Charlie Hepburn
Summary
I am a PhD student at the Mathematics for Real-World Systems (Mathsys) CDT under the supervision of Prof. Giovanni MontanaLink opens in a new window. I am interested in data-driven reinforcement learning, where insights from a fixed dataset can be used to supplement the learning process. This includes reinforcement learning with demonstrations, where expert trajectories are used to guide the agent increasing sample efficiency. At the extreme, this is offline (or batch) reinforcement learning, where online interaction with the environment is prohibited and the agent must learn an optimal policy solely from a static sub-optimal dataset.
Publications
Hepburn, C.A. and Montana, G. Model-based trajectory stitching for improved behavioural cloning and its applications. Machine Learning (2023) (PublicationLink opens in a new window) (ArxivLink opens in a new window)
Hepburn, C.A. and Montana, G. Model-based Trajectory Stitching for Improved Offline Reinforcement Learning. 3rd Offline RL Workshop at Neural Information Processing Systems (2022) (PublicationLink opens in a new window) (ArxivLink opens in a new window)
Education
2021 - Present: PhD | University of Warwick
Supervised by Prof. Giovanni Montana.
2020 - 2021: MSc in Mathematics of Systems (Distinction) | University of Warwick
Individual Project: A Critical Analysis of Selected Offline Reinforcement Learning Algorithms. Supervised by: Prof. Giovanni Montana.
Group Project: Modelling substantia-nigra neurons to quantify the effects of alpha-synuclein in Parkinson’s disease.
2016 - 2020: MMath Mathematics (First Class Hons.) | University of Edinburgh
Dissertation: Randomized Iterative Methods for Linear Systems. Supervised by Dr. Aretha Teckentrup.
Group Project: Approximate Bayesian Computation.
Contact
Office: B1.29, Zeeman Building