Microsoft extends Artificial Intelligence in Azure
Cognitive services are intended to make artificial intelligence accessible to developers without machine learning skills. Microsoft has introduced four new services that analyze text, work with forms, and communicate with users in natural language.
According to IDC market researchers, three out of four companies will use solutions based on artificial intelligence by 2022. They hope that this will help them improve operational efficiency and customer service. Microsoft wants to enable companies of all sizes and in all industries to use artificial intelligence simply and comprehensively. To this end, the provider has added four services to its more than 30 existing Cognitive Services.
Azure Cognitive Services are interfaces for the development of intelligent apps. They enable machines to see, hear, cognitively capture their environment and convert it into machine-readable data. According to Microsoft, developers do not need machine learning skills to work with these interfaces. More than 30 of these services are currently available in the areas of decision making, speech recognition and input, image analysis and web search. Microsoft has now introduced four new services.
Text analytics, form recognition and speech processing
The Text Analytics for Health feature enables researchers and healthcare organizations to use unstructured medical data to make informed decisions. In response to the Corona crisis and in collaboration with the Allen Institute of AI, Microsoft has created a free collection of more than 47,000 scientific documents, the COVID 19 Open Research Dataset. In addition, Microsoft offers a newly developed COVID-19 cognitive search engine to help researchers gain new insights from researching and combating the corona virus.
Advances in Natural Language Processing (NLP) make it possible for Microsoft to offer an opinion mining feature for text analysis. This recognizes moods in texts and thus enables, for example, more precise analysis of customer voices in social media.
According to the company, the Form Recognizer, which is available immediately, is capable of recognizing unstructured data contained in forms with tables, objects and other elements, for example. Normally, companies need manual classification for this type of data.
Communication via 15 natural sounding voices
The Custom Commands are now generally available. This would enable developers to integrate customer-specific speech functions into applications that process spoken language ("Speech to Text") and understand ("Language Understanding"). The applications communicate with their users via voice response and text to speech. According to Microsoft, custom commands require relatively little programming effort ("low code").
The Neural Text to Speech function extends the linguistic capabilities of Cognitive Services by 15 natural sounding voices based on modern models for neural speech synthesis: Arabic (Egypt, Saudi Arabia), Catalan (Spain), Danish (Denmark), English (India), Hindi (India), Dutch (Netherlands), Polish (Poland), Portuguese (Portugal), Russian (Russia), Swedish (Sweden), Thai (Thailand), Chinese (Cantonese, Traditional and Taiwanese Mandarin).
The first release can be found at www.isreport.de.
With kind permission of the author we were allowed to publish the article in this blog.