Timezone: »

Unsupervised Curricula for Visual Meta-Reinforcement Learning
Allan Jabri · Kyle Hsu · Abhishek Gupta · Benjamin Eysenbach · Sergey Levine · Chelsea Finn

Wed Dec 11 10:30 AM -- 10:35 AM (PST) @ West Exhibition Hall A

Meta-reinforcement learning algorithms leverage experience across many tasks to learn fast and effective reinforcement learning (RL) algorithms. However, current meta-RL methods depend critically on a manually-defined distribution of meta-training tasks, and hand-crafting these task distributions is challenging and time-consuming. We develop an unsupervised algorithm for inducing an adaptive meta-training task distribution, i.e. an automatic curriculum, by modeling unsupervised interaction in a visual environment. Crucially, the task distribution is scaffolded by the meta-learner's behavior, with density-based exploration driving the evolution of the task distribution. We formulate unsupervised meta-RL with an information-theoretic objective optimized via expectation-maximization over trajectory-level latent variables. Repeating this procedure leads to iterative reorganization of behavior, allowing the task distribution to adapt as the meta-learner becomes more competent. In our experiments on vision-based navigation and manipulation domains, we show that our algorithm allows for unsupervised meta-learning of skills that transfer to downstream tasks specified by human-provided reward functions, as well as pre-training for more efficient meta-learning on user-defined task distributions. To understand the nature of the curricula, we provide visualizations and analysis of the task distributions discovered throughout the learning process, finding that the emergent tasks span a range of environment-specific exploratory and exploitative behavior.

Author Information

Allan Jabri (UC Berkeley)
Kyle Hsu (University of Toronto)
Abhishek Gupta (University of California, Berkeley)
Benjamin Eysenbach (Carnegie Mellon University)
Benjamin Eysenbach

Assistant professor at Princeton working on self-supervised reinforcement learning (scaling, algorithms, theory, and applications).

Sergey Levine (UC Berkeley)
Sergey Levine

Sergey Levine received a BS and MS in Computer Science from Stanford University in 2009, and a Ph.D. in Computer Science from Stanford University in 2014. He joined the faculty of the Department of Electrical Engineering and Computer Sciences at UC Berkeley in fall 2016. His work focuses on machine learning for decision making and control, with an emphasis on deep learning and reinforcement learning algorithms. Applications of his work include autonomous robots and vehicles, as well as applications in other decision-making domains. His research includes developing algorithms for end-to-end training of deep neural network policies that combine perception and control, scalable algorithms for inverse reinforcement learning, deep reinforcement learning algorithms, and more

Chelsea Finn (Stanford University)

Related Events (a corresponding poster, oral, or spotlight)

More from the Same Authors