Biography
I am a second-year Ph.D. Student at Criteo AI Lab and ENSAE-CREST, under the supervision of Anna Korba, Victor-Emmanuel Brunel and David Rohde. Previously, I was an Applied Scientist at Amazon and a Scholar at Google DeepMind. I am mainly focusing on designing and analyzing practical reward-optimizing algorithms for interactive systems. Precisely, I am interested in Off-Policy Evaluation and Learning, Practical Bandit Algorithms, Bayesian Analysis and Learning, Recommender Systems, and Generative Models.
Publications
2024
Logarithmic Smoothing for Pessimistic Off-Policy Evaluation, Selection and Learning. Neurips 2024 (Spotlight) and RecSys 2024 CONSEQUENCES Workshop (Oral)
Otmane Sakhi, Imad Aouali, Pierre Alquier, Nicolas ChopinUnified PAC-Bayesian Study of Pessimism for Offline Policy Learning with Regularized Importance Sampling. UAI 2024 (Poster)
Imad Aouali, Victor-Emmanuel Brunel, David Rohde, Anna KorbaBayesian Off-Policy Evaluation and Learning for Large Action Spaces. Preprint
Imad Aouali, Victor-Emmanuel Brunel, David Rohde, Anna KorbaDiffusion Models Meet Contextual Bandits with Large Action Spaces. Preprint
Imad AoualiPrior-Dependent Allocations for Bayesian Fixed-Budget Best-Arm Identification in Structured Bandits. Preprint
Nicolas Nguyen*, Imad Aouali*, András György, Claire Vernade
(*Equal contribution)
2023
Linear Diffusion Models Meet Contextual Bandits With Large Action Spaces. Foundation Models for Decision Making Workshop at NeurIPS 2023
Imad AoualiExponential Smoothing for Off-Policy Learning. ICML 2023 (Oral & Poster)
Imad Aouali, Victor-Emmanuel Brunel, David Rohde, Anna KorbaMixed-Effect Thompson Sampling. AISTATS 2023 (Poster)
Imad Aouali, Branislav Kveton, Sumeet KatariyaProbabilistic Rank and Reward: A Scalable Model for Slate Recommendation. Preprint
Imad Aouali*, Achraf Ait Sidi Hammou*, Sergey Ivanov, Otmane Sakhi, David Rohde, Flavian Vasile (*Equal contribution)
2022
Reward Optimizing Recommendation using Deep Learning and Fast Maximum Inner Product Search. Tutorials at KDD 2022 and ECML 2022
Imad Aouali, Amine Benhalloum, Martin Bompaire, Achraf Ait Sidi Hammou, Sergey Ivanov, Benjamin Heymann, David Rohde, Otmane Sakhi, Flavian Vasile, Maxime Vono (All authors contributed equally)Combining Reward and Rank Signals for Slate Recommendation. BCIRWIS Workshop at KDD 2022
Imad Aouali, Sergey Ivanov, Mike Gartrell, David Rohde, Flavian Vasile, Victor Zaytsev, Diego LegrandOffline Evaluation of Reward-Optimizing Recommender Systems: The Case of Simulation. SimuRec Workshop at RecSys 2022 (Position Paper)
Imad Aouali, Amine Benhalloum, Martin Bompaire, Benjamin Heymann, Olivier Jeunen, David Rohde, Otmane Sakhi, Flavian Vasile (All authors contributed equally)