For an up-to-date list see my Google Scholar page
Papers
- Distributed Second-order Optimization using Kronecker-Factored Approximations
Jimmy Ba, Roger Grosse, James Martens
In International Conference on Learning Representations (ICLR), 2017
[PDF] - A Kronecker-factored approximate Fisher matrix for convolution layers
Roger Grosse, James Martens
In Proceedings of the 33rd International Conference on Machine Learning (ICML), 2016
[arXiv] - Adding Gradient Noise Improves Learning for Very Deep Networks
Arvind Neelakantan, Luke Vilnis, Quoc V. Le, Ilya Sutskever, Lukasz Kaiser, Karol Kurach, James Martens
[arXiv] - Optimizing Neural Networks with Kronecker-factored Approximate Curvature
James Martens, Roger Grosse
In Proceedings of the 32nd International Conference on Machine Learning (ICML), 2015
[Full version] [ICML version] [Appendices for ICML version] [Code] - New insights and perspectives on the natural gradient method
James Martens
[arXiv] - On the Expressive Efficiency of Sum Product Networks
James Martens, Venkatesh Medabalimi
[arXiv] - On the Representational Efficiency of Restricted Boltzmann Machines
James Martens, Arkadev Chattopadhyay, Toniann Pitassi, Richard Zemel
In Proceedings of the 27th Annual Conference Neural Information Processing Systems (NIPS), 2013
[PDF] - On the importance of momentum and initialization in deep learning
Ilya Sutskever, James Martens, George Dahl, and Geoffery Hinton
In Proceedings of the 30th International Conference on Machine Learning (ICML), 2013
[PDF] - Training Deep and Recurrent Neural Networks with Hessian-Free Optimization
James Martens, Ilya Sutskever
In Neural Networks: Tricks of the Trade, 2012
[PDF] [Publisher link (out of date version)] - Estimating the Hessian by Back-propagating Curvature
James Martens, Ilya Sutskever, and Kevin Swersky
In Proceedings of the 29th International Conference on Machine Learning (ICML), 2012
[PDF] [Supplement/appendix] - Learning Recurrent Neural Networks with Hessian-Free Optimization
James Martens, Ilya Sutskever
In Proceedings of the 28th International Conference on Machine Learning (ICML), 2011
[PDF] [Supplement] [Video] - Generating Text with Recurrent Neural Networks
Ilya Sutskever, James Martens, Geoffrey Hinton
In Proceedings of the 28th International Conference on Machine Learning (ICML), 2011
[PDF] - Normalization for probabilistic inference with neurons
Chris Eliasmith, James Martens
In Biological Cybernetics, 2011
- Deep Learning via Hessian-free Optimization
James Martens
In Proceedings of the 27th International Conference on Machine Learning (ICML), 2010
[PDF] [Slides] [Code] - Learning the Linear Dynamical System with ASOS
James Martens
In Proceedings of the 27th International Conference on Machine Learning (ICML), 2010
[PDF] [Supplement and code] [Slides] - Parallelizable Sampling of Markov Random Fields
James Martens, Ilya Sutskever
In Proceedings of Artificial Intelligence and Statistics (AISTATS), 2010
[PDF] - Novel Lead Configurations for Robust Bio-Impedance Acquisition
Joel Ironstone, Milan Graovac, James Martens, Martin Rozee, K.C. Smith
In Proceedings of the 30th Annual International Conference of the IEEE Engineering in Medicine and Biology Society (EMBC), 2007
Theses
- Second-order Optimization for Neural Networks
James Martens
Ph.D. Thesis, Dept. of Computer Science, University of Toronto
[PDF] - A New Algorithm for Rapid Parameter Learning in Linear Dynamical Systems
James Martens
Master's Thesis, Dept. of Computer Science, University of Toronto
[PDF]
Posters/Presentations
- A neurologically plausible implementation of statistical inference applied to random dot motion
James Martens, Chris Eliasmith
Submitted poster, Computational Neuroscience (CNS), 2007
- Challenges for biological bayes: Solving normalization
Chris Eliasmith, James Martens
Presentation, COSYNE Workshop on Statistical Inference in the Brain, 2007
- A biologically realistic model of statistical inference applied to random dot motion
James Martens, Chris Eliasmith
Submitted poster, COSYNE, 2007
Patents
- Weighted gradient method and system for diagnosing disease.
Milan Graovac, James Martens, Zorin Pavlovic, Joel Ironstone
U.S. patent
News :