Skip to main content Skip to navigation

James Price

I am a PhD student in the MathSys II CDT programme. My project is under the supervision of Professor Colm ConnaughtonLink opens in a new window and Dr Susana GomesLink opens in a new window and in collaboration with the London Mathematical LaboratoryLink opens in a new window. My mathematical interests are random processes, algorithms that do interesting things and investigating anything that goes against my intuition.

I am currently focused on applying the concept of growth rate optimality to Reinforcement Learning problems. The most common decision making rule in Reinforcement Learning is to prefer the policy with the highest expected future discounted returns. However, in random processes where it is possible for an agent to accumulate a large amount of wealth expectation can be misleading as it is effected by low probability high impact events. The research field of Ergodicity Economics deals with this by instead preferring the policy with the highest growth rate. My research is focused on creating a Reinforcement Learning framework for Ergodicity Economics based decision making.

Research Output

Baumann, Dominik, Erfaun Noorani, James Price, Ole Peters, Colm Connaughton, and Thomas B. Schön. "Non-ergodicity in reinforcement learning: robustness via ergodicity transformations." ICML 2024 Workshop: Foundations of Reinforcement Learning and Control (2024) (linkLink opens in a new window)

Price, James, and Colm Connaughton. "Distinguishing Risk Preferences using Repeated Gambles." arXiv preprint arXiv:2308.07054 (2023). (arXiv linkLink opens in a new window)

Hulme, O., Vanhoyweghen, A., Connaughton, C., Peters, O., Steinkamp, S., Adamou, A., Baumann, D., Ginis, V., Verbruggen, B., Price, J. and Skjold, B., 2023. Reply to" The Limitations of Growth-Optimal Approaches to Decision Making Under Uncertainty". Econ Journal Watch, 20(2), pp.335-348. (journal linkLink opens in a new window)

Education

University of Bath (2014-2019): Mmath Mathematics with Industrial Placement

University of Warwick (2019-2020): MSc Mathematics of Real-World Systems

University of Warwick (2020-2024): PhD Mathematics of Real-World Systems

Past Research Projects

Heuristics for Real-time Railway Rescheduling (Summer Internship - STOR-i Lancaster University) Designed and evaluated heuristics based algorithms for minimising the overall passenger delay in a railway network. Supervised by Dr Edwin Reynolds at Lancaster University.

Estimation and Inference on Undirected Graphical Models (Masters Dissertation - University of Bath) Investigated various machine learning methods for estimating the structure of a graph whose nodes are random variables and edges represent conditional dependencies. Methods were applied to genomic data to infer which components of E.coli genomes are closely related. Supervised by Dr Sandipan Roy at the University of Bath.

The Dynamics of Wealth (MSc Group Project - University of Warwick) Explored properties of a stochastic process, based on Geometric Brownian Motion, which allows transfer of wealth between agents and how varying parameters affect long-term inequality between agents. Supervised by Prof. Colm Connaughton at the University of Warwick

Determining the Motion in a Skeletal Join from the Articular Surface Geometry (MSc Individual Project - University of Warwick) Modelling forces in a skeletal joint requires the expensive process of turning the surface of a bone into a mesh grid. In this the project bones were instead represented as a collection of unstructured points, the work focusing especially on how to estimate differential geometric properties of the bone surface. Supervised by Dr Shreyas Mandre at the University of Warwick.

Past Work Experience

Actuarial Placement Student (ArgoGlobal 2016-2017) Year-long placement as a pricing actuary at a Lloyd's of London insurance syndicate. Aspects of the role included: Developing pricing models, assisting with regulatory reporting and giving actuarial advice on the performance of larger contracts. I was also a member of the Lloyd's of London Choir.

Teaching

First Year Undergraduate Supervisor (Term 1,2&3 2020/2021, 2021/2022, 2022/2023)

MA241 Combinatorics - Teaching Assistant (Term 1 2020/2021)

MA2K3 Consolidation - Supervisor (Term 1 2021/2022)

MA482 Stochastic Analysis - Teaching Assistant (Term 2 2021/2022)

*University of Bath* MA10211 Probability & Statistics 1A - Undergraduate Tutor (Semester 1 2017/2018, 2018/2019)

*University of Bath* MA10212 Probability & Statistics 1B - Undergraduate Tutor (Semester 2 2017/2018, 2018/2019)

Talks

DR@W Forum | University of Warwick, 6 June 2024
Talk: "The Dynamics of Decision Making with Uncertain Outcomes"

Ergodicity Economics Seminar | online, 17 April 2024
Talk: "Growth Optimal Approaches for Finite-time and Value Dependent Problems"

Pre-Viva talk | University of Warwick, 4 March 2024
Talk: "The Dynamics of Decision Making with Uncertain Outcomes"

EE Summer Workshop | Lisbon, 25 July 2023
Talk: "Growth Optimal Approaches in Continuous Time"

MathSys Annual Retreat | Chester, 28 April 2023
Talk: "Estimating Ergodicity in Decision Making"

EE2023 | online, 30 January 2023
Talk: "Extending EE: Ruin and Finite-time problems"

Experimental Psychology Reading Group | University of Warwick, 3 November 2022
Talk: "Using Growth Rates for Decision Making"

SPAAM Seminar | University of Warwick, 27 October 2022
Talk: "Decision Making for Repeated Games"

EE2022 | online, 19 January 2022
Talk: "Parameter Estimation of a Reallocating Geometric Brownian Motion"

MathSys PhD Seminar | online, 6 October 2021
Talk: "Ergodicity Economics in Decision Making"

Maths in Social Sciences Reading Group | online, 13 September 2020
Talk: "Dynamics of Wealth Inequality"

Other Interests

My non-academic interests include running, reading and cooking with obscure ingredients. I also enjoy playing the clarinet (quietly) and percussion (badly).

Email: james dot price dot 2 at warwick dot ac dot uk

Google Scholar: James Price - google scholar

GitHub: JRP4 - GitHubLink opens in a new window