Active Reinforcement Learning -- A Roadmap Towards Curious Classifier Systems for Self-Adaptation

TL;DR

This paper proposes a research agenda for active reinforcement learning, aiming to develop self-adaptive intelligent systems that overcome traditional RL limitations like trial-and-error and reactive behavior.

Contribution

It introduces the concept of active reinforcement learning as a unified approach to improve self-adaptive systems by integrating multiple learning techniques.

Findings

Identifies limitations of traditional reinforcement learning in real-world applications.

Proposes a roadmap for developing active reinforcement learning systems.

Highlights potential benefits of proactive and integrated learning strategies.

Abstract

Intelligent systems have the ability to improve their behaviour over time taking observations, experiences or explicit feedback into account. Traditional approaches separate the learning problem and make isolated use of techniques from different field of machine learning such as reinforcement learning, active learning, anomaly detection or transfer learning, for instance. In this context, the fundamental reinforcement learning approaches come with several drawbacks that hinder their application to real-world systems: trial-and-error, purely reactive behaviour or isolated problem handling. The idea of this article is to present a concept for alleviating these drawbacks by setting up a research agenda towards what we call "active reinforcement learning" in intelligent systems.