Machine learning process intelligent assembly method (for deep learning) based on semantic network

Abstract

The invention provides a machine learning process intelligent assembly method (for deep learning) based on a semantic network. The method comprises the steps of S1, randomly selecting several preprocessing methods from preprocessing methods supported by different frameworks to serve as preprocessing methods of image data in an automatic or manual mode, S2, selecting a network structure from modelstructures supported by different frameworks in an automatic or manual mode, S3, selecting an optimizer to be used in the training process from optimizers supported by different frameworks in an automatic or manual mode, and S4, establishing a semantic model to describe the functions of the algorithm modules of the different frameworks selected in the three steps, performing assembly, constructinga machine learning model training process, and performing model training and result evaluation. The intelligent assembly and training evaluation process of the process can be repeatedly carried out,and the model with the best effect is reserved as a final model. Thus, intelligent assembly and automatic exploration can be effectively carried out on a deep learning process in machine learning.

Description

technical field [0001] The invention relates to the technical field of machine learning and deep learning, in particular to a semantic web-based intelligent assembly method for machine learning processes (oriented to deep learning). Background technique [0002] Machine learning infrastructure is the core prerequisite for the widespread application and efficient integration of big data in various industries. However, the heterogeneity of machine learning algorithm frameworks, the diversity of machine learning process assembly, and the complexity of software and hardware resource management all pose great challenges to the construction of machine learning infrastructure. Contents of the invention [0003] The present invention proposes a semantic web-based machine learning process intelligent assembly method (oriented to deep learning), which can effectively perform intelligent assembly and automatic exploration for the deep learning process in machine learning. [0004] T...

Images