Model-Based Planning with Discrete and Continuous Actions

Abstract

Action planning using learned and differentiable forward models of the world is a general approach which has a number of desirable properties, including improved sample complexity over model-free RL methods, reuse of learned models across different tasks, and the ability to perform efficient gradient-based optimization in continuous action spaces. However, this approach does not apply straightforwardly when the action space is discrete. In this work, we show that it is in fact possible to effectively perform planning via backprop in discrete action spaces, using a simple paramaterization of the actions vectors on the simplex combined with input noise when training the forward model. Our experiments show that this approach can match or outperform model-free RL and discrete planning methods on gridworld navigation tasks in terms of performance and/or planning time while using limited environment interactions, and can additionally be used to perform model-based control in a challenging new task where the action space combines discrete and continuous actions. We furthermore propose a policy distillation approach which yields a fast policy network which can be used at inference time, removing the need for an iterative planning procedure.

References in corpus (4)

• Distilling the Knowledge in a Neural Network
• Categorical Reparameterization with Gumbel-Softmax
• The Concrete Distribution: A Continuous Relaxation of Discrete Random Variables
• Learning model-based planning from scratch

Cited by in corpus (12)

• Dream to Control: Learning Behaviors by Latent Imagination
• Mapping State Space using Landmarks for Universal Goal Reaching
• Mastering Atari with Discrete World Models
• Model-Predictive Control via Cross-Entropy and Gradient-Based Optimization
• Bridging Imagination and Reality for Model-Based Deep Reinforcement Learning
• Continual Learning Using World Models for Pseudo-Rehearsal
• Iterative Amortized Policy Optimization
• Swoosh! Rattle! Thump! -- Actions that Sound
• Data-Driven Reinforcement Learning for Virtual Character Animation Control
• High Performance Across Two Atari Paddle Games Using the Same Perceptual Control Architecture Without Training
• Improving Hyperparameter Optimization by Planning Ahead
• Towards an Interpretable Latent Space in Structured Models for Video Prediction