diverse applications, ranging from learning game-playing strategies to designing sorting networks and automated software testing. CoEAs can be particularly effective in game-theoretic scenarios where the strategy space is so vast that computing the full payoff matrix is computationally intractable. Our focus is on the conditions under which these algorithms provably find Nash equilibria efficiently, in particular logarithmic in the number of strategies. We present recent theoretical results that establish these runtime bounds for selected algorithms, including PDCoEA.

Session 1 (13th): ECR Short Talks

This session is structured as an online seminar series of short talks (30 minutes + 10 minutes for questions) from colleagues working on topics related to game theory, joining us from outside the UK. The session will be streamed on Zoom and the recording, title/abstract will be shared on this page before and after the workshop.

Dr. Barbara Franci

Barbara Franci has a Master's Degree in Mathematics from Università degli Studi di Siena and she received her PhD degree in Pure and Applied Mathematics from Università di Torino - Politecnico di Torino (joint program) in 2018. Then, she was a Post Doctoral researcher at DCSC (Delft Center for Systems and Control) in Technische Universiteit Delft until 2021. After that, she moved to Maastricht University where she was Assistant Professor at the Department of Advanced Computing Sciences till April 2025. Currently, she is Assistant Professor at Politecnico di Torino in the Department of Mathematical Sciences, being the winner of 'Programma Rita Levi Montalcini'. Her research topics are mostly related to Game Theory but she also interested in opinion dynamics and the wisdom of crowds.

Title: GANs and Games - Training Generative Adversarial Networks via Operator Theory.

Abstract: Generative adversarial networks (GANs) are a class of generative models with two antagonistic neural networks: a generator and a discriminator. These two neural networks compete against each other, often with the aim of generating realistic images. Because of this competition, the training process can be modelled as a stochastic Nash equilibrium seeking problem. Starting from this game, this talk showcases the problems of GANs training and the possible tricks, borrowed from game theory and operator theory, that can be used to address them. In particular, it will cover possible training mechanism involving stochastic approximation and regularization steps that ensure convergence (under suitable assumptions) to the equilibrium and therefore allow to obtain nicely generated images.

Dr. Emilio Benenati

Emilio Benenati is a post-doctoral researcher at KTH Stockholm, Sweden. He received his Ph.D. degree in 2025 from the Delft Center for Systems and Control, TU Delft, The Netherlands, his Master's degree in 2019 from ETH Zürich, Switzerland, and his Bachelor's degree in 2016 from the University of Catania, Italy. In 2019-2020, he held a research position at the Italian Institute of Technology in Genova, Italy. He visited the Colorado University at Boulder, in 2019, and the University of California, Santa Barbara, in 2024. His research interests include game theoretic control problems for the real-time control of complex interactive systems, with application to traffic control and multi-agent systems.

Title: Game-theoretic Model Predictive Control: Design and Computational Methods.

Abstract: Some control problems, which emerge for example in vehicle traffic routing, autonomous driving, and markets clearing, are characterized, by interactions between multiple autonomous agents with coupled dynamics and non-aligned control objectives. Game-theoretic model predictive control addresses such systems by determining, at each time-step, the control action as the solution to a finite dynamic game, namely, a Nash equilibrium. This approach enables the development of interaction-aware autonomous systems.

In this talk, I will show a novel design method of the agents' terminal objectives that guarantees stability and agent-wise infinite-horizon optimality of the nominal control action. I will then present two computational contributions: an algorithmic method for offline derivation of the complete state-to-control map, and an active-set refinement method for the Douglas-Rachford fixed-point algorithm. Both significantly reduce computation time, enabling high-frequency control sampling and improving on current methods by several orders of magnitude.

Dr. Julian Barreiro-Gomez

Julian Barreiro-Gomez is an Assistant Professor in the Department of Computer and Information Engineering (CIE), College of Computing and Mathematical Sciences at Khalifa University, Abu Dhabi, UAE. He is the Internship Coordinator in his Department (CIE), and actively serves in the Undergraduate Curriculum Committee, the Faculty Recruitment Committee, Research & Outreach Committee, Program Improvement Committee, and the College Seminar Committee. He received the Best Ph.D. Thesis in Europe in the Field of Control for Complex and Heterogeneous Systems 2017 awarded by the European Embedded Control Institute (EECI), and the Best Ph.D. Thesis in Control Engineering 2017 award from the Spanish National Committee of Automatic Control (CEA) and Springer. Jointly with his Ph.D. thesis advisors, he received the ISA Transactions (top 3% journal Scopus) Best Paper Award 2018 in Recognition to the best paper published in the previous year. Since 2018, he is a Senior Member of the Institute of Electrical and Electronics Engineers (IEEE). He is author of the book "The Role of Population Games in the Design of Optimization-Based Controllers: A Large-scale Insight" published by Springer Cham in 2019, and co-author of the book "Mean-Field-Type Games for Engineers" published by CRC Press in 2021. His main research interests are: optimal control, dynamic games, machine learning, and engineering applications with emphasis on multi-agent systems. Julian lives in Abu Dhabi with his wife and son, and he enjoys spending time with his family, having a picnic during the nice weather, and expressing gratitude to Almighty God for the miracle of health and life

Title: On the Interplay Between Game Theory and Machine Learning.

Abstract: This talk explores two complementary perspectives on the interplay between machine learning and game theory. First, we consider machine learning in game theory through a crowd dynamics model based on game-theoretic principles, closely related to the Hughes model. We discuss how learning-based methods can be leveraged to compute solutions to evacuation problems arising in this setting. Second, we investigate game theory in machine learning by establishing a novel connection between a class of recurrent neural networks and evolutionary dynamics in population games. In particular, we show that, in classification tasks, the output of a recurrent neural network coincides with the evolution of the population state under suitable game-theoretic dynamics.

The first part of this talk is joint work with Vicente Vargas-Panesso, Yuepeng Wang, and Xun Shen. The second part is joint work with Jorge I. Poveda.

Session 2 (14th): Network+ UK Funding

This session is designed to allow networking among the participants of the workshop, with opportunities for discussion, breakout rooms and collaboration. Jon Kudge will give an overview of the UK funding opportunities, including EPSRC, UKRI and ARIA, relevant to the audience of the workshop. Here is an overview of the session:

1:00pm-1:30pm - Overview of EPSRC, UKRI and ARIA funding opportunities, as well as information about the Research Professional system at UoB (Jon Kudge, Leonardo Stella).
1:30pm-1:45pm - Overview of SHARE, an EPSRC strategic infrastructure grant application that wants to build an environment for human/embodied AI/robots interaction for research purposes (Max Di Luca).
2:00pm-2:30pm - Overview of funding opportunities for early career researchers, including fellowships (Jon Kudge).

Session 3 (15th): PhD students Lightning Talks

This session is inspired by the yearly three-minute thesis competition (3MT) that is held in many universities worldwide, where PhD students present their thesis in 3 minutes with the aid of a single slide. In this session, colleagues, postdocs and PhD students will give an overview of their research in 10/20-minute short talks. Here is a list of the participants that volunteered to present and the title of their talks:

Dr. Jens Christian Claussen, "Evolutionary dynamics in finite populations: Drift reversals in cyclic games".
Tao Shan, "Ensuring Semantic Consistency in Multi-agent systems via DDD".
Spyridon Giagtzoglou, "Asymmetric Regularization for GAN Training via Operator Theory".
Aju Ani Justus, "Inter-agent Communication in Multi Agent Reinforcement Learning".
Shanshan Mao, "Multi-agent Simulation Perspective".
Seyedehnegar Seyedmonir, "Lyapunov-based Policy Synthesis for Interval MDPs".
Calina Durbac, "Nontrivial Effects of the Lotka-Volterra Replicator Dynamics Mapping in Random Interactions".

Workshop Organisers

Dr. Leonardo Stella

Assistant Professor, University of Birmingham. He obtained his PhD at the University of Sheffield (UK) in automatic control and systems engineering. His work focuses on game theory, systems and control, multi-agent learning, and optimisation with applications in bio-chemical and cyber-physical systems. He received £630k (Principal Investigator with Dr. Mirco Giacobbe) in funding from ARIA to work on safeguarded AI in biopharmaceutical manufacturing with collaborators from AstraZeneca. Recently, the project has been selected for extension with a budget of nearly £1M until November 2027.

Dr. Jens Christian Claussen

Associate Professor, University of Birmingham. He obtained his Dr. rer. nat. (PhD) in Theoretical Physics at the Christian-Albrechts-University Kiel (Germany) in control of delayed-measured nonlinear systems, and received a second doctorate (Dr. habil.) from research contributions in dynamical systems, cellular automata, and evolutionary game theory. He was leading (as chair and co-chair) the Division of Physics of Socio-Economic Systems (SOE) of the German Physical Society (DPG) and (co)organizing the annual conferences 2012-2022. He received recent EPSRC New Horizon 200k funding on Evolving Networks Towards Complexity.

Organisation Committee

Dr Leonardo Stella - Assistant Professor (l.stella@bham.ac.uk).
Dr Jens Christian Claussen - Associate Professor (j.c.claussen@bham.ac.uk).
Mr Ziyue Chu (zxc332@student.bham.ac.uk).
Miss Suzannah Gebbett - PhD student (sxg179@student.bham.ac.uk).
Mr Tuo Zhang - PhD student (txz257@student.bham.ac.uk).

For any inquiries, feel free to reach out to the organisation committee for more info.

Page updated

Google Sites

Report abuse

Workshop: Game Theory in the Era of AI University of Birmingham - COSIMO - IDAI - Network+

Workshop: Game Theory in the Era of AI University of Birmingham COSIMO - IDAI - Network+Organisers: Leonardo Stella, Jens Christian Claussen

Date: 13th-15th May 2026.Location: Elm House, Business Park, UoB.351 Bristol Rd, B5 7SW.Registration: closed. Confirmation has been sent to participants via email.Accommodation: Edgbaston Park Hotel (suggested).