This is a revolutionary product enabling data oriented, multi model, neural network designs for every data owner. With your dataset, sql or nosql tables, you can create huge neural network ensembles to simulate a data defined environment. There are multiple good use cases for these ensembles, for example:
1. These ensembles can be used to increase accuracy or reduce error of a table data. You can create multiple models for the same table and label column. Then make queries to generate a result table for each data line. With the generated results, train a reducer model to complete the ensemble.
2. Different models or ensembles can be created to simulate a multi table environment such as a complete database with many tables or a combination of different data sources such as image, sound and a feature table together.
3. Existing machine learning practices can be simulated. Such as Generative Adversarial Network, LSTM, deep reinforcement learning simulator ensembles can be created.
Start trying bicedeep data science ai to discover more use cases.