Parameter-Based Value Functions
arXiv:2006.09226
Abstract
Traditional off-policy actor-critic Reinforcement Learning (RL) algorithms learn value functions of a single target policy. However, when value functions are updated to track the learned policy, they forget potentially useful information about old policies. We introduce a class of value functions called Parameter-Based Value Functions (PBVFs) whose inputs include the policy parameters. They can generalize across different policies. PBVFs can evaluate the performance of any policy given a state, a state-action pair, or a distribution over the RL agent's initial states. First we show how PBVFs yield novel off-policy policy gradient theorems. Then we derive off-policy actor-critic algorithms based on PBVFs trained by Monte Carlo or Temporal Difference methods. We show how learned PBVFs can zero-shot learn new policies that outperform any policy seen during training. Finally our algorithms are evaluated on a selection of discrete and continuous control tasks using shallow policies and deep neural networks. Their performance is comparable to state-of-the-art methods.
Published as a conference paper at ICLR 2021
References in corpus (10)
- Deep Learning in Neural Networks: An Overview
- Practical Bayesian Optimization of Machine Learning Algorithms
- Soft Actor-Critic: Off-Policy Maximum Entropy Deep Reinforcement Learning with a Stochastic Actor
- Evolution Strategies as a Scalable Alternative to Reinforcement Learning
- Scalable Bayesian Optimization Using Deep Neural Networks
- AlgaeDICE: Policy Gradient from Arbitrary Experience
- Off-Policy Policy Gradient with State Distribution Correction
- Predicting Neural Network Accuracy from Weights
- Policy Optimization via Importance Sampling
- Policy Evaluation Networks