New Healthcare AI Models, More Coming to Azure Cloud
As part of sweeping updates to its Cloud for Healthcare product, Microsoft this week unveiled a handful of AI capabilities coming to the Azure cloud designed specifically for medical scenarios.
The first is now generally available: support for more languages in Azure Text Analytics. Besides English, the natural language processing service can now interpret and analyze text written in in Spanish, French, German, Italian, Portuguese and Hebrew.
The other AI features are currently in preview. They include three new models in Azure AI Health Insights:
- Patient Timeline: Pulls date-related information from unstructured data using generative AI, and presents that data in chronological order so care providers get a more complete picture of a patient's care hisory.
- Clinical Report Simplification: Translates esoteric medical terminology into content that patients and their caregivers can more easily understand.
- Radiology Insights: As its name suggests, pulls data such as "follow-up recommendations and clinical findings" from radiology reports.
Also in preview are generative AI capabilities in the Azure AI Health Bot product. This means healthcare organizations can now build intelligent chatbots trained on their internal data and workflows, as well as data from outside sources like the FDA.
The Azure-powered Microsoft Fabric service is also getting a "de-identification" capability that uses machine learning to strip sensitive and private patient information from unstructured text. This is particularly useful for researchers who need to reference large volumes of patient data for their work while remaining compliant with HIPAA requirements.
Finally, the Text Analytics service, which uses machine learning to apply labels to parts of unstructured medical text, is getting more templates for "population health, patient population Q&A using Azure OpenAI Service, clinical trials patient cohorts and mass historic data processing."
Combined, these features "help health organizations improve patient experience, gain new insights with machine learning and AI, and handle health information securely," said Microsoft's Hadas Bitran in a blog post.
Microsoft made these Azure AI announcements on Tuesday, coinciding with the HLTH 2023 conference currently taking place in Las Vegas. The company also announced several healthcare-related data management upgrades coming to its Fabric product.