The articles in this section described the different aspects of the automatic learning of the models. The specific case of the automatic learning is AutoML.
AutoML is the concept of automating the process of applying machine learning to real-world problems. AutoML solves problems of parameter optimization, feature selection, model type selection, etc.
Generative design solves the problems of growing new data-driven models, new composite models, ensembles or other compositions from already existing models, etc. These days, ideas of low-level “assembly” of the model for custom problem setting are implemented only in the context of NAS (Neural Architecture Search) direction and mainly for problems of pattern recognition. However, for all other classes of tasks and models (the application of which may be more effective than neural networks), these approaches are not implemented.
The goal of the Generative AutoML implementation is to improve the quality of analysis of various natural, technical and social processes by identifying composite models based on available data sets.