Real-world data is messy. It often spans media types (e.g. text documents, PDF files, images, databases), changes constantly, and carries valuable knowledge in ways that is not readily usable.
Cognitive Search adds data extraction, natural language processing (NLP), and image processing skills to an Azure Search indexing pipeline, making previously unsearchable or unstructured content more searchable. Information created by Cognitive Search Skills, such as entity recognition or image analysis, gets added to an index in Azure Search.
This solution alleviates the large effort needed to accomplish the typical solution pattern needed for this: ingest-enrich-explore. Sidesteps usual challenges like large scale change tracking to file format support, and even composition of multiple AI models. Today it takes a huge amount of effort, requires branching into multiple unrelated domains (from cracking PDFs to handling AI model composition). This is where Cognitive Search comes in.
This one day training will focus on hands-on activities that develop proficiency with Cognitive Search, an Azure Search AI-oriented capability. These labs assume an introductory to intermediate knowledge of Visual Studio, the Azure Portal, Azure Functions and Azure Search. If you are not at that skill level, we have prerequisite materials below that you need to complete prior to beginning this training.
We will focus on hands-on activities to learn how to create a Cognitive Search solution for all types of business documents. The documents include pdfs, docs, ppts and images, as well as documents with multiple languages. In this training, you will create a data flow that uses cognitive skills to enrich your business documents. These enrichments will become part of an Azure Search index.
At the end of this workshop, you should have learned:
- What Cognitive Search is
- How to implement this Cognitive Search Solution
- Why to use this solution with demos, POCs and other business scenarios
Since this is an AI training on top of Microsoft Azure Services, before we start you need:
- If if you don’t have prior experience:
- To Create: You need a Microsoft Azure account to create the services we use in our solution. You can create a free account, use your MSDN account or use any other subscription where you have permission to create services.
- To Install: Visual Studio 2017 version version 15.5 or later, including the Azure development workload.
- To Install: Postman. To call the labs APIs.
Since you have finished the prerequisites, let’s start the training. You just need to follow the workshop structure presented below.
- Introduction - 1 hour - Motivation, context, key concepts
- Solution Architecture - 1 hour - Diagram, use cases, deployment options and costs
- Environment Creation - 1 hour - Using the Azure Portal, we will create the services we need fo the workshop
- Lab 1 - 2 hours - Create a Cognitive Search Enrichment Process: Text Skills
- Lab 2 - 1 hour - Create a Cognitive Search Skillset: Image Skills
- Lab 3 - 2 hours - Create a Cognitive Search Skillset with Custom Skills
- Final Case - 0.5 hour - Brainstorm - Create a Cognitive Search Solution