A Deep Neural Network Based Resource Configuration Framework for Human-Machine Computing System

The collaborative computing between humans and machines has reached a new level in recent decades being a result of the increasing convergence of technology and progress in different scientific fields. By combining the strengths of humans and machines, Human-Machine Computing (HMC) integrates the complex cognitive reasoning capabilities of humans and the high-performance computing capabilities of computer clusters to tackle complex tasks that are difficult to accomplish by machines alone. However, combining humans with intelligent machines to obtain more efficient scheduling and management is a nontrivial task. Considering the heterogeneity of human-machine resources, we proposed a deep neural network-based resource configuration framework for the HMC system. In particular, we firstly describe the architecture of the HMC system. On this basis, we present the modeling details of the human-machine computing resource. Secondly, we analyze the optimization problem to be solved by the framework and propose a deep neural network-based scheduler for solving the resource configuration problem. Finally, the performance of our proposed framework is evaluated through simulation experiments, and the results show that our solution can make remarkable effectiveness, which can serve as guidelines for future research on HMC systems.