sjkp · June 2, 2024 16:42
diff --git a/gistfile1.txt b/gistfile1.txt
 New Microsoft Azure reference architectures and implementation guidance are now generally available for customers to confidently design and deploy intelligent apps. Customers can easily leverage patterns and practices to create private chatbots that are reliable, cost-efficient and compliant — adhering to both the functional and nonfunctional requirements of an organization.

 The new guidance helps customers adopt well-architected best practices and includes:

 A reference architecture and reference implementation for Microsoft Azure OpenAI Service based on Azure landing zones, which helps jumpstart and scale app deployment.
 Service guides for machine learning that gives precise configuration instructions for Azure services used to deliver intelligent apps.
 Patterns for designing and developing a RAG solution: While the architecture is straightforward, designing, experimenting with, and evaluating RAG solutions that fit into this architecture involves many complex considerations that benefit from a rigorous, scientific approach.
 Additional resources:

 Email: Contact the Microsoft Media and Analyst Events Team for more information
 Breakout: Take an Azure OpenAI Service chat application from PoC to enterprise-ready
 1.1.2. ANNOUNCING CUSTOM GENERATIVE MODE IN PREVIEW SOON
 Custom generative is a new model type, coming soon to preview, that will start with a single document and will guide the user through the schema definition and model creation process with minimal labeling, allowing the user to process complex documents with a variety of formats and templates.

 The model will use large language models (LLMs) to extract the fields, and users will only need to correct the output when the model does not get a field right. The model will adapt to each sample added to the training dataset. Add new labeled documents and rebuild the model to continually improve the model after deployment.

 Additional resources:

 Email: Contact the Microsoft Media and Analyst Events Team for more information
 Breakout: Revamping the Document Automation Workflow with Generative AI
 Breakout: Going big with multimodal GenAI experiences with Azure AI
 1.1.3. AZURE AI SEARCH FEATURES SEARCH RELEVANCE UPDATES AND NEW INTEGRATIONS
 Microsoft Azure AI Search is a full-featured information retrieval system built to run superior retrieval-augmented generation (RAG) and enterprise search. Users can streamline indexing and development with deep data and platform integrations and scale major workloads on an enterprise-ready foundation. State-of-the-art search technology like hybrid search and re-ranking helps deliver the best experience for every user and interaction. Azure AI Search has dramatically increased storage capacity and vector index size for new services at no additional cost, helping customers scale their generative AI apps without compromising cost or performance. New services will have additional compute to support more vectors at high performance.

 Updates to Azure AI Search, in preview now, include:

 Performing RAG at scale with more capacity: Vector search will support binary vector types and other vector search features, helping customers improve storage efficiency.
 Returning more relevant results with enhancements to vector and hybrid search. These new capabilities for vector and hybrid search, including vector weighting, score threshold control and allowing maximum text recall size, will give customers more options and flexibility to improve the accuracy of their responses.
 New seamless data and processing integrations with updates to integrated vectorization, now with built-in image vectorization via Microsoft Azure AI Vision, the latest Microsoft Azure OpenAI Service embedding models, and additional models available in Microsoft Azure AI Studio model catalog. Customers will be able to easily process, vectorize and search images natively, not just text embeddings, in Azure AI Search.
 New platform integration with Azure AI Search’s OneLake connector for files. Organizations will be able to directly connect their data in Microsoft Fabric to Azure AI Search with the new integration with OneLake, expanding the range of data sources that can be indexed and searched
	New Microsoft Azure reference architectures and implementation guidance are now generally available for customers to confidently design and deploy intelligent apps. Customers can easily leverage patterns and practices to create private chatbots that are reliable, cost-efficient and compliant — adhering to both the functional and nonfunctional requirements of an organization.

	The new guidance helps customers adopt well-architected best practices and includes:

	A reference architecture and reference implementation for Microsoft Azure OpenAI Service based on Azure landing zones, which helps jumpstart and scale app deployment.
	Service guides for machine learning that gives precise configuration instructions for Azure services used to deliver intelligent apps.
	Patterns for designing and developing a RAG solution: While the architecture is straightforward, designing, experimenting with, and evaluating RAG solutions that fit into this architecture involves many complex considerations that benefit from a rigorous, scientific approach.
	Additional resources:

	Email: Contact the Microsoft Media and Analyst Events Team for more information
	Breakout: Take an Azure OpenAI Service chat application from PoC to enterprise-ready
	1.1.2. ANNOUNCING CUSTOM GENERATIVE MODE IN PREVIEW SOON
	Custom generative is a new model type, coming soon to preview, that will start with a single document and will guide the user through the schema definition and model creation process with minimal labeling, allowing the user to process complex documents with a variety of formats and templates.

	The model will use large language models (LLMs) to extract the fields, and users will only need to correct the output when the model does not get a field right. The model will adapt to each sample added to the training dataset. Add new labeled documents and rebuild the model to continually improve the model after deployment.

	Additional resources:

	Email: Contact the Microsoft Media and Analyst Events Team for more information
	Breakout: Revamping the Document Automation Workflow with Generative AI
	Breakout: Going big with multimodal GenAI experiences with Azure AI
	1.1.3. AZURE AI SEARCH FEATURES SEARCH RELEVANCE UPDATES AND NEW INTEGRATIONS
	Microsoft Azure AI Search is a full-featured information retrieval system built to run superior retrieval-augmented generation (RAG) and enterprise search. Users can streamline indexing and development with deep data and platform integrations and scale major workloads on an enterprise-ready foundation. State-of-the-art search technology like hybrid search and re-ranking helps deliver the best experience for every user and interaction. Azure AI Search has dramatically increased storage capacity and vector index size for new services at no additional cost, helping customers scale their generative AI apps without compromising cost or performance. New services will have additional compute to support more vectors at high performance.

	Updates to Azure AI Search, in preview now, include:

	Performing RAG at scale with more capacity: Vector search will support binary vector types and other vector search features, helping customers improve storage efficiency.
	Returning more relevant results with enhancements to vector and hybrid search. These new capabilities for vector and hybrid search, including vector weighting, score threshold control and allowing maximum text recall size, will give customers more options and flexibility to improve the accuracy of their responses.
	New seamless data and processing integrations with updates to integrated vectorization, now with built-in image vectorization via Microsoft Azure AI Vision, the latest Microsoft Azure OpenAI Service embedding models, and additional models available in Microsoft Azure AI Studio model catalog. Customers will be able to easily process, vectorize and search images natively, not just text embeddings, in Azure AI Search.
	New platform integration with Azure AI Search’s OneLake connector for files. Organizations will be able to directly connect their data in Microsoft Fabric to Azure AI Search with the new integration with OneLake, expanding the range of data sources that can be indexed and searched