Matt J. Kusner

I am an Associate Professor in Machine Learning at the University of Oxford.
My work is in fair algorithms, discrete generative models, document distances, privacy, dataset compression, budgeted learning, and Bayesian optimization.

Previously I was a member of Kilian Weinberger's machine learning research group. CV, Contact

I was recently awarded the Turner Dissertation Award for best Computer Science & Engineering doctoral dissertation.

Preprints

PDF Matt J. Kusner, Chris Russell, Joshua R. Loftus, Ricardo Silva
Causal Interventions for Fairness
arXiv:1806.02380, 2018

PDF Joshua R. Loftus, Chris Russell, Matt J. Kusner, Ricardo Silva
Causal Reasoning for Algorithmic Fairness
arXiv:1805.05859, 2018

PDF John Bradshaw, Matt J. Kusner, Brooks Paige, Marwin H. S. Segler, José Miguel Hernández-Lobato
Predicting Electron Paths
arXiv:1805.10970, 2018

Publications

PDF CODE Amartya Sanyal, Matt J. Kusner, Adrià Gascón, Varun Kanade
TAPAS: Tricks to Accelerate (encrypted) Prediction As a Service
The International Conference on Machine Learning (ICML), 2018

PDF Niki Kilbertus, Adrià Gascón, Matt J. Kusner, Michael Veale, Krishna Gummadi, Adrian Weller
Blind Justice: Fairness with Encrypted Sensitive Attributes
The International Conference on Machine Learning (ICML), 2018

PDF David Janz, Jos van der Westhuizen, Brooks Paige, Matt J. Kusner, José Miguel Hernández-Lobato
Learning a Generative Model for Validity in Complex Discrete Structures
International Conference on Learning Representations (ICLR), 2018

PDF Chris Russell*, Matt J. Kusner*, Joshua R. Loftus, Ricardo Silva
When Worlds Collide: Integrating Different Counterfactual Assumptions in Fairness
Neural Information Processing Systems (NIPS), 2017
*=authors contributing equally

PDF POSTER TALK CODE Matt J. Kusner*, Joshua R. Loftus*, Chris Russell*, Ricardo Silva
Counterfactual Fairness [Oral Presentation]
Neural Information Processing Systems (NIPS), 2017
*=authors contributing equally

PDF TALK CODE Matt J. Kusner*, Brooks Paige*, José Miguel Hernández-Lobato
Grammar Variational Autoencoder
The International Conference on Machine Learning (ICML), 2017
*=authors contributing equally

PDF Matt J. Kusner
Learning in the Real World: Constraints on Cost, Space, and Privacy
Ph.D. Thesis, 2016

PDF TALK CODE Gao Huang*, Chuan Guo*, Matt J. Kusner, Yu Sun, Kilian Q. Weinberger, Fei Sha
Supervised Word Mover's Distance [Oral Presentation]
Neural Information Processing Systems (NIPS), 2016
*=authors contributing equally

PDF Matt J. Kusner, Yu Sun, Karthik Sridharan, Kilian Q. Weinberger
Private Causal Inference [Oral Presentation]
Artificial Intelligence and Statistics (AISTATS), 2016

PDF Gustavo Malkomes, Matt J. Kusner, Wenlin Chen, Kilian Q. Weinberger, Benjamin Moseley
Fast Distributed k-Center Clustering with Outliers on Massive Data
Neural Information Processing Systems (NIPS), 2015

PDF SLIDES POSTER TALK CODE Matt J. Kusner, Yu Sun, Nicholas I. Kolkin, Kilian Q. Weinberger
From Word Embeddings To Document Distances
The International Conference on Machine Learning (ICML), 2015

PDF SLIDES POSTER TALK Matt J. Kusner, Jacob R. Gardner, Roman Garnett, Kilian Q. Weinberger
Differentially Private Bayesian Optimization
International Conference on Machine Learning (ICML), 2015

PDF Zhixiang (Eddie) Xu, Matt J. Kusner, Kilian Q. Weinberger, Minmin Chen, Olivier Chapelle
Classifier Cascades and Trees for Minimizing Feature Evaluation Cost
Journal of Machine Learning Research (JMLR), 2014

PDF POSTER Matt J. Kusner, Wenlin Chen, Quan Zhou, Zhixiang (Eddie) Xu, Kilian Q. Weinberger, Yixin Chen
Feature-Cost Sensitive Learning with Submodular Trees of Classifiers
AAAI Conference on Artificial Intelligence (AAAI), 2014

PDF SLIDES POSTER CODE Matt J. Kusner, Stephen Tyree, Kilian Q. Weinberger, Kunal Agrawal
Stochastic Neighbor Compression
International Conference on Machine Learning (ICML), 2014

PDF Jacob R. Gardner, Matt J. Kusner, Zhixiang (Eddie) Xu, Kilian Q. Weinberger, John P. Cunningham
Bayesian Optimization with Inequality Constraints
International Conference on Machine Learning (ICML), 2014

PDF Zhixiang (Eddie) Xu, Matt J. Kusner, Gao Huang, Kilian Q. Weinberger
Anytime Feature Learning
International Conference on Machine Learning (ICML), 2013

PDF Zhixiang (Eddie) Xu, Matt J. Kusner, Kilian Q. Weinberger, Minmin Chen
Cost-Sensitive Tree of Classifiers
International Conference on Machine Learning (ICML), 2013


Technical Reports

PDF Jacob R. Gardner*, Paul Upchurch*, Matt J. Kusner, Yixuan Li, Kilian Q. Weinberger, Kavita Bala, John E. Hopcroft
Deep Manifold Traversal: Changing Labels with Convolutional Features
*=authors contributing equally

PDF Matt J. Kusner, Nicholas I. Kolkin, Stephen Tyree, Kilian Q. Weinberger
Image Data Compression for Covariance and Histogram Descriptors

Code

Counterfactual Fairness

R/Stan code can be found here: CODE

Grammar Variational Autoencoder

Keras/Tensorflow code for the GVAE can be found here: CODE

Supervised Word Mover's Distance

Matlab code for the SWMD written by the brilliant Gao Huang: CODE

Word Mover's Distance

Code for scikit-learn compatible WMD written by the prolific Vlad Niculae: CODE
Renaud Richardet has generously created a very nice GitHub repository from the code I released: CODE
WMD has also recently been added to the GENSIM Python library: CODE

Stochastic Neighbor Compression

Here is the first release of the code: CODE See the README within for details.