Interpretable Modeling of Deep Reinforcement Learning Driven Scheduling

arXiv:2403.16293v1 Announce Type: new
Abstract: In the field of high-performance computing (HPC), there has been recent exploration into the use of deep reinforcement learning for cluster scheduling (DRL scheduling), which has demonstrated promising outcomes. However, a significant challenge arises from the lack of interpretability in deep neural networks (DNN), rendering them as black-box models to system managers. This lack of model interpretability hinders the practical deployment of DRL scheduling. In this work, we present a framework called IRL (Interpretable Reinforcement …

abstract arxiv challenge cluster computing cs.dc cs.lg dnn exploration however hpc interpretability modeling networks neural networks performance reinforcement reinforcement learning rendering scheduling them type