About Me

Houssam Zenati

Hi! I currently work on causal inference at Criteo AI Lab and Inria Thoth under Julien Mairal's supervision in Grenoble, France. I am a Master's student at CentraleSupélec and Ecole Normale Supérieure Paris-Saclay. Previously, I have been working on Deep Reinforcement Learning in Tokyo, Japan and Computer Vision and Deep Learning in Singapore.

Research

Criteo AI Research/Inria Thoth

I currently work on causal inference, more specifically on counterfactual risk minization. I work both at Criteo AI Lab and Inria Thoth in Grenoble.

Apr. 2019 - Present
Research Intern

IBM Research

I worked at IBM Research in the AI laboratory in Tokyo. My research was on deep reinforcement learning, especially on developping algorithms to learn control policies for constrained robotics problems.

Jun. 2018 - Aug. 2018
Research Intern

Institute for Infocomm Research, A*STAR

I worked at the Computer Vision Lab of I2R in Singapore. I worked mainly on anomaly detection using deep learning models. I also actively participated to semi-supervised learning research projects as well as algorithms on convergence of multi-agents games and saddle points.

Aug. 2017 - Jun. 2018
Research Intern - Research Engineer

Publications

Optimistic mirror descent in saddle-point problems: Going the extra (gradient) mile [arXiv]
P. Mertikopoulos, B. Lecouat, H. Zenati, C.S Foo, V. Chandrasekhar and G. Piliouras. International Conference on Learning Representations, (ICLR) (2019), New Orleans, USA

Semi-Supervised Deep Learning for Abnormality Classification in Retinal Images [arXiv]
B. Lecouat, K. Chang, C.S Foo, B. Unnikrishnan, J. Brown, H. Zenati, A. Beers, V. Chandrasekhar, J. Kalpathy-Cramer and P. Krishnaswamy. Neural Information Processing Systems, (NeurIPS) ML4H Workshop (2018), Montréal, Canada.

Adversarially Learned Anomaly Detection [arXiv] [CODE]
H. Zenati, M. Romain, C.S Foo, B. Lecouat and V. Chandrasekhar. IEEE International Conference on Data Mining (ICDM) (2018), Singapore.

Semi-Supervised Learning With GANs: Revisiting Manifold Regularization [arXiv] [CODE]
B. Lecouat, C.S Foo, H. Zenati and V. Chandrasekhar. International Conference on Learning Representations, (ICLR) Workshop Track (2018), Vancouver, Canada

Efficient GAN-Based Anomaly Detection [arXiv] [CODE]
H. Zenati, C.S Foo, B. Lecouat, G. Manek and V. Chandrasekhar.