Sotabase

Career

· PhD in Computer Science, UC Berkeley2020–
· AI Resident, Google Brain, Montréal2019–2020
· Head Undergraduate Student Instructor, UC Berkeley Statistics2017–2019
· Visualization Lead, Berkeley Institute for Data Science2016–2016
· Research Assistant, Lawrence Berkeley National Laboratory2016–2019
· Lead Developer, Preminon Inc2016–2018
· Course Developer, UC Berkeley2016–2016
· Course Development Assistant, UC Berkeley2016–2016
· Undergraduate: Computer Science (honors) and Applied Mathematics (honors), UC Berkeley2015–2019
· Member of the Technical Staff, Pretraining Team, Anthropic

Publications (32)

Octo: An Open-Source Generalist Robot Policy
Robotics: Science and Systems · 2024
928
cited
π0.5: a Vision-Language-Action Model with Open-World Generalization
arXiv.org · 2025
431
cited
Learning to Reach Goals via Iterated Supervised Learning
International Conference on Learning Representations · 2019
208
cited
Why Generalization in RL is Difficult: Epistemic POMDPs and Implicit Partial Observability
Neural Information Processing Systems · 2021
146
cited
Implicit Under-Parameterization Inhibits Data-Efficient Deep Reinforcement Learning
International Conference on Learning Representations · 2020
142
cited
Divide-and-Conquer Reinforcement Learning
International Conference on Learning Representations · 2017
131
cited
Variational Inverse Control with Events: A General Framework for Data-Driven Reward Definition
Neural Information Processing Systems · 2018
130
cited
Learning Actionable Representations with Goal-Conditioned Policies
International Conference on Learning Representations · 2018
113
cited
HIQL: Offline Goal-Conditioned RL with Latent States as Actions
Neural Information Processing Systems · 2023
102
cited
Reinforcement Learning from Passive Data via Latent Intentions
International Conference on Machine Learning · 2023
65
cited
Offline RL Policies Should be Trained to be Adaptive
International Conference on Machine Learning · 2022
56
cited
Representations for Stable Off-Policy Reinforcement Learning
International Conference on Machine Learning · 2020
45
cited
Learning To Reach Goals Without Reinforcement Learning
arXiv.org · 2019
37
cited
Robotic Offline RL from Internet Videos via Value-Function Pre-Training
arXiv.org · 2023
29
cited
An operator view of policy gradient methods
Neural Information Processing Systems · 2020
28
cited
On Catastrophic Interference in Atari 2600 Games
arXiv.org · 2020
27
cited
Accelerating Exploration with Unlabeled Prior Data
Neural Information Processing Systems · 2023
18
cited
What Do Learning Dynamics Reveal About Generalization in LLM Reasoning?
arXiv.org · 2024
18
cited
Distributionally Adaptive Meta Reinforcement Learning
Neural Information Processing Systems · 2022
17
cited
Robotic Offline RL from Internet Videos via Value-Function Learning
IEEE International Conference on Robotics and Automation · 2024
11
cited
Sotabase
Dibya Ghosh | Researcher Profile | Sotabase | Sotabase