During the Ignite 2021 event in November this year, Microsoft unveiled its new ‘Azure OpenAI Services‘. Microsoft brings together OpenAI API and Azure enterprise-level security, compliance, and regional availability, giving Azure customers the ability to utilise OpenAI’s machine learning models. OpenAI’s API provides access to GPT-3, which performs a wide variety of natural language tasks, and Codex, which translates natural language to code.
The GPT-3 Benefit
According to Dominic Divakaruni, Microsoft group product manager leading Azure OpenAI, “The potential enterprise uses for GPT-3 range from summarising common complaints in customer service logs to helping developers code faster without having to stop and search for examples or generating new content as starting points for blog posts”.
Similarly, Codex is like GPT-3, but instead of being trained on natural languages, it is trained on code, which allows users to generate new code, and even create entire sections. GPT-3 derivative, Codex, helps developers write code more efficiently and avoid repetitive tasks with automatic code completion and suggestions.
Azure OpenAI Service offers customers direct access to GPT-3 in a format that is designed to be intuitive enough for developers to use yet robust enough for machine learning experts to work with the models as they wish. The service allows the GPT-3 model and its derivative to be used together to handle applications that require a deep understanding of language, for example – converting natural language into software code, summarising large amounts of text, and generating answers to questions.
Features of Azure OpenAI Services
Unlocking new language scenarios
The services allow customers to apply large, pretrained models to new use cases. The services provide easy access to the groundbreaking GPT-3 models that have been pretrained with trillions of words. This feature can be applied to new scenarios, from summarisation to content and code generation.
Tailoring models to specific needs
Models can be fine-tuned by using one’s data and adjusting hyperparameters to ensure accurate results. Furthermore, models can be customised with labelled data for a specific scenario using a simple REST API. Customers can also use the few-shot learning capability to provide the API with examples and achieve more relevant results.
Applying AI responsibly
Microsoft is ensuring ethical and responsible use of the model by limiting access to invitation-only services. Additionally, Azure services offer tools to empower customers to moderate generated content. Finally, guidance and implementation of best practices are provided to help customers design their applications while keeping safety front and centre. For example, large training models include racial stereotypes or vulgar languages; hence it is important for Microsoft to monitor how the services are being used.
Access enterprise-grade security
The services provide increased security through role-based access control (RBAC) and private networks. OpenAI Service runs on the Azure global infrastructure to meet the customer’s production needs, such as critical enterprise security, compliance, and regional availability. It provides the capacity to support the needs of a customer’s application and scale for demand over time. It makes deployment more secure and trusted with role-based authentication and private network connectivity. In addition, the services allow one to train their model with full control of their data.
