Multiple complimentary approaches are available for modelling the adaptive behaviour of individual agents in complex systems, and in this work reinforcement learning is the focus. A core problem here is that unambiguous identification of rewards driving the behaviour of entities operating in complex (open-ended) real-world environments is at least difficult, if not impossible. In part this is because the true goals of agents are not observable; also, reward-driven behaviours emerge endogenously over longer timescales and are dynamically updated as environments change. Defining a reliable reward function to use in models therefore remains a challenge. Reproducing the emergence of rewards is a potential solution, and would be have application in many domains. Simulation experiments will be described which assess a candidate algorithm for the dynamic updating of rewards, RULE: Reward Updating through Learning and Expectation. The approach is tested in a simplified ecosystem-like setting where manipulated conditions challenge the survival of an entity population, calling for significant behavioural change.
About the speaker
Richard Bailey is Professor of Environmental Systems at the University of Oxford. His academic life in began in Earth Sciences and he holds a PhD from London University in solid state physics, with applications to long-term environmental change. Over many years, Richard developed a fascination with complex systems and eventually moved out of physics and in to the modelling of large scale environmental systems, with applications first in ecology and more recently in coupled human-environmental systems. He is very much interested in theoretical issues, but also tries to be useful in helping solve significant environmental problems. His applied work over the last 10 years has focused on fisheries and ocean ecology, plastic pollution, and; agriculture.
Richard works with various international bodies: UN on plastics; governments, NGOs, charities on plastic and on oceans; a small number of US start-ups on AI/ML applications for environmental solutions.
Please register to attend online or in person by filling in the form at https://app.onlinesurveys.jisc.ac.uk/s/oxford/seminar-registration-complex-open-ended-environment
This event will be in hybrid format but will not be recorded. For more information, please email complexity@inet.ox.ac.uk or call 01865 610403.