Overview

This offering teaches you how IBM Watson OpenScale on IBM Cloud Pak for Data lets business analysts, data scientists, and developers build monitors for artificial intelligence (AI) models to manage risks. You will understand how to use Watson OpenScale to build monitors for quality, fairness, and drift, and how monitors impact business KPIs. You will also learn how monitoring for unwanted biases and viewing explanations of predictions helps provide business stakeholders confidence in the AI being launched into production. Note: This course contains the same topics as W7069G Watson OpenScale Methodology WBT.

Audience

Analysts, Developers, Data Scientists and others who need to monitor machine learning jobs

Prerequisites

• Basic knowledge of cloud platforms, for example IBM Cloud  • Basic understanding of machine learning models, and how they are used • IBM Cloud Pak for Data (V2.5.x): Foundations - 6X236G (recommended)

Objective

• Introduction to IBM Watson OpenScale • IBM Watson OpenScale architecture • Get started with IBM Watson OpenScale on IBM Cloud Pak for Data • Overview of Watson OpenScale monitors • Explore a use case • Build and configure the fairness monitor • Configure the quality monitor • Detect drift and configure the drift monitor • Configure application monitors

Course Outline

Introduction to IBM Watson OpenScale • Describe the problem that Watson OpenScale solves • Describe models, monitors, workflow • Describe AIF and AIE 360 toolkits • Describe workflow IBM Watson OpenScale architecture • Describe OpenScale architecture on IBM Cloud and on IBM Cloud Pak for Data • Describe how Watson OpenScale works with other cloud services Get started with IBM Watson OpenScale on IBM Cloud Pak for Data • Install the Watson OpenScale service • Work with Watson OpenScale on Cloud Pak for Data Overview of Watson OpenScale monitors • Identify the different Watson OpenScale monitors • Describe how the monitors are used Explore a use case • Prepare the model for monitoring Build and configure the fairness monitor • Features to monitor • Values  that represent a favorable outcome of the model • Reference and monitored groups • Fairness thresholds • Sample size • Insights and explainability Configure the quality monitor • Quality alert threshold • Sample size • Insights and explainability Detect drift and configure the drift monitor • Alert threshold • Sample size • Insights and explainability Configure application monitors • Configure application monitors • Configure KPI metrics in Watson OpenScale • Configure event details • Access and visualize custom metrics