The big Microsoft Event for Developers was streamed last week during 48H « follow the sun » with a lot of content in AI (especially speech and vision). Here is some takeover if you missed the event:
BUILD 2020 : Conversational AI
- Bot Framework Composer is now GA – open source and available on Windows, Mac and Linux
- Virtual Assistant Solution Accelerator 1.0 moves to GA
- DirectLine App Service Extension now GA
- Public preview for new Amazon Alexa channel in Azure Bot Service now available
- New LUIS portal with dramatically improved labelling experience
- QnA Maker improvements – Rich Text Editor and role based access support (RBAC)
- And more…..
So a ton’s of news we already saw at the Global MVP Summit 2020. I’m still confused about Microsoft low-code UI not very sexy, that should be more inspired of Node-RED UI (yes I know IBM is a competitor)
- Conversational AI powered Customer and Employee Virtual Assistants
- Accelerate bot development in Power Virtual Agents
- Bot Framework Composer: Bot Framework’s new collaborative Conversational AI development environment
- Use the Efficiency of Low-Code with the Extensibility to Azure to Design World-Class Chatbots
- Conversational AI and human agents working together
BUILD 2020 : Cognitive Speech
Updates to the Speech Service:
- Speech to Text—Quickly transcribe audio to text. Speech to Text is expanding to 27 new locales (coming soon), with 30 percent average improvement in speech transcription accuracy.
- Neural Text to Speech—Converts text to lifelike speech for more natural interfaces. Neural TTS is extending support to 11 new locales with 15 new voices, with pronunciation error rate reduced by 50 percent for 13 locales, enabling more customers to benefit from a broad range of natural-sounding voices.
New preview features:
- Pronunciation assessment: The Speech service pronunciation assessment capability evaluates speech pronunciation and gives speakers feedback on the accuracy and fluency of spoken audio. With pronunciation assessment, language learners can practice, get instant feedback, and improve their pronunciation so that they can speak and present with confidence. Educators can use the capability to evaluate pronunciation of multiple speakers in real time. The feature currently supports American English and correlates highly with speech assessments conducted by experts.
- Speaker Recognition—Speaker Recognition is extending language support to 8 languages. Use either free-form speech or passphrases to verify speakers with the new text-independent speaker verification API. The new Speaker Recognition feature will be available on June 1, 2020. Documentation will be available here on June 1.
- Improve Customer Service with Call Center Analytics – by Panos Periorellis
- Building Voice Assistants with Custom Commands – by Vishesh Oberoi
- Conversational AI powered Customer and Employee Virtual Assistant
- Add AI to your solutions with Cognitive Services
- Unlock speech interactions and insights in 10 minutes
About AI & Cognitive Services (edit)
- Language Understanding (LUIS) and Text Analytics (TA) sentiment analysis 3.0 are now GA in containers
- Computer Vision READ API 3.0 GA with support of Spanish, German, French, Italian, Portuguese and Dutch.
- Azure Bot Services updates
- Announcing accelerated training with ONNX Runtime
- Use AI responsibly
One more thing …
Here is all the nerdy easter eggs during the show 🙂