In order to build this cloud and the advantages of AI, the giants usually show their technical strength through a grand conference, AWS has the largest re:invent conference of the year, Google has Google cloud next conference, and Microsoft developers for developers The conference is the best stage for its "show muscles".
However, before the opening of Microsoft Build 2019, which began on May 6th, US time, Microsoft officials could not wait to announce a series of "heavy" products and technologies on the official website. Of course, the release of the site may be even more important, but the information released before the meeting is still worthy of attention, especially the functions related to AI.
Cognitive services re-expanded, more personalized
Microsoft Azure's Cognitive Services is a collection of APIs for intelligent algorithms that developers can use to perform image recognition, speech recognition, natural language processing, anomaly detection, and other intelligent tasks.
Today, Microsoft announced a new set of pre-built machine learning models for its cognitive services platform, including APIs for building personalized features, form recognizers for automated data entry, handwriting recognition APIs, and focus on conversational transcription Enhanced speech recognition service.
The API for personalization is unique. After all, this is a rare thing, because the data in various information silos is difficult to get through, and it is very difficult to build data on it. However, Microsoft is now betting on reinforcement learning, which does not require markup training data that is commonly used for machine learning. Instead, it will continually try to find the best way to achieve a given goal based on user behavior. Microsoft said it is the first in its industry (alwaysXboxTested).
Conversation transcription is part of Microsoft's existing speech-to-text functionality in the cognitive services lineup, which can tag different speakers, transcribe conversations in real time, and even handle crosstalk. It has been integrated with Microsoft Teams and other conferencing software.
First launched the edge database Azure SQL
In the field of IoT and edge computing, Microsoft is introducing a new variant of the Azure SQL database called Azure SQL Edge Database, which will run on edge computing devices.
Microsoft said the new database shares the same core programming environment as Azure SQL Database and SQL Server. Azure SQL Database Edge includes support for ARM and x64-based devices and edge gateways, and will support cloud-attached and fully disconnected edge devices. Developers can request access to the Early Adopter Program to access the preview.
Scott Guthrie, executive vice president of Microsoft's cloud computing and AI team, said:
"Smart cloud and smart edge applications have evolved from low-computing IoT devices that primarily use cloud computing to powerful computing at the edge, which requires a new modern hybrid application approach... this cross-cloud and edge database The consistency of the programming and control planes is essential to not only run safely but also manage well-mixed applications."
Launched IoT Plug and Play Software to remove barriers to IoT development
Microsoft will also introduce IoT plug-and-play software, officially called a new open modeling language, developers do not need to write any code.
The principle is that users can plug in almost any peripheral deviceWindowsPC without having to worry about driver and IoT device compatibility issues. Although it is common to start and run with hardware and software configuration issues, Microsoft is committed to simplifying this process.
Microsoft Azure VP said that one of the obstacles to the development of IoT is the heterogeneity of today's global devices, which can't really connect millions of IoT devices to the cloud because they exist in form, processing power, memory, operating system, etc. difference.
In order to do this, Microsoft has signed cooperation with Aso, Brainium, Compal, and Kyocera's IoT upstream and downstream vendors to create their own "ecology" in hardware and software.
Microsoft has created a website dedicated to IoT plug-and-play. VIA driving recorders, industrial sensor network SenceCAP, IoT endpoint SensorTile.box have been displayed on the homepage to attract more partners.
Deepen machine learning in a "no code" way
Microsoft has also added some new features for machine learning to help developers build high-quality models more easily, and it's commendable to introduce a "no code" approach.
The new Azure ML interface makes creating a model as easy as importing a dataset and telling the service which value to predict. Developers don't have to write code. On the back end, this updated version also supports optimization of many new algorithms. In addition, Microsoft has not yet built a visual interface with Azure ML, and developers can build, train, and deploy machine learning models without touching the keyboard.
Although this is very similar to the existing Azure ML Studio, Microsoft still said that the new version combines the best features of Azure ML Studio with Azure machine learning services, extending the possibilities of ML Studio.
Public preview of Azure blockchain services
This service provides the creation, management, and governance of a specific blockchain network. It is reported that Microsoft announced the launch of Azure Blockchain Workbench in 2018, providing developers with a way to model blockchain applications on pre-configured Azure-enabled networks.
Microsoft said that the first company to use it was JP Morgan Chase and its Quorum ledger based on Ethereum.
In addition to the above items, Microsoft also released some small service update information, but more suspense or choose to stay at the day's conference.
Smart cloud business becomes Microsoft's "new trump card"
Microsoft said that more than 1.3 million developers are using Azure awareness services, and 3,000 new robots are creating features every week. On the scale of digital agents, Microsoft has nearly 400,000 in total.
In the latest release of the 2019 Q1 earnings report, from the perspective of business revenue, cloud computing and gaming business have become the biggest contributors to Microsoft's business revenue growth. Led by the Smart Cloud business unit, the company has achieved healthy revenue and earnings growth in recent years, and it continues to drive growth in its first quarter.
This quarter, Azure's revenue soared 73%, and Microsoft's commercial cloud business, including Azure, grew 41% quarter-to-quarter to $9.6 billion.
However, despite the acceleration of Microsoft's cloud business expansion and the intensification of data center construction around the world, China has also passed the long "five anniversary", but neither the global market, the Asian market and even the Chinese market can surpass the current AWS. And there are Google Cloud andAli CloudThe ups and downs of Microsoft’s construction of technical barriers are not large.
Obviously, if Microsoft Corporation regards smart cloud as a "new trump card", it must re-create its own specific advantages in technology and channels, and can maintain a large traditional customer base. Only this is to narrow down with AWS. The best way to bridge the gap and expand the advantage between the “following the soldiers”.