SmartMLchain implements a patented model, CALSTATDN, that uses a calculus-based approach (e.g., convolutions in an image using filters is a calculus-based approach) for model training to generate aggregated (e.g., integration in calculus) outputs. The aggregated outputs undergo statistics-based computations to generate scores for predictions and eventually stored as properties in normalized tables in a relational database. These steps of calculus-based and statistics-based computing, and data normalization can have multiple iterations.
This CALSTATDN model can be further extended for collusions in a network leading to a technique which combines deep learning with blockchain in a decentralized autonomous organization. We call this technique Deep Chain Learning (DCL). The technology uses the concept of blockchain with smart contracts, where each block in a blockchain can represent a component neural network, reading/writing data from/to relational databases. The CALSTATDN model can add a lot of value to DCL.
About the cAlstatdn model
SmartMLchain has implemented a patented model, CALSTATDN (US Patent# 10176435) for machine learning by iterating over a sequence of computing methods of calculus (CAL), statistics (STAT) and database normalization (DN) respectively, to reduce error and processing times of extremely large volumes of streaming data. The model has been applied to a Smart Home Analytics system to improve performance by several orders of magnitude.