doi.bio/bilal_piot
Bilal Piot
Bilal Piot is a researcher in the field of machine learning and artificial intelligence. He is currently affiliated with Google and has previously been associated with the University of Lille and the Centre national de la recherche scientifique. Piot's research interests include reinforcement learning, Markov decision processes, and self-supervised learning. He has co-authored 60 publications, which have received over 5,000 citations.
Education and Career
Piot's educational background is not readily available, but his career as a researcher has been notable. He has worked with several institutions and collaborated with numerous researchers in the field of machine learning and AI.
Research and Publications
Piot has contributed to a range of research topics, with a focus on reinforcement learning and self-supervised learning. Some of his most notable publications include:
- "Bootstrap Your Own Latent: A New Approach to Self-Supervised Learning" (2020): This paper introduces Bootstrap Your Own Latent (BYOL), a method for self-supervised image representation learning that achieves state-of-the-art performance.
- "Rainbow: Combining Improvements in Deep Reinforcement Learning" (2017): The paper examines extensions to the DQN algorithm and their combination, demonstrating state-of-the-art performance on the Atari 2600 benchmark.
- "Leveraging Demonstrations for Deep Reinforcement Learning on Robotics Problems with Sparse Rewards" (2017): This publication proposes a model-free approach for reinforcement learning on robotics tasks with sparse rewards, building upon the Deep Deterministic Policy Gradient algorithm.
- "Deep Q-learning from Demonstrations" (2017): The paper presents an algorithm, Deep Q-learning from Demonstrations (DQfD), which accelerates the learning process using small sets of demonstration data.
Co-authors and Collaborations
Throughout his career, Bilal Piot has collaborated with many researchers in the field, including Jean-Bastien Grill, Florian Strub, Florent Altché, Corentin Tallec, Pierre H. Richemond, Matteo Hessel, Joseph Modayil, Hado van Hasselt, and Todd Hester, among others.
Bilal Piot
Bilal Piot is a researcher in the field of machine learning and artificial intelligence, currently affiliated with Google and DeepMind. He previously held positions at the University of Lille and the Centre national de la recherche scientifique. Piot's research interests include reinforcement learning, Markov decision processes, and self-supervised learning.
Biography
Piot has an h-index of 30 and has co-authored 60 publications, receiving 5481 citations as of February 2024. His notable works include:
- "Bootstrap Your Own Latent: A New Approach to Self-Supervised Learning" (BYOL), published in 2020, which introduces a new method for self-supervised image representation learning, achieving state-of-the-art performance on transfer and semi-supervised benchmarks.
- "Rainbow: Combining Improvements in Deep Reinforcement Learning" (2017), which examines the combination of six extensions to the DQN algorithm, demonstrating state-of-the-art performance on the Atari 2600 benchmark.
- "Leveraging Demonstrations for Deep Reinforcement Learning on Robotics Problems with Sparse Rewards" (2017), where he proposes a general and model-free approach for reinforcement learning on real robotics tasks with sparse rewards, building upon the Deep Deterministic Policy Gradient algorithm.
- "Deep Q-learning from Demonstrations" (2017), which presents an algorithm that leverages small sets of demonstration data to accelerate the learning process and automatically assesses the necessary ratio of demonstration data.
Publications
- "Bootstrap Your Own Latent: A New Approach to Self-Supervised Learning" (2020) with Jean-Bastien Grill, Florian Strub, Florent Altché, Corentin Tallec, Pierre H. Richemond, and others.
- "Rainbow: Combining Improvements in Deep Reinforcement Learning" (2017) with Matteo Hessel, Joseph Modayil, Hado van Hasselt, Tom Schaul, and others.
- "Leveraging Demonstrations for Deep Reinforcement Learning on Robotics Problems with Sparse Rewards" (2017) with Matej Vecerík, Todd Hester, Jonathan Scholz, Fumin Wang, Olivier Pietquin, and others.
- "Deep Q-learning from Demonstrations" (2017) with Todd Hester, Matej Vecerík, Olivier Pietquin, Marc Lanctot, Tom Schaul, and others.
- "The Edge of Orthogonality: A Simple View of What Makes BYOL Tick" with Pierre H. Richemond, Allison Tam, Yunhao Tang, Florian Strub, Felix Hill.
- "Nash Learning from Human Feedback" with Rémi Munos, Michal Valko, Daniele Calandriello, Mohammad Gheshlaghi Azar, Mark Rowland, and others.
- "Robust Exploration via Clustering-based Online Density Estimation" with Alaa Saade, Steven Kapturowski, Daniele Calandriello, Charles Blundell, Michal Valko, Pablo Sprechmann.
- "Acme: A Research Framework for Distributed Reinforcement Learning" with Matthew W. Hoffman, Bobak Shahriari, John Aslanides, Gabriel Barth-Maron, Nikola Momchev, and others.
- "Mastering the Game of Stratego with Model-Free Multiagent Reinforcement Learning" with Julien Perolat, Bart de Vylder, Daniel Hennes, Eugene Tarassov, Florian Strub, and others.
Youtube Videos
Youtube Title: RLVS 2021 - Day 3 - Deep Q-Networks and its variants (Part 1)
Youtube Link: link
Youtube Channel Name: ANITI Toulouse
Youtube Channel Link: https://www.youtube.com/@anititoulouse8186
RLVS 2021 - Day 3 - Deep Q-Networks and its variants (Part 1)
Youtube Title: DQfD playing Amidar
Youtube Link: link
Youtube Channel Name: Todd Hester
Youtube Channel Link: https://www.youtube.com/@toddhester1
DQfD playing Amidar
Youtube Title: RLVS 2021 - Day 3 - Deep Q-Networks and its variants (Part 3)
Youtube Link: link
Youtube Channel Name: ANITI Toulouse
Youtube Channel Link: https://www.youtube.com/@anititoulouse8186
RLVS 2021 - Day 3 - Deep Q-Networks and its variants (Part 3)
Youtube Title: DQfD playing Private Eye
Youtube Link: link
Youtube Channel Name: Todd Hester
Youtube Channel Link: https://www.youtube.com/@toddhester1
DQfD playing Private Eye
Youtube Title: DQfD playing Pitfall
Youtube Link: link
Youtube Channel Name: Todd Hester
Youtube Channel Link: https://www.youtube.com/@toddhester1
DQfD playing Pitfall
Youtube Title: Leveraging Demonstrations for Deep Reinforcement Learning on Robotics Problems with Sparse Rewards
Youtube Link: link
Youtube Channel Name: Jon Scholz
Youtube Channel Link: https://www.youtube.com/@jonscholz237
Leveraging Demonstrations for Deep Reinforcement Learning on Robotics Problems with Sparse Rewards
Youtube Title: DQfD playing Hero
Youtube Link: link
Youtube Channel Name: Todd Hester
Youtube Channel Link: https://www.youtube.com/@toddhester1
DQfD playing Hero
Youtube Title: Agent57: Outperforming the Atari Human Benchmark
Youtube Link: link
Youtube Channel Name: Yannic Kilcher
Youtube Channel Link: https://www.youtube.com/@YannicKilcher
Agent57: Outperforming the Atari Human Benchmark
Youtube Title: DeepMind's AI Plays 57 Different Games With Super-Human Performance | Game Futurology #11
Youtube Link: link
Youtube Channel Name: DeepGamingAI
Youtube Channel Link: https://www.youtube.com/@DeepGamingAI
DeepMind's AI Plays 57 Different Games With Super-Human Performance | Game Futurology #11
Youtube Title: Deep Reinforcement Learning: Meta-Learning | Analytics Club at ETH | Data Science Meetup
Youtube Link: link
Youtube Channel Name: Analytics Club at ETH
Youtube Channel Link: https://www.youtube.com/@analyticsclubateth4081
Deep Reinforcement Learning: Meta-Learning | Analytics Club at ETH | Data Science Meetup
Youtube Title: DQfD playing Montezuma's Revenge
Youtube Link: link
Youtube Channel Name: Todd Hester
Youtube Channel Link: https://www.youtube.com/@toddhester1
DQfD playing Montezuma's Revenge
Youtube Title: BYOL: Bootstrap Your Own Latent: A New Approach to Self-Supervised Learning (Paper Explained)
Youtube Link: link
Youtube Channel Name: Yannic Kilcher
Youtube Channel Link: https://www.youtube.com/@YannicKilcher
BYOL: Bootstrap Your Own Latent: A New Approach to Self-Supervised Learning (Paper Explained)
Youtube Title: ODI Summit 2021: The Data Game
Youtube Link: link
Youtube Channel Name: Open Data Institute
Youtube Channel Link: https://www.youtube.com/@OpenDataInstituteUK
ODI Summit 2021: The Data Game
Youtube Title: NIPS: Spotlight Session 7: Reinforcement Learning Spotlights
Youtube Link: link
Youtube Channel Name: Microsoft Research
Youtube Channel Link: https://www.youtube.com/@MicrosoftResearch
NIPS: Spotlight Session 7: Reinforcement Learning Spotlights
Youtube Title: ITRE New Members Seminar Series - The Geography of Unemployment
Youtube Link: link
Youtube Channel Name: CEPR & VideoVox Economics
Youtube Channel Link: https://www.youtube.com/@VOXViewsCEPR
ITRE New Members Seminar Series - The Geography of Unemployment
Youtube Title: Charles Blundell - Agent57: Outperforming the Atari Human Benchmark
Youtube Link: link
Youtube Channel Name: London Machine Learning Meetup
Youtube Channel Link: https://www.youtube.com/@LondonMachineLearningMeetup
Charles Blundell - Agent57: Outperforming the Atari Human Benchmark
Youtube Title: Learning and Transferring Visual Representations with Few Labels - Carl Doersch
Youtube Link: link
Youtube Channel Name: Oxford VGG
Youtube Channel Link: https://www.youtube.com/@oxfordvgg5972
Learning and Transferring Visual Representations with Few Labels - Carl Doersch
Youtube Title: Harvard Doctor: As States Rush to Reopen, Lack of COVID-19 Testing Is “Achilles Heel” for U.S.
Youtube Link: link
Youtube Channel Name: Democracy Now!
Youtube Channel Link: https://www.youtube.com/@DemocracyNow
Harvard Doctor: As States Rush to Reopen, Lack of COVID-19 Testing Is “Achilles Heel” for U.S.
Youtube Title: Lbenj - Galaxy
Youtube Link: link
Youtube Channel Name: Lbenj
Youtube Channel Link: https://www.youtube.com/channel/UCSj3fnfKHCI5G2VYUcx3DHg
Lbenj - Galaxy