Deep Reinforcement Learning (DRL) is a method where an agent, through trial and error, interacts with an environment to gain knowledge about it. DRL has proven to be a successful candidate solution for optimizing real-time networking decisions. However, these DRL agents are closed boxes and inherently difficult to explain, which prevents their use in production settings. In our work, we take a step forward towards removing this critical barrier with SYMBXRL, a novel technique that uses Symbolic AI to generate human-interpretable explanations, where key concepts and their relationships are expressed by intuitive symbols and logical rules.
I am pursuing my PhD with the Network Data Science and Resilient AI Networks groups at IMDEA Networks. My research interests include Reinforcement Learning and Explainable AI.
This event will be conducted in English