Agents, interaction and complexity

About

We are investigating the theoretical underpinnings of safe, verifiable, and trustworthy autonomous systems and their use in addressing real-world challenges from smart cities to disaster response. Our research addresses models of safety, responsibility and optimisation for AI systems, reinforcement learning, game theory and negotiation, mechanism design, reasoning and learning under uncertainty, and ethical and responsible Artificial Intelligence.

Exciting research challenges around human-in-the-loop interaction include exploring how to build better human-in-the-loop AI systems which use Artificial Intelligence to support and augment human performance, both enhancing interaction experiences for humans and utilizing human expertise to enhance AI performance.

Our research includes:

• human-system/robot interaction,
• natural language processing,
• machine listening,
• human-agent collaboration
• citizen centric AI.

How systems evolve, learn and adapt enable us to better understand complex networks of individual actors. Our research in this area spans evolutionary computation and evolutionary biology, complex economic systems and social networks, and complex multi-robot systems.