GridGain Continuous Learning Framework

Continuous Learning Framework for In-Memory Machine and Deep Learning

The GridGain Continuous Learning Framework includes the Apache Ignite integrated machine learning and multilayer perceptron for deep learning features. As part of the GridGain in-memory computing platform, these features make continuous learning based on the in-memory operational dataset possible using machine learning and deep learning directly in GridGain. These machine learning and deep learning libraries have been optimized for massively parallel processing (MPP) against the data in the GridGain cluster. By leveraging MPP, large-scale machine learning can be greatly accelerated to train models much faster.

A continuous learning workflow can be established by processing data directly in the GridGain cluster. This offers the ability to train models without moving transactional data into a separate database before model training. GridGain offers real-time model training or even continuous model training with less complexity and substantially lower cost than traditional approaches which rely on an ETL process to load data into an analytical database for training.

Continuous Learning Framework for In-Process HTAP

The GridGain Continuous Learning Framework can be used as a building block for in-process HTAP (hybrid transactional/analytical processing) applications which continuously train a data model based on incoming data. For next generation applications, in-process HTAP offers the ability to react to and benefit from real-time model training which can power better real-time decision making. This can enable a new range of business applications built on real-time model updates such as fraud prevention, ecommerce recommendation engines, credit approvals, logistics, and transportation system maintenance decisions.

The GridGain Continuous Learning Framework is included in the GridGain Community Edition 2.7 and higher and the GridGain Enterprise and Ultimate Editions 8.4 and higher.

Continuous Learning Framework

For further information,
contact GridGain Systems →