# Curious Meta-Controller: Adaptive Alternation between Model-Based and Model-Free Control in Deep Reinforcement Learning

## TL;DR

The paper introduces Curious Meta-Controller, an adaptive system that switches between model-based and model-free reinforcement learning using curiosity feedback, significantly improving sample efficiency in robotic tasks from raw pixels.

## Contribution

It presents a novel adaptive control method that combines model-based and model-free RL using curiosity-driven feedback based on neural dynamics models.

## Key findings

- Improves sample efficiency in robotic control tasks.
- Achieves near-optimal performance on pixel-based learning tasks.
- Effective in both dense and sparse reward scenarios.

## Abstract

Recent success in deep reinforcement learning for continuous control has been dominated by model-free approaches which, unlike model-based approaches, do not suffer from representational limitations in making assumptions about the world dynamics and model errors inevitable in complex domains. However, they require a lot of experiences compared to model-based approaches that are typically more sample-efficient. We propose to combine the benefits of the two approaches by presenting an integrated approach called Curious Meta-Controller. Our approach alternates adaptively between model-based and model-free control using a curiosity feedback based on the learning progress of a neural model of the dynamics in a learned latent space. We demonstrate that our approach can significantly improve the sample efficiency and achieve near-optimal performance on learning robotic reaching and grasping tasks from raw-pixel input in both dense and sparse reward settings.