COBRA: Data-Efficient Model-Based RL through Unsupervised Object Discovery and Curiosity-Driven Exploration

Agent solving a clustering task after 100 steps


Data efficiency and robustness to task-irrelevant perturbations are long-standing challenges for deep reinforcement learning algorithms. Here we introduce a modular approach to addressing these challenges in a continuous control environment, without using hand-crafted or supervised information. Our Curious Object-Based seaRch Agent (COBRA) uses task-free intrinsically motivated exploration and unsupervised learning to build object-based models of its environment and action space. Subsequently, it can learn a variety of tasks through model-based search in very few steps and excel on structured hold-out tests of policy robustness.


This project also led to the open-sourcing of the Spriteworld environment/rendering framework, a flexible and configurable Python-based learning environment.

Loic Matthey
Loic Matthey
Staff Research Scientist in Machine Learning

ex-Neuroscientist working on Artificial General Intelligence at Google DeepMind. Unsupervised learning, structured generative models, concepts and how to make AI actually generalize is what I do.