I'm a PhD student in the Department of Machine Learning within the School of Computer Science at Carnegie Mellon University. I am advised by Professor Geoffrey Gordon. My research focuses on developing statistically principled models which are also highly practical.

I'm a native of Wellington, New Zealand. I completed my Masters Degree in Computer Science with Professor Mengjie Zhang at Victoria University of Wellington. I received my Undergraduate Degree in Computer Science and Mathematics at Victoria University of Wellington. When I get the chance I spend my free time on Badminton, Ultimate Frisbee, and Ballroom Dance.

You can download my full CV here

Get in Touch
Email: cmdowney@cs.cmu.edu
Office: GHC 8007
Address: Gates Bulding, 5000 Forbes Avenue, Pittsburgh, PA, 15213

Curriculum Vitae

You can download my full CV here

Research Interests

My research focuses on developing statistically principled models which are also highly practical.

One example is our recent work on Predictive State Recurrent Neural Networks. PSRNNs are a new approach for modelling time-series data that hybridizes Bayes filters with RNNs. PSRNNs have a similar function form to widely used RNN architectures such as GRUs, however because they are derived from bayes rule they can be initialized using a consisten method of moments algorithm.

My research interests include, but are not limited to: Machine Learning, Optimization, Neural Networks, Kernel Methods, Reinforcement Learning, Information Theory, Spectral Algorithms, and Spectral Graph Theory.

Selected Publications

Carlton Downey*, Krzysztof Choromanski*, Byron Boots
Initialization Matters: Orthogonal Predictive State Recurrent Neural Networks
International Conference on Learning representations (ICLR). 2018.

Quan Wang, Carlton Downey, Li Wan, Philip Andrew Mansfield, Ignacio Lopez Moreno
Speaker Diarization with LSTM
International Conference on Acoustics, Speech and Signal Processing (ICASSP). 2018.

Ahmed Hefny, Carlton Downey, Geoff Gordon
An Efficient, Expressive and Local Minima-free Method for Learning Controlled Dynamical Systems
Association for the Advancement of Artificial Intelligence (AAAI). 2018.

Carlton Downey, Ahmed Hefny, Geoff Gordon
Predictive State Recurrent Neural Networks
Advances in Neural Information Processing Systems (NIPS). 2017.

Carlton Downey, Ahmed Hefny, Geoff Gordon
Practical Learning of Predictive State Representations
arXiv. 2017.

Ahmed Hefny, Carlton Downey, Geoff Gordon
Supervised Learning for Dynamical System Learning
Advances in Neural Information Processing Systems (NIPS). 2015.

Sashank Reddi, Ahmed Hefny, Carlton Downey, Avinava Dubey, Suvrit Sra
Large-scale randomized-coordinate descent methods with non-separable linear constraints
Association for Uncertainty in Artificial Intelligence (UAI). 2015.

Carlton Downey, Mengjie Zhang
Caching for Parallel Linear Genetic Programming
Proceeding of Genetic and Evolutionary Computation Conference. GECCO (Companion) 2011, ACM Press. pp. 201-202.

Carlton Downey, Mengjie Zhang
Execution Trace Caching for Linear Genetic Programming
Proceeding of the 2011 IEEE Congress on Evolutionary Computation. IEEE Press. New Orleans, USA. June 5-8, 2011. pp. 1191-1198.

Carlton Downey, Mengjie Zhang
Parallel Linear Genetic Programming
Proceedings of the 14th European Conference on Genetic Programming. Lecture Notes in Computer Science. Vol. 6621. Springer. Torino, Italy 2011. pp. 178-189.
(Nominated for the Best Paper Award)

Carlton Downey, Scott Sanner
Temporal Difference Bayesian Model Averaging: A Bayesian Perspective on Adapting Lambda
In Proceedings of the 27th International Conference on Machine Learning (ICML-10); Haifa, Israel

Carlton Downey, Mengjie Zhang, Will Browne
New Crossover Operators in Linear Genetic Programming for Multiclass Object Classification
Proceedings of 2010 Genetic and Evolutionary Computation Conference. Portland, USA. 2010. ACM.

Carlton Downey, Mengjie Zhang
Multiclass Object Classification for Computer Vision using Linear Genetic Programming
Proceeding of the 24th International Conference on Image and Vision Computing New Zealand. Wellington, 2009. IEEE Press. pp. 73-78.

Teaching

Carnegie Mellon University (2013-2014)
10-701: Introduction to Machine Learning
10-601: Introduction to Machine Learning

Victoria University of Wellington (2007 - 2011)
COMP 307: Introduction to Artificial Intelligence
COMP 261: Algorithms and Data Structures
COMP 303: Design and Analysis of Algorithms
COMP 202: Formal Methods of Computer Science
COMP 205: Software Design and Engineering
COMP 103: Introduction to Data Structures and Algorithms
SWEN 102: Introduction to Software Modelling
COMP 102: Introduction to Computer Program Design

An implementation of Predictive State Recurrent Neural Networks can be found here.

Get in touch

Email: cmdowney@cs.cmu.edu
Office: GHC 8007
Address: 5000 Forbes Avenue, Pittsburgh, PA, 15213