AI ENRICHMENT IN AZURE COGNITIVE SEARCH

In this article we will cover the following :

• What is Azure AI Enrichment?
• Built-in skills.
• Applications.
• Steps in enrichment pipeline.
• Workflow.
• Resources.

What is Azure AI Enrichment?

It is an extension that can be used to extract text from images, blobs, and other unstructured data sources. Enrichment and extractions make your content more searchable. Extraction and Enrichment are applied using cognitive skills attached to the indexer driven pipeline.

Built-in skills

• Image Processing kills include Optical Character Recognition and identification of visual features.
• Natural Langauge Processing NLP skills include but not limited to keyphrase extraction, sentiment detection, etc.

Built-in skills in Azure Cognitive Search are based on machine learning models in API. Users can attach Cognitive Services resources.

Applications

• Users can attach an optical OCR skill to identify, extract, and ingest text from JPEG files.
• Users can apply language detection and possibly text translation.
• Built-in skills can be also used to restructured content through text split, and shape operations.

Steps in Enrichment Pipeline

1. Document Cracking phrase

2.Cognitive Skills and enrichment phrase

3. Search index and query-based access

Workflow

• Create Data Source Object
• Create Skillset
• Define Index Schema
• Define Indexer
• Add outputFieldMappins within the indexer
• Create an Indexer Request
• Run Queries
• Reset Indexer

Resources

https://docs.microsoft.com/en-us/azure/search/knowledge-store-create-rest

https://docs.microsoft.com/en-us/azure/search/search-howto-reindex

https://docs.microsoft.com/en-us/azure/cognitive-services/form-recognizer/overview