Sotabase

Career

· Master of Engineering, Massachusetts Institute of Technology2023–
· PhD Student, University of California, Berkeley2023–
· Bachelor of Science, Massachusetts Institute of Technology2022–
· AI Research Scientist, Cross Labs2020–2021
· Research Assistant, MIT CSAIL Embodied Intelligence Lab2018–
· AI Research Intern, Office of the CTO, Autodesk Research
· AI Research Intern, Reinforcement Learning Group, OpenAI
· AI Research and Engineering Intern, Sizigi Studios

Publications (17)

Meta Learning Shared Hierarchies
International Conference on Learning Representations · 2017
372
cited
CLIPDraw: Exploring Text-to-Drawing Synthesis through Language-Image Encoders
Neural Information Processing Systems · 2021
267
cited
OGBench: Benchmarking Offline Goal-Conditioned RL
International Conference on Learning Representations · 2024
86
cited
Outline Colorization through Tandem Adversarial Networks
arXiv.org · 2017
63
cited
Is Value Learning Really the Main Bottleneck in Offline RL?
Neural Information Processing Systems · 2024
50
cited
Horizon Reduction Makes RL Scalable
arXiv.org · 2025
18
cited
Unsupervised Zero-Shot Reinforcement Learning via Functional Reward Encodings
International Conference on Machine Learning · 2024
18
cited
Unsupervised Image to Sequence Translation with Canvas-Drawer Networks
arXiv.org · 2018
17
cited
Diffusion Guidance Is a Controllable Policy Improvement Operator
arXiv.org · 2025
13
cited
AI Charades: Language Models as Interactive Game Environments
2021 IEEE Conference on Games (CoG) · 2021
10
cited
Powderworld: A Platform for Understanding Generalization via Rich Task Distributions
International Conference on Learning Representations · 2022
10
cited
Leveraging Skills from Unlabeled Prior Data for Efficient Online Exploration
International Conference on Machine Learning · 2024
7
cited
Population-Based Evolution Optimizes a Meta-Learning Objective
arXiv.org · 2021
6
cited
Parallel Trust Region Policy Optimization with Multiple Actors
2016
2
cited
Questions for the Open-Ended Evolution Community: Reflections from the 2021 Cross Labs Innovation Science Workshop
2021
2
cited
Selecting for Selection: Learning To Balance Adaptive and Diversifying Pressures in Evolutionary Search
arXiv.org · 2021
Sotabase
Kevin Frans | Researcher Profile | Sotabase | Sotabase