Joel Dyer
Address
Mathematical Institute
University of Oxford
Andrew Wiles Building
Radcliffe Observatory Quarter
Woodstock Road
Oxford
OX2 6GG
TA for B8.2 Continuous Martingales and Stochastic Calculus, 2022
TA for B8.1 Probability, Measure, and Martingales, 2021
Tutor for C5.4 Networks, 2021
TA for C5.4 Networks, 2020
TA for B8.5 Graph Theory, 2019
Best Proposal Paper Award, AI4ABM Workshop at ICML 2022
Alan Turing Institute Enrichment Student, 2021-2022
EPSRC InFoMM CDT Studentship, University of Oxford, 2018-2022
I am a DPhil student in Mathematics and member of Cohort 5 of the InFoMM CDT. I am affiliated with the Institute for New Economic Thinking at Oxford and supervised by Prof J Doyne Farmer. My research interests include:
Complexity economics
Opinion dynamics
Computational social science
Bayesian inference
Calibrating agent-based models to microdata with graph neural networks
Dyer, J; Cannon, P; Farmer, JD; Schmon, SM
2022 (Spotlight Talk & Best Proposal Award at the AI4ABM Workshop, ICML)
Black-box Bayesian inference for economic agent-based models
Dyer, J; Cannon, P; Farmer, JD; Schmon, SM
2022 (arxiv)
Amortised Likelihood-free Inference for Expensive Time-series Simulators with Signatured Ratio Estimation
Dyer, J; Cannon, P; Schmon, SM
International Conference on Artificial Intelligence and Statistics 25 (2022)
Approximate Bayesian Computation with Path Signatures
Dyer, J; Cannon, P; Schmon, SM
2021 (arxiv)
Deep Signature Statistics for Likelihood-free Time-series Models
Dyer, J; Cannon, P; Schmon, SM
2021 (INNF+ Workshop, ICML)
Public risk perception and emotion on Twitter during the Covid-19 pandemic
Dyer, J; Kolic, B
Applied Network Science issue 1 volume 5 (16 December 2020)