I am a Ph.D. candidate with the Computational and Biological Learning group at the University of Cambridge, supervised by Dr José Miguel Hernández-Lobato and advised by Dr Richard Turner. My research is focused on developing probabilistic models (typically parameterized by deep neural networks) and associated scalable approximate inference procedures.
Recent advances in stochastic backpropagation enable us to efficiently train increasingly complex models. Some notable examples are the variational autoencoder and Bayesian neural networks. In my research I aim to examine how we can design these (and similar) models to enable more efficient learning (statistical and computational efficiency) as well as learning in more complex settings. I am especially interested in leveraging generative abilities of models to these ends.
In my previous studies I examined applications of machine learning to specific domains such as healthcare and manufacturing. Specifically, I examined how machine learning might be useful in prognosis and prediction of ALS progression based on clinical trial data. I also worked on a project to develop optimal sampling strategies for quality assurance processes in semi-conductor manufacturing environments.
In my spare time I like to cook, run, listen and play music, watch movies etc. I also love spending time with the best part of me: my beautiful and perfect wife R.C.
- Present - PhD in Machine learning – University of Cambridge
- 2017 - MPhil in Machine Learning, Speech and Language Technology – University of Cambridge (distinction).
- 2016 - MSc. in Applied Statistics and Engineering – Ben-Gurion University (magna cum laude).
- 2014 - BSc. in Engineering – Ben-Gurion University (magna cum laude).
Bayesian Batch Active Learning as Sparse Subset Approximation
R. Pinsler, J. Gordon, E. Nalisnick, J. M. H. Lobato, NeurIPS 2019, arXiv:1908.02144
Fast and Flexible Multi-Task Classification Using Conditional Neural Adaptive Processes
J. R. Requima, J. Gordon, J. F. Bronskill, S. Nowozin, and R. E. Turner, NeurIPS 2019 (spotlight), arXiv:1906.07697
Probabilistic Neural Architecture Search
F. P. Casale, J. Gordon, N. Fusi, arXiv:1902.05116
Meta-Learning Probabilistic Inference for Prediction
J. Gordon, J. F. Bronskill, M. Bauer, S. Nowozin, and R. E. Turner, ICLR 2019 arXiv:1805.09921
Consolidating the Meta-Learning Zoo: a Unifying Perspective as Posterior Predictive Inference
J. Gordon, J. F. Bronskill, M. Bauer, S. Nowozin, and R. E. Turner, Meta-Learning Workshop, NeurIPS 2018
Versa: Versatile and Efficient Few-Shot Learning
J. Gordon, J. F. Bronskill, M. Bauer, S. Nowozin, and R. E. Turner, Bayesian Deep Learning Workshop, NeurIPS 2018
Cambridge MPhil thesis
MPhil thesis on Bayesian adaptations of VAEs and VAE-BNN hybrids (Download)
Bayesian Semi-Supervised and Active Learning with Deep Generative Models
J. Gordon and J. M. H. Lobato, ICML Workshop on Principled Approaches to Deep Learning, arXiv:1706.09751
Exposing and Modeling Underlying Mechanisms in ALS with Machine Learning
J. Gordon and B. Lerner, International Conference on Pattern Recognition, 2016
Disease State Prediction, Knowledge Representation, and Heterogeneity Decomposition for ALS
J. Gordon and B. Lerner, UAI Workshop on Machine Learning in Healthcare, 2016
Awards and Scholarships
Funding for my Ph.D. research at the University of Cambridge, generously provided by Samsung ltd., the University of Cambridge, and the CBL.
Yearly scholarship awarded for outstanding research, generously provided by the AJA foundation for two years.
Yearly scholarship awarded for outstanding research, generously provided by the Kenneth Lindsay foundation for one year.
Funding and tuition provided for two years of MSc. studies for outstanding students, generously provided by Ben-Gurion University and the department of Engineering.