Kelvin Xu

I am a second year master student in the MILA lab at the Université de Montréal studying machine learning (deep learning in particular) under the supervision of Prof. Yoshua Bengio/ Aaron Courville. I did my undergraduate in the Engineering Science Program at the University of Toronto. I did my senior year thesis in the machine learning group where I was very fortunate to work with Prof. Richard Zemel.

My current research focuses on using using neural networks for sequence-to-sequence/image-to-sequence tasks. But there are a lot of interesting problems in machine learning and I aspire to work on at least a few of them at some point.

From Oct 2015-Mar 2016, I visited the machine learning group at the Courant Institute at NYU. Starting in June 2016, I will be a resident at Google Research as part of their Brain Residency Program.

email / google scholar

Kelvin Xu

An Actor-Critic Algorithm for Structured Prediction
D. Bahdanau, P. Brakel, K. Xu, A. Goyal, R. Lowe,
J. Pineau, A. Courville, Y. Bengio
International Conference for Learning Representations (ICLR), 2017
[open review] [arxiv]

Unsupervised Perceptual Rewards for Imitation Learning
P. Sermanet, K. Xu , S. Levine
NIPS Workshop for Action and Interaction, 2016
[open review]

A Controller-Recognizer Framework: How necessary is recognition for control?
M. Moczulski, K. Xu, A. Courville, K. Cho
International Conference for Learning Representations (ICLR) Workshop, 2016
[arxiv preprint]

On Using Monolingual Corpora in Neural Machine Translation
C. Gulchere, O. Firat, K. Xu, K. Cho, L. Barrault,
H. Lin, F. Bougares, H. Schwenk, Y. Bengio
Computer Speech and Language, 2016
[arxiv preprint] [project page]

Show, Attend and Tell: Neural Image Caption Generation with Visual Attention
K. Xu, J. Ba, R. Kiros, K. Cho, A. Courville,
R. Salakhutdinov, R. Zemel, Y. Bengio (2015)
International Conference for Machine Learning (ICML), 2015
[paper] [project page] [inceptionism-captions] [abstract-captions] [code]

(This website sure is swell)