Hindsight Policy Gradients

Abstract

A reinforcement learning agent that needs to pursue different goals across episodes requires a goal-conditional policy. In addition to their potential to generalize desirable behavior to unseen goals, such policies may also enable higher-level planning based on subgoals. In sparse-reward environments, the capacity to exploit information about the degree to which an arbitrary goal has been achieved while another goal was intended appears crucial to enable sample efficient learning. However, reinforcement learning agents have only recently been endowed with such capacity for hindsight. In this paper, we demonstrate how hindsight can be introduced to policy gradient methods, generalizing this idea to a broad class of successful algorithms. Our experiments on a diverse selection of sparse-reward environments show that hindsight leads to a remarkable increase in sample efficiency.

Stats

• citations0
• S2 citations—
• github stars0
• HF likes0
• heat score0.00
• arxiv keyrauber2017hindsight

Related papers

• Rewriting History With Inverse RL: Hindsight Inference For Policy Improvement (2020)0.00
• Learning Self-imitating Diverse Policies (2018)0.00
• Hindsight Trust Region Policy Optimization (2019)0.00
• In Hindsight: A Smooth Reward For Steady Exploration (2019)0.00
• Improving Policy Gradient By Exploring Under-appreciated Rewards (2016)0.00
• Bias-reduced Hindsight Experience Replay With Virtual Goal Prioritization (2019)9.41
• Policy Gradient From Demonstration And Curiosity (2020)0.00
• Hindsight Value Function For Variance Reduction In Stochastic Dynamic Environment (2021)2.26