Hierarchical Skills for Efficient Exploration

Abstract

In reinforcement learning, pre-trained low-level skills have the potential to greatly facilitate exploration. However, prior knowledge of the downstream task is required to strike the right balance between generality (fine-grained control) and specificity (faster learning) in skill design. In previous work on continuous control, the sensitivity of methods to this trade-off has not been addressed explicitly, as locomotion provides a suitable prior for navigation tasks, which have been of foremost interest. In this work, we analyze this trade-off for low-level policy pre-training with a new benchmark suite of diverse, sparse-reward tasks for bipedal robots. We alleviate the need for prior knowledge by proposing a hierarchical skill learning framework that acquires skills of varying complexity in an unsupervised manner. For utilization on downstream tasks, we present a three-layered hierarchical learning algorithm to automatically trade off between general and specific skills as required by the respective task. In our experiments, we show that our approach performs this trade-off effectively and achieves better results than current state-of-the-art methods for end- to-end hierarchical reinforcement learning and unsupervised skill discovery. Code and videos are available at https://facebookresearch.github.io/hsd3 .

To appear in 35th Conference on Neural Information Processing Systems (NeurIPS 2021)

References in corpus (19)

• Soft Actor-Critic Algorithms and Applications
• Emergence of Locomotion Behaviours in Rich Environments
• #Exploration: A Study of Count-Based Exploration for Deep Reinforcement Learning
• FeUdal Networks for Hierarchical Reinforcement Learning
• Data-Efficient Hierarchical Reinforcement Learning
• dm_control: Software and Tasks for Continuous Control
• Learning human behaviors from motion capture by adversarial imitation
• Neural probabilistic motor primitives for humanoid control
• Variational Option Discovery Algorithms
• A Laplacian Framework for Option Discovery in Reinforcement Learning
• Reinforcement Learning with Parameterized Actions
• Why Does Hierarchy (Sometimes) Work So Well in Reinforcement Learning?
• Discrete and Continuous Action Representation for Practical RL in Video Games
• Hierarchical Reinforcement Learning By Discovering Intrinsic Options
• Swing-twist decomposition in Clifford algebra
• Learning Robot Skills with Temporal Variational Inference
• D2RL: Deep Dense Architectures in Reinforcement Learning
• Composing Task-Agnostic Policies with Deep Reinforcement Learning
• Planning in Learned Latent Action Spaces for Generalizable Legged Locomotion