Explainable Reinforcement Learning: Towards Trustworthy Autonomous Network Operations

Other conference contribution, 2024

Reinforcement learning represents a promising solution for complex network automation tasks. However, it also poses unique challenges when it comes to extracting explainable decisions. In this keynote, present an overview of the current architectures for network automation. Then, we explore how reinforcement learning has been used, with a specific use case example. Then, we discuss the challenges in extracting explanations, analyzing some preliminary results. Finally, the open challenges are discussed.