Robin Strudel

I am a Research Scientist at DeepMind in Paris, where I work on generative models. I obtained my PhD from INRIA and DI ENS, where I was part of the Willow team, and developed learning methods to perform visually-guided tasks on a robot with my advisors Ivan Laptev and Cordelia Schmid.

Before my PhD, I earned a master's degree in mathematics, machine learning and computer vision (MVA) from ENS Paris-Saclay and in probability from ENS Lyon. I also had the opportunity to work as visiting student researcher in the UC Berkeley Department of Statistics for a year under the supervision of Steven N. Evans and as an intern in the Oxford Department of Statistics with Julien Berestycki.

Email / CV / Google Scholar / Github

Research

	Self-conditioned Embedding Diffusion for Text Generation Robin Strudel, Corentin Tallec, Florent Altché, Yilun Du, Yaroslav Ganin, Arthur Mensch, Will Grathwohl, Nikolay Savinov, Sander Dieleman, Laurent Sifre, Rémi Leblond Preprint, 2022 arXiv / bibtex A general-purpose and flexible language diffusion model.
	Assembly Planning from Observations under Physical Constraints Thomas Chabal, Robin Strudel, Etienne Arlaud, Jean Ponce, Cordelia Schmid IROS, 2022 arXiv / bibtex Assembling structures from a single photograph.
	Weakly-supervised segmentation of referring expressions Robin Strudel, Ivan Laptev, Cordelia Schmid Preprint, 2022 arXiv / bibtex Learning segmentation from referring expressions, without pixel-level supervision.
	Segmenter: Transformer for Semantic Segmentation Robin Strudel, Ricardo Garcia, Ivan Laptev, Cordelia Schmid ICCV, 2021 arXiv / code / bibtex Semantic segmentation as a sequence-to-sequence mapping with Vision Transformers.
	Learning Obstacle Representations for Neural Motion Planning Robin Strudel, Ricardo Garcia, Justin Carpentier, Jean-Paul Laumond, Ivan Laptev, Cordelia Schmid CoRL, 2020 arXiv / project / code / bibtex Visually guided motion planning in unstructured and dynamically changing environments.
	Learning to combine primitive skills: A step towards versatile robotic manipulation Robin Strudel, Alexander Pashevich, Igor Kalevatykh, Ivan Laptev, Josef Sivic, Cordelia Schmid ICRA, 2020 arXiv / project / code / bibtex Learning to perform manipulation tasks with a hierarchical approach. A vocabulary of simple skills is learned from demonstrations then combined with a planning policy to perform more complex tasks.
	Learning to Augment Synthetic Images for Sim2Real Policy Transfer Alexander Pashevich, Robin Strudel*, Igor Kalevatykh, Ivan Laptev, Cordelia Schmid IROS*, 2019 arXiv / project / code / bibtex Learning sim-to-real data augmentation automatically with MCTS and then transferring policies learned in simulation to a real robot.

Design and source code from Jon Barron's website.