Intelligent Storage System of Machine Learning Model Based on Task Similarity

With the closer integration of database and machine learning, machine learning task in database can reduce the data transmission, thus dramatically boosting the runtime performance of the whole task. Moreover, if there is a chance of storing machine learning models involved in similar tasks in the system intelligently, the computation resource and time cost of repeated training will be greatly reduced. However, the intelligent storage system of machine learning model has not been developed yet. In order to achieve this goal, a method is proposed to measure the similarity of machine learning tasks. Second, the intelligent storage system of machine learning model was designed to manage models. Finally, it introduced the overall architecture and key technologies of intelligent storage system of machine learning model based on task similarity (ISSMLM), and describe three demonstration scenarios of the system. The results show the validity of the proposed method.