Continuous Adaptation via Meta-Learning in Nonstationary and Competitive Environments

Introduction

In this paper, a probabilistic framework for meta learning is derived then applied to tasks involving simulated robotic spiders.