Dr. Iain Dunning

Engineer and researcher, with a focus on applying techniques from machine learning and optimization to solve difficult decision problems - plan a logistics network, run a power grid, or solve a video game.
Currently living in Manhattan, NY, 🇺🇸.
Contact: iaindunning 📧 gmail.


Recent Work History

Hudson River Trading

Algorithm Developer (August '18 —)
Researching and developing automated trading algorithms using advanced mathematical techniques.

DeepMind Technologies

Senior Research Engineer (July '16 — July '18)
Applying large-scale artificial intelligence techniques like deep reinforcement learning to complex environments, using cutting-edge deep learning hardware. Team lead for five engineers, tech lead for multiagent research engineering.

Massachusetts Institute of Technology

Teaching & Research Assistant (Sep. '11 — May '16)
See research below. Co-created X0,000-person EdX class The Analytics Edge. Taught MBA and executive MBA residential versions of the class at MIT Sloan School of Management.

Google

Decision Support Engineer intern (June 13 — Aug. '13)
Worked on search engine crawler. Designed & implemented algorithms to improve the crawl prioritization, analyzed impacts on very large (O(1010) rows, O(103) TB) datasets with MapReduce/Flume.


Research

Learning Fast Optimizers for Contextual Stochastic Integer Programs
V. Nair, K. Dvijotham, I. Dunning, O. Vinyals.
UAI, 2018.
[paper]

Human-level performance in first-person multiplayer games with population-based deep reinforcement learning
M. Jaderberg*, W. Czarnecki*, I. Dunning*, L. Marris, G. Lever, A. García Castañeda, C. Beattie, N. Rabinowitz, A. Morcos, A. Ruderman, N. Sonnerat, T. Green, L. Deason, J. Z. Leibo, D. Silver, D. Hassabis, K. Kavukcuoglu, T. Graepel.
Submitted, 2018.
[arXiv] [video] [DeepMind blog]

Inequity aversion resolves intertemporal social dilemmas
E. Hughes*, J. Z. Leibo*, M. Philips, K. Tuyls, E. Duéñez-Guzmán, A. García Castañeda, I. Dunning, T. Zhu, K. McKee, R. Koster, H. Roff, T. Graepel.
Submitted, 2018.
[arXiv]

IMPALA: Scalable distributed deep-RL with importance weighted actor-learner architectures
L. Espeholt*, H. Soyer*, R. Munos*, K. Simonyan, V. Mnih, T. Ward, Y. Doron, V. Firoiu, T. Harley, I. Dunning, S. Legg, K. Kavukcuoglu.
ICML, 2018.
[arXiv] [DeepMind blog]

Population based training of neural networks
M. Jaderberg, V. Dalibard, S. Osindero, W. Czarnecki, J. Donahue, A. Razavi, O. Vinyals, T. Green, I. Dunning, K. Simonyan, C. Fernando, K. Kavukcuoglu.
DeepMind tech report, 2017.
[arXiv] [DeepMind blog]

Extended formulations in mixed integer conic quadratic programming
J. P. Vielma*, I. Dunning, J. Huchette, M. Lubin.
Mathematical Programming Computation, 2017.
[arXiv]

JuMP: A modeling language for mathematical optimization
I. Dunning*, J. Huchette*, M. Lubin*.
SIAM Review, 2017.
[arXiv] [JuMP package]

Multistage robust mixed-integer optimization with adaptive partitions
D. Bertsimas, I. Dunning*.
Operations Research, 2016.
[Optimization Online]

Reformulation versus cutting-planes for robust optimization
D. Bertsimas, I. Dunning*, M. Lubin*.
Computational Management Science, 2016.
[Optimization Online]

Computing in operations research using Julia
M. Lubin, I. Dunning.
INFORMS Journal on Computing, 2015.
[arXiv]

See Google Scholar for other references.
Some papers use alphabetical ordering - asterix indicates "first" author.


Education

Massachusetts Institute of Technology

Ph.D., Operations Research (Sep. '11 — May '16)
At the MIT Operations Research Center, supervised by Prof. Dimitris Bertsimas

University of Auckland

B.E.(Hons), Engineering Science (Mar. '07 — Dec. '10)