Computer systems regularly act for us, from the autofocus of our cameras to the autonomous high-frequency trading algorithms of financial markets. In addition, recent advances in online market places and digital currencies (such as Bitcoin) enable automated trading and negotiation — but this requires deep reasoning and machine learning, especially about future uncertainties and adversarial opponents.
Our Intelligent Systems group applies a range of tools, from economic risk models to evolutionary game theory and multi-agent learning, in order to understand the uncertainties and interaction dynamics, especially in settings with both competitive and cooperative elements. We pay specific attention to the information that is available to each participant a priori or by observation. We then develop algorithms to exploit partial and uncertain information signals for autonomous planning, execution and strategy improvement.
The outcomes of this research include risk estimation algorithms for energy markets, coordination mechanisms for establishing fair and efficient outcomes, and algorithms that are able to interactively represent end-users by learning from experience and feedback.
Digital market places are highlighted in the national research and innovation agenda route Smart Industry and the Digital Single Market strategy adopted by the European Commission in 2015.
Contact person: Michael Kaisers
Research group: Intelligent Autonomous Systems (IAS)
Research partners: Spectral Utilities, Greenchoice, Engineering Ingegneria Informatica S.p.A., SEITA Energy Flexibility B.V., University of Southampton, University of Liverpool