Learning To Cooperate: Representation Learning for Social Decision Making (PennyWU20PSY)

Classic strategic games defined in game theory and neuroeconomics allow for mathematical and neuroscientific studies of the tensions, trade-offs and computational bases of decisions as to whether to cooperate or compete. In most of these games, however, agents have a very limited set of actions to choose from. In more realistic settings, people often have a very large number of ways of interacting with one another – even if people would like to cooperate it may not be clear how to. In this project we will develop agent models that embody Representation Learning methods from the field of machine learning. These agents reduce the large product space of actions to a lower-dimensional latent space. We hypothesise that it is the use of this latent space that allows people to efficiently learn how to best interact with one another across multiple related contexts. This hypothesis will be tested using computational models and behavioural experiments. If the project goes well we will also use brain imaging experiments to find out which neural circuits support the computations underlying these behaviours.

For more information on the supervisor for this project, please go here: https://people.uea.ac.uk/w_penny

This is a PhD programme.

The start date of the project is October 2020.

The mode of study is full-time/part-time. The studentship length is 3 years for full-time students and 6 years for part-time students.