Merge pull request #1495 from MicrosoftDocs/release-ignite-2024-aoai

Release ignite 2024 aoai

Author: thomps23

articles/ai-services/openai/concepts/abuse-monitoring.md

@@ -13,19 +13,26 @@ manager: nitinme
 
 # Abuse Monitoring
 
-Azure OpenAI Service detects and mitigates instances of recurring content and/or behaviors that suggest use of the service in a manner that may violate the [Code of Conduct](/legal/cognitive-services/openai/code-of-conduct?context=/azure/ai-services/openai/context/context) or other applicable product terms. Details on how data is handled can be found on the [Data, Privacy and Security page](/legal/cognitive-services/openai/data-privacy?context=/azure/ai-services/openai/context/context). Apply for modified abuse monitoring using this form: [Azure OpenAI Limited Access Review: Modified Abuse Monitoring](https://ncv.microsoft.com/3a140V2W0l).
+Azure OpenAI Service detects and mitigates instances of recurring content and/or behaviors that suggest use of the service in a manner that might violate the [Code of Conduct](/legal/cognitive-services/openai/code-of-conduct?context=/azure/ai-services/openai/context/context) or other applicable product terms. Details on how data is handled can be found on the [Data, Privacy, and Security](/legal/cognitive-services/openai/data-privacy?context=/azure/ai-services/openai/context/context) page.
 
 ## Components of abuse monitoring
 
 There are several components to abuse monitoring:
 
-- **Content Classification**: Classifier models detect harmful language and/or images in user prompts (inputs) and completions (outputs). The system looks for categories of harms as defined in the [Content Requirements](/legal/cognitive-services/openai/code-of-conduct?context=/azure/ai-services/openai/context/context), and assigns severity levels as described in more detail on the [Content Filtering page](content-filter.md).
+- **Content Classification**: Classifier models detect harmful text and/or images in user prompts (inputs) and completions (outputs). The system looks for categories of harms as defined in the [Content Requirements](/legal/cognitive-services/openai/code-of-conduct?context=/azure/ai-services/openai/context/context), and assigns severity levels as described in more detail on the [Content Filtering](/azure/ai-services/openai/concepts/content-filter) page. The content classification signals contribute to pattern detection as described below.  
+- **Abuse Pattern Capture**: Azure OpenAI Service’s abuse monitoring system looks at customer usage patterns and employs algorithms and heuristics to detect and score indicators of potential abuse. Detected patterns consider, for example, the frequency and severity at which harmful content is detected (as indicated in content classifier signals) in a customer’s prompts and completions, as well as the intentionality of the behavior. The trends and urgency of the detected pattern will also affect scoring of potential abuse severity.
+    For example, a higher volume of harmful content classified as higher severity, or recurring conduct indicating intentionality (such as recurring jailbreak attempts) are both more likely to receive a high score indicating potential abuse. 
+- **Review and Decision**: Prompts and completions that are flagged through content classification and/or identified as part of a potentially abusive pattern of use are subjected to another review process to help confirm the system’s analysis and inform actioning decisions. Such review is conducted through two methods: human review & AI review.
+    - By default, if prompts and completions are flagged through content classification as harmful and/or identified to be part of a potentially abusive pattern of use, they may be sampled for automated, eyes-off review by using an LLM instead of a human reviewer. The LLM used for this purpose processes prompts and completions only to confirm the system’s analysis and inform actioning decisions; prompts and completions that undergo such LLM review are not stored by the system or used to train the LLM or other systems.
+    - In some cases, when automated review does not meet applicable confidence thresholds in complex contexts or if LLM review systems are not available, human eyes-on review may be introduced to make an extra judgment. This can help improve the overall abuse analysis accuracy. Authorized Microsoft employees may assess flagged content, and either confirm or correct the classification or determination based on predefined guidelines and policies. Prompts and completions can be accessed for human review only by authorized Microsoft employees via Secure Access Workstations (SAWs) with Just-In-Time (JIT) request approval granted by team managers. For Azure OpenAI Service resources deployed in the European Economic Area, the authorized Microsoft employees are located in the European Economic Area. This human review process will not take place if the customer has been approved for modified abuse monitoring. 
+- **Notification and Action**: When a threshold of abusive behavior has been confirmed based on the preceding steps, the customer is informed of the determination by email. Except in cases of severe or recurring abuse, customers typically are given an opportunity to explain or remediate—and implement mechanisms to prevent recurrence of—the abusive behavior. Failure to address the behavior—or recurring or severe abuse—may result in suspension or termination of the customer’s access to Azure OpenAI resources and/or capabilities.
 
-- **Abuse Pattern Capture**: Azure OpenAI Service’s abuse monitoring looks at customer usage patterns and employs algorithms and heuristics to detect indicators of potential abuse. Detected patterns consider, for example, the frequency and severity at which harmful content is detected in a customer’s prompts and completions.
+## Modified abuse monitoring 
 
-- **Human Review and Decision**: When prompts and/or completions are flagged through content classification and abuse pattern capture as described above, authorized Microsoft employees may assess the flagged content, and either confirm or correct the classification or determination based on predefined guidelines and policies. Data can be accessed for human review <u>only</u> by authorized Microsoft employees via Secure Access Workstations (SAWs) with Just-In-Time (JIT) request approval granted by team managers. For Azure OpenAI Service resources deployed in the European Economic Area, the authorized Microsoft employees are located in the European Economic Area.
+Some customers may want to use the Azure OpenAI Service for a use case that involves the processing of highly sensitive or highly confidential data, or otherwise may conclude that they do not want or do not have the right to permit Microsoft to store and conduct human review on their prompts and completions for abuse detection. To address these concerns, Microsoft allows customers who meet additional Limited Access eligibility criteria to apply to modify abuse monitoring by completing [this ](https://customervoice.microsoft.com/Pages/ResponsePage.aspx?id=v4j5cvGGr0GRqy180BHbR7en2Ais5pxKtso_Pz4b1_xUOE9MUTFMUlpBNk5IQlZWWkcyUEpWWEhGOCQlQCN0PWcu)form. Learn more about applying for modified abuse monitoring at [Limited access to Azure OpenAI Service](/legal/cognitive-services/openai/limited-access?context=%2Fazure%2Fai-services%2Fopenai%2Fcontext%2Fcontext), and about the impact of modified abuse monitoring on data processing at [Data, privacy, and security for Azure OpenAI Service](/legal/cognitive-services/openai/data-privacy?context=%2Fazure%2Fai-services%2Fopenai%2Fcontext%2Fcontext&tabs=azure-portal).    
 
-- **Notification and Action**: When a threshold of abusive behavior has been confirmed based on the preceding three steps, the customer is informed of the determination by email. Except in cases of severe or recurring abuse, customers typically are given an opportunity to explain or remediate—and implement mechanisms to prevent recurrence of—the abusive behavior. Failure to address the behavior—or recurring or severe abuse—may result in suspension or termination of the customer’s access to Azure OpenAI resources and/or capabilities.
+> [!NOTE]
+> When abuse monitoring is modified and human review is not performed, detection of potential abuse may be less accurate. Customers will be notified of potential abuse detection as described above, and should be prepared to respond to such notification to avoid service interruption if possible.  
 
 ## Next steps
 

articles/ai-services/openai/concepts/provisioned-migration.md

@@ -6,7 +6,7 @@ manager: nitinme
 ms.service: azure-ai-openai
 ms.custom: 
 ms.topic: how-to
-ms.date: 08/23/2024
+ms.date: 11/11/2024
 author: mrbullwinkle
 ms.author: mbullwin
 recommendations: false
@@ -250,7 +250,7 @@ With your commitment plan ready, the next step is to create the commitments. Com
 
 For each new commitment you need to create, follow these steps:
 
-1. Launch the Provisioned Throughput purchase dialog by selecting  **Quotas** > **Provisioned** > **Manage Commitments**.
+1. Launch the Provisioned Throughput purchase dialog by selecting  **Quota** > **Azure OpenAI Provisioned** > **Manage Commitment plans**.
 
 :::image type="content" source="../media/how-to/provisioned-onboarding/quota.png" alt-text="Screenshot of the purchase dialog." lightbox="../media/how-to/provisioned-onboarding/quota.png":::
 
@@ -301,7 +301,7 @@ Commitment renewal settings can be changed at any time before the expiration dat
 
 ## Monitor commitments and prevent unexpected billings
 
-The manage commitments pane provides a subscription wide overview of all resources with commitments and PTU usage within a given Azure Subscription. Of particular importance interest are:
+The manage commitments pane provides a subscription wide overview of all resources with commitments and PTU usage within a given Azure Subscription. Of particular importance are:
 
 - **PTUs Committed, Deployed and Usage** – These figures provide the sizes of your commitments, and how much is in use by deployments. Maximize your investment by using all of your committed PTUs.
 - **Expiration policy and date** - The expiration date and policy tell you when a commitment will expire and what will happen when it does. A commitment set to autorenew will generate a billing event on the renewal date. For commitments that are expiring, be sure you delete deployments from these resources prior to the expiration date to prevent hourly overage billingThe current renewal settings for a commitment. 

articles/ai-services/openai/concepts/use-your-data.md

@@ -8,7 +8,7 @@ ms.service: azure-ai-openai
 ms.topic: quickstart
 author: aahill
 ms.author: aahi
-ms.date: 04/08/2024
+ms.date: 10/25/2024
 recommendations: false
 ms.custom: references_regions 
 ---
@@ -26,14 +26,14 @@ Azure OpenAI On Your Data enables you to run advanced AI models such as GPT-35-T
 :::image type="content" source="../media/use-your-data/workflow-diagram.png" alt-text="A diagram showing an example workflow.":::
 
 Typically, the development process you'd use with Azure OpenAI On Your Data is:
-1. **Ingest**: Upload files using either Azure OpenAI Studio or the ingestion API. This enables your data to be cracked, chunked and embedded into an Azure AI Search instance that can be used by Azure Open AI models. If you have an existing [supported data source](#supported-data-sources), you can also connect it directly.
+1. **Ingest**: Upload files using either Azure OpenAI Studio or the ingestion API. This enables your data to be cracked, chunked and embedded into an Azure AI Search instance that can be used by Azure OpenAI models. If you have an existing [supported data source](#supported-data-sources), you can also connect it directly.
 
 1. **Develop**: After trying Azure OpenAI On Your Data, begin developing your application using the available REST API and SDKs, which are available in several languages. It will create prompts and search intents to pass to the Azure OpenAI service.
 
 1. **Inference**: After your application is deployed in your preferred environment, it will send prompts to Azure OpenAI, which will perform several steps before returning a response:
     1. **Intent generation**: The service will determine the intent of the user's prompt to determine a proper response.
 
-    1. **Retrieval**: The service retrieves relevant chunks of available data from the connected data source by querying it. For example by using a semantic or vector search. [Parameters](#runtime-parameters) such as strictness and number of documents to retreive are utilized to influence the retrieval.
+    1. **Retrieval**: The service retrieves relevant chunks of available data from the connected data source by querying it. For example by using a semantic or vector search. [Parameters](#runtime-parameters) such as strictness and number of documents to retrieve are utilized to influence the retrieval.
 
     1. **Filtration and reranking**: Search results from the retrieval step are improved by ranking and filtering data to refine relevance.    
 
@@ -307,6 +307,55 @@ Mapping these fields correctly helps ensure the model has better response and ci
 
 Along with using Elasticsearch databases in Azure OpenAI Studio, you can also use your Elasticsearch database using the [API](../references/elasticsearch.md). 
 
+# [MongoDB Atlas (preview)](#tab/mongo-db-atlas)
+
+You can connect your MongoDB Atlas vector index with Azure OpenAI On Your Data for inferencing. You can use it through the Azure AI Studio, API and SDK.
+
+### Prerequisites 
+
+* A [MongoDB Atlas account](https://account.mongodb.com/account/register)
+* An [Azure OpenAI ada002 embedding model](./models.md#embeddings)
+    * To achieve good retrieval quality, make sure your vector index is created with Azure OpenAI ada002 embedding model.
+
+We recommend using one of the following models for MongoDB Atlas
+* gpt-4 (0613)
+* gpt-4 (turbo-2024-04-09) 
+* gpt-4o (2024-05-13) 
+* gpt-35-turbo (1106) 
+
+### Configuration  
+
+Only public network access is supported. Please make sure the database allows public access 
+:::image type="content" source="../media/use-your-data/mongo-db-network-access.png" alt-text="A screenshot showing the network access screen for Mongo DB.":::
+
+### Data preparation
+
+If you want to create a new vector search index with your documents, you can use the [available script on GitHub](https://github.com/microsoft/sample-app-aoai-chatGPT/blob/rawan/mongodbdataprep/scripts/mongo_vector_db_data_preparation.py) to prepare your data for use with Azure OpenAI On Your Data. 
+
+### Connection to MongoDB account
+
+To add your data source, you first need to create a connection to MongoDB Atlas. This connection includes information such as authentication (username and password). Enter the endpoint of your MongoDB Atlas connection string using the following format: `mongodb+srv://{user_name}:{password}@{endpoint}/?appName={application_name}`. See the [MongoDB documentation](https://aka.ms/mongodb-connection-string) for more information about connection string methods.
+
+:::image type="content" source="../media/use-your-data/mongo-db-atlas-connection.png" alt-text="A screenshot showing the connection screen for Mongo DB." lightbox="../media/use-your-data/mongo-db-atlas-connection.png":::
+
+### Source index
+
+Once you have created a connection or chosen an existing connection, you can enter the information to connect to a specific vector index within this connected account. You need to input the name of your database, collection and vector index. Make sure you have entered the information correctly to successfully build the connection.
+
+:::image type="content" source="../media/use-your-data/mongo-db-atlas-source-index.png" alt-text="A screenshot showing the field required information for Mongo DB Atlas." lightbox="../media/use-your-data/mongo-db-atlas-source-index.png":::
+
+To use MongoDB Atlas, you'll need an Azure OpenAI ada002 embedding model. This model will be created for you if you don't already have one, which will incur [usage](https://go.microsoft.com/fwlink/?linkid=2264246) on your account.
+
+### Index field mapping
+
+When you add your MongoDB Atlas data source, you can specify data fields to properly map your data for retrieval.
+
+* Content data (required): This is the main text content of each document. For multiple fields, separate the values with commas, with no spaces.
+* Vector field (required): The field name in your MongoDB Atlas search index that contains the vectors.
+* File name/title/URL: Used to display more information when a document is referenced in the chat.
+
+:::image type="content" source="../media/use-your-data/mongo-db-atlas-field-mapping.png" alt-text="A screenshot showing the field mapping options for Mongo DB Atlas." lightbox="../media/use-your-data/mongo-db-atlas-field-mapping.png":::
+
 ---
 
 ## Deploy to a copilot (preview), Teams app (preview), or web app