Azure - Building Multimodal Generative Experiences. Part 2

Previous Post Link

Create a composed Document Intelligence Model

• Composed models in Azure AI document intelligence enable users to submit a form when they don't know which is the best model to use.
• Composed Models
  • when you have forms with unusual or unique formats, you can create and train your own custom models in Azure AI Document Intelligence.
  • You can create custom model of 2 types (custom template model and custom neural models) refer to previous post to know more about them.
  • Once you have created a set of custom models, you must assemble them into a composed model. you can do this on the Azure AI Studio.
  • Custom model Compatibility
  • Custom template models are responsible with other custom template models across 3.0 and 2.1 API versions
  • Custom neural models are composable with other custom neural models.
  • Custom neural models can't be composed with custom template models.
  • Custom models

Build a document intelligence custom skill for azure search.

• If you integrate AI Search with an Azure AI Document intelligence solution, you can enrich your index with fields that your Azure AI Document Intelligence models are trained to extract.
• Azure AI Search is a search service hosted in Azure that can index content on your permises or in a cloud location.

• There are 5 stages in Indexing process

  • Document Cracking. In document cracking, the indexer opens the content files and extracts their content.
  • Field Mappings. Fields such as titles, names, dates, and more are extracted from the content. You can use field mappings to control how they're stored in the index.
  • Skillset Execution. In the optional skillset execution stage, custom AI processing is done on the content to enrich the final index.
  • Output field mappings. If you're using a custom skillset, its output is mapped to index fields in this stage.
  • Push to index. The results of the indexing process are stored in the index in Azure AI Search.

• AI Search Skillset

  • Key Phrase extraction
  • Language Detection
  • Merge
  • Sentiment
  • Translation
  • Image Analysis
  • Optical character recognition

• we can use custom skills too and they can be used for 2 reasons

• The list of built-in skills doesn't include the type of AI Enrichment you need.

• you want to train your own model to analyze the data

• 2 types of custom skills that you can create

  • Azure Machine Learning Custom Skills
  • Custom Web API Skills Refer to this link for building an Azure AI Document Intelligence Custom Skill