Please note that this project is self-funded.
This project will aim to build on research by exploring novel reinforcement learning algorithms and/or other techniques from machine learning (such as deep learning).
The prisoner's dilemma continues to be a prime area of research used to understand emergent social behaviour as well as biologic behaviour such as tumour growth. This research project aims to contribute to this study by applying modern research techniques from other fields. For example, recent work within the group has shown the potential for self-recognition mechanisms to evolve using realistic reinforcement learning algorithms.
Project aims and methods
The scope of the project is not limited to the prisoner's dilemma but will almost certainly make use of a lot of computational expertise. Candidates willing to increase their skills in research software engineering should find this particularly interesting. Furthermore, there is scope to apply analytical game theoretic method to not only understand the potential for training mechanism but find bounds on performance as well as optimise certain micro behaviours.
This work has the potential to impact not only our understanding of social/biological behaviour but also develop novel training methodologies with applications to machine learning and artificial intelligence.
I gael gwybodaeth am strwythur y rhaglen, gofynion mynediad a sut i wneud cais ewch i’r rhaglen Mathemateg.Gweld y Rhaglen