HIRL: Hierarchical Inverse Reinforcement Learning for Long-Horizon Tasks with Delayed Rewards

Abstract

Reinforcement Learning (RL) struggles in problems with delayed rewards, and one approach is to segment the task into sub-tasks with incremental rewards. We propose a framework called Hierarchical Inverse Reinforcement Learning (HIRL), which is a model for learning sub-task structure from demonstrations. HIRL decomposes the task into sub-tasks based on transitions that are consistent across demonstrations. These transitions are defined as changes in local linearity w.r.t to a kernel function. Then, HIRL uses the inferred structure to learn reward functions local to the sub-tasks but also handle any global dependencies such as sequentiality. We have evaluated HIRL on several standard RL benchmarks: Parallel Parking with noisy dynamics, Two-Link Pendulum, 2D Noisy Motion Planning, and a Pinball environment. In the parallel parking task, we find that rewards constructed with HIRL converge to a policy with an 80% success rate in 32% fewer time-steps than those constructed with Maximum Entropy Inverse RL (MaxEnt IRL), and with partial state observation, the policies learned with IRL fail to achieve this accuracy while HIRL still converges. We further find that that the rewards learned with HIRL are robust to environment noise where they can tolerate 1 stdev. of random perturbation in the poses in the environment obstacles while maintaining roughly the same convergence rate. We find that HIRL rewards can converge up-to 6x faster than rewards constructed with IRL.

arXiv comment12 pages

References in corpus (3)

• End-to-End Training of Deep Visuomotor Policies
• A Two-round Variant of EM for Gaussian Mixtures
• Learning Contact-Rich Manipulation Skills with Guided Policy Search

Cited by in corpus (9)

• An Algorithmic Perspective on Imitation Learning
• Composing Meta-Policies for Autonomous Driving Using Hierarchical Deep Reinforcement Learning
• Planning on the fast lane: Learning to interact using attention mechanisms in path integral inverse reinforcement learning
• Inverse Reinforcement Learning via Nonparametric Spatio-Temporal Subgoal Modeling
• Online Baum-Welch algorithm for Hierarchical Imitation Learning
• Human-Interactive Subgoal Supervision for Efficient Inverse Reinforcement Learning
• oIRL: Robust Adversarial Inverse Reinforcement Learning with Temporally Extended Actions
• Learning a Skill-sequence-dependent Policy for Long-horizon Manipulation Tasks
• Developing cooperative policies for multi-stage tasks