Visualization-based interactive machine learning method based on visualization

Abstract

The invention discloses a visualization-based interactive machine learning method. The method comprises the steps: firstly abstracting neural network data into structural data, implicit value data andmodel evaluation data; according to attributes of the data, carrying out visual mapping and encoding through visual channels such as colors, positions and transparency; designing a colored circular visual code to represent a neural network structure; designing a visual code fusing the thermodynamic diagram and the parallel coordinates, and representing the change trend of the implicit state value; designing a time sequence confusion matrix visual code to represent a training result; completing global presentation in combination with multi-view linkage, wherein an implicit state value analysisview based on a thermodynamic diagram and parallel coordinates serves as a main view, training of the model can be evaluated through assistance of a time sequence confusion matrix view, conjecture can be rapidly verified through a network structure view, and new training data are generated for the next round of analysis. The functions of the three views are relatively independent and complementary to each other, and the interactive machine learning method based on visualization is jointly completed.

Description

technical field [0001] The invention relates to the technical fields of information visualization analysis and machine learning, in particular to a visualization-based interactive machine learning method. Background technique [0002] Machine learning is the science of how to use computers to simulate or realize human learning activities. It is one of the most intelligent and cutting-edge research fields in artificial intelligence. In the era of big data, with the continuous increase in the demand for data analysis in various industries, the efficient acquisition of knowledge through machine learning has gradually become the main driving force for the development of today's machine learning technology. How to conduct in-depth multi-dimensional analysis of complex and diverse data based on machine learning methods, and obtain hidden, effective, and understandable knowledge from massive data has become the main direction of machine learning research in the current big data env...

AI Technical Summary

Problems solved by technology

There is also less exploration of the information contained in hidden state values, mostly the discovery of certain rules, lack of work on optimizing neural network training based on the similarity of hidden state values, and the visual design of hidden state values ​​​​is not very good. Reflects the trend of changes in both individual neurons and overall neurons

Images