I am a senior research scientist at DeepMind.
Previously I was a research scientist at Lyric Labs (Analog Devices), and before that, a PhD student at the Operations Research Center at MIT.

Research Interests

I am interested in machine learning and artificial intelligence, deep learning, reinforcement learning and model-based RL, probabilistic modeling (and probabilistic programming), and variational methods for inference, among other topics.

Publications and Working Papers

Imagination-Augmented Agents for Deep Reinforcement Learning, with Sebastien Racaniere, David P. Reichert, Lars Buesing, Arthur Guez, Adria Puigdomenech Badia, Oriol Vinyals, Nicolas Heess, Yujia Li, Razvan Pascanu, Peter Battaglia, David Silver, and Daan Wierstra [abstract] [paper].

Learning model-based planning from scratch, with Razvan Pascanu, Yujia Li, Oriol Vinyals, Nicolas Heess, Lars Buesing, Sebastien Racaniere, David Reichert, Daan Wierstra, and Peter Battaglia [abstract] [paper].

Visual Interaction Networks, with Nicholas Watters, Andrea Tacchetti, Razvan Pascanu, Peter Battaglia and Daniel Zoran, [abstract] [paper].

Stochastic Gradient Estimation with Finite Differences, with Lars Buesing and Shakir Mohamed, NIPS 2016 workshop on Advances in Approximate Bayesian Inference [abstract] [paper].

Attend, Infer, Repeat: Fast Scene Understanding with Generative Models, with Ali Eslami, Nicolas Heess, Yuval Tassa, David Szepesvari, Koray Kavukcuoglu, and Geoffrey Hinton, NIPS 2016 [abstract] [paper].

Gradient Estimation Using Stochastic Computation Grapsh, with John Schulman, Nicolas Heess, and Pieter Abbeel, NIPS 2015 [abstract] [paper].

Deep Reinforcement Learning in Large Discrete Action Spaces, with Gabriel Dulac-Arnold, Richard Evans, Hado Van Hasselt, Peter Sunehag, Timothy Lillicrap, Jonathan Hunt, Timothy Mann, Thomas Degris, and Ben Coppin, [abstract] [paper].

Reinforced Variational Inference, with Nicolas Heess, Ali Eslami, John Schulman, David Wingate, and David Silver, NIPS 2015 workshop on Advances in Approximate Bayesian Inference [paper].

Automated Variational Inference in Probabilistic Programming, with D. Wingate, NIPS 2012 workshop on probabilistic programming [abstract] [paper].

Accelerating Inference: towards a full Language, Compiler and Hardware stack, with S. Hershey, J. Bernstein, B. Bradley, A. Schweitzer, N. Stein and Ben Vigoda, presented at the NIPS 2012 workshop on probabilistic programming [abstract] [paper].

Random Decision Networks: Correlation Decay and Decentralized Optimization, with D. Gamarnik, accepted to Mathematics of Operations Research (preliminary version best submission to the 2008 Conference on Stochastic Networks). [abstract] [paper] [online complement]

Quantifying Statistical Interdependence by Message Passing on Graphs: Algorithms and Application to Neural Signals: Part I, with J.Dauwels, F.Vialatte, T. Musha and A.Cichocki, published in Neural Computation. [abstract] [paper]

Quantifying Statistical Interdependence by Message Passing on Graphs: Algorithms and Application to Neural Signals, Part II: multidimensional point processes, with J.Dauwels, F.Vialatte, T. Musha and A.Cichocki, published in Neural Computation. [abstract] [paper]

Quantifying Statistical Interdependence by Message Passing on Graphs: Algorithms and Application to Neural Signals, Part III: N>2 point processes, with J.Dauwels, F.Vialatte, T. Musha and A.Cichocki, published in Neural Computation. [abstract] [paper]

State-relevance Weights for the Linear Programming Approach to Dynamic Programming, with D. Pucci de Farias, in preparation (preliminary version at CDC 2008). [abstract] [paper]

To Wave or Not to Wave? Order Release Policies for Warehouses with an Automated Sorter, with J. Gallien, Mathematics of Operations Research [abstract] [paper].

Theophane Weber

Contact Information

Théophane Weber-de Tonquedec

e-mail: theophaneatgmail.com

twitter: attheophaneweber