Bring Your Own Intelligence

This release introduces ONE's new Bring Your Own Intelligence suite of features which allows customers to connect their own artificial intelligence (AI) models with NEO. This suite of features enables NEO to make decisions using data gathered from machine learning (ML). Customers use or create pre-trained ML models for NEO to use in making decisions based on past data. The ML models are trained using one of ONE's supported ML frameworks; examples of currently supported frameworks include scikit-learn (random forest) and light gradient boost.

The machine learning decision points supported in NEO are called ML plug points and central to these features are Machine Learning (ML) Models and Machine Learning (ML) Pipelines. NEO's ML Pipelines feature enables users to build, manage, and operate their ML Pipelines inside NEO's UI.

The following provides an overview of the steps to leverage machine learning (ML) to make decisions in the ONE system. Following the steps, the links to specific workflows for the machine learning process are listed as well as definitions for common ML terms.

Bring Your Own Intelligence (Machine Learning) Workflow Overview

  1. Prior to the steps below, the data science team must build, train, and save an ML serialized model. Users must have the created serialized, trained model file saved to their computer before proceeding.

  2. Using the serialized, trained model file, create an ML model in NEO.

  3. Add, view, or edit the schema on the ML Models screen.

  4. From the ML Plug Points screen, create a new plug point for the appropriate subnet using the desired ML model. Subnets include an enterprise, an organization, or a specific site.

  5. Once the new plug point is created, click the icon for the ML Pipeline Editor to view or edit the ML Pipeline for that plug point.

  6. From the ML Pipeline Editor screen, add nodes to the pipeline as required to adapt the data between ONE's inputs and the ML model's inputs. On the ML Pipeline Editor screen, users can also upload files (if necessary) and test the pipeline. Users can view the Node Log to view the log from a test execution of the pipeline.

Machine Learning Definitions

Machine Learning Plug Points

Plug points are situations where users can specify that machine learning should be used to make a decision. As a prerequisite, users must have a trained machine learning serialized model from the data science team to create a plug point.

Machine Learning Models

A machine learning model is a model that is trained to recognize certain types of patterns using sample data and an algorithm. The model is provided with a set of data and an algorithm and is trained to recognize the desired patterns over time. The training process continues until the model achieves a desired level of accuracy on the training data.

Machine Learning Pipelines

Pipelines are an end-to-end framework that directs data flow into and out of a machine learning model.

Machine Learning Schemas

ML Schemas are a set of classes, properties, and restrictions, including input and output formats, related to the information used by machine learning algorithms.