NashTech Insights

Azure OpenAI service introduction

Phi Huynh
Phi Huynh
Table of Contents
Cloud patterns south of the Aleutian Islands

What is Azure OpenAI service?

Azure OpenAI service is a great new AI service that help us working with OpenAI service in a protected and secured environment in Azure platform. Azure OpenAI Service provides REST API access to OpenAI’s powerful language models, including the GPT-3, Codex, and Embeddings model series. Additionally, the new GPT-4 and ChatGPT (gpt-35-turbo) model series have now reached general availability. These models can be easily adapted to specific tasks, such as content generation, summarization, semantic search, and natural language to code translation. Users can access the service through REST APIs, Python SDK, or our web-based interface in the Azure OpenAI Studio.

Azure OpenAI service is part of the Azure AI, that includes other services such as Azure Cognitive Services, Azure Machine Learning, and Azure Bot Service, etc. By using Azure OpenAI service, you can leverage the power of large-scale models with the reliability & high security of Azure .

Azure OpenAI Service in Azure AI ecosystem
Azure OpenAI Service in Azure AI ecosystem

What can you do with Azure OpenAI service?

There are lot of use cases that we can use Azure OpenAI service, such as customer support, code generation, and query over your own data. With Azure OpenAI service, you can:

  • Access the latest and most advanced language models from OpenAI, including GPT-3.5/GPT-4, Codex, DALL-E 2, Embeddings, etc..
  • Customize and fine-tune these models with your own data and hyperparameters to achieve more relevant and accurate results for your specific scenarios.
  • Use the zero shot, one shot and few-shot learning capability (with prompting engineering) to provide the API with examples and get outputs without any additional training.
  • Benefit from the enterprise-grade security, reliability, and compliance of Azure, with role-based authentication, private network connectivity, and data control .

Some examples of use cases with the name of the models are:

  • Writing assistance: You can use gpt3.5/gpt-4 models to generate high-quality text for various domains and purposes, such as blog posts, product descriptions, summaries, headlines, etc. You can also use the Embeddings model to find similar or related texts based on semantic similarity. I’ll introduce about Vector Database for similarity search in another post.
  • Code generation: You can use Codex models to translate natural language into code for different programming languages and frameworks, such as Python, JavaScript, HTML, CSS, etc. Additionally, you can use Codex models to generate documentation, comments, or tests for your code .
  • Reasoning over data: You can use gpt3.5/gpt-4 models to perform complex queries and analysis over structured or unstructured data, such as tables, graphs, documents, etc. You can also use Codex models to generate SQL queries from natural language .

Is Azure OpenAI service more secure than OpenAI API?

Security is one of key concerns while using large models that can generate and process sensitive data. However, Azure OpenAI service provides many advanced features over OpenAI API in terms of security and compliance.

Azure OpenAI service runs on the Azure global infrastructure, which meets the highest standards of reliability, availability, and data protection. Azure OpenAI service supports virtual network and private link connectivity, which allows you to isolate your network traffic from the public internet and ensure that only authorized users can access your AI models. You can also use Azure Active Directory to manage user identities and roles.

Additionally, Azure OpenAI service follows Microsoft’s principles for responsible AI use, which include fairness, reliability, privacy, security, inclusiveness, transparency, and accountability. Microsoft has invested in content filtering and moderation features to help you detect and mitigate harmful use of AI models. You can also leverage Microsoft’s responsible AI guidance and tools to assess and mitigate potential risks and harms of your AI applications.

Microsoft Responsible AI Principles
Microsoft Responsible AI Principles

Azure regions support Azure OpenAI service

For now, Azure OpenAI service is available in the following regions: West Europe, East US and South Central US. Hope Microsoft will expand this service to other regions soon.

Only West Europe, East US and South Central US support Azure OpenAI service (updated on 21th May 2023)
Phi Huynh

Phi Huynh

Technical Manager

Leave a Comment

Your email address will not be published. Required fields are marked *

Suggested Article

%d bloggers like this: