“A bit about platforms”: The future of AIaaS
Every 10 years there is a new platform (in the last issue on this topic - our post about BaaS ), which changes the sphere of computer technology. Already today, the largest players on the market offer the CaaS (cognition as a service) service, or AIaaS.
According to an IBM survey, 55% of respondents working with AI prefer it. In addition, the first users of the services note that the introduction of artificial intelligence in cloud platforms is a necessary element of their business strategy.
/ Flickr / jd hancock / cc
')
Today, a number of advantages that companies get when switching to cloud platforms can safely add the ability to work with artificial intelligence services. For example, in Amazon cloud services, you can train AI models on the data used to predict customer demand and behavior in certain situations.
The development of this direction can be judged by personnel decisions of market leaders. For example, Fei Fei Li, director of artificial intelligence at Stanford University, joined the Google cloud computing team in November. The team’s task is not to conduct research, but to introduce artificial intelligence into Google’s products so that customers can use the software, for example, to forecast sales.
But not only corporations are engaged in these promising technologies. For example, StackStorm presented its solution for integrating the development and operation of software for IaaS. StackStorm CEO Evan Powell explains that the largest cloud service operators are about 10-100 times more productive than any average cloud provider due to internal automation services.
StackStorm will allow automation to be expanded to a wider market, and thanks to integrated machine learning, automation technology will respond faster to growing infrastructure needs. By the way, the cofounder of the startup is our compatriot, Dmitry Zimin, who previously worked at VMware.
There are other companies: some, like DataRobot, are developing as startups, offering software for creating predictive models based on machine learning and cloud computing. And the 2-year-old Nervana Solutions in August was bought by Intel. The startup also offers a cloud platform for deep learning, which is designed to create solutions in different areas of business and research.
If we talk about existing plans in this area, then in 2017, Japan plans to build a supercomputer with a capacity of 130 petaflops. The ABCI project (AI Bridging Cloud Infrastructure) is aimed, among other things, at deep learning of neural networks. Access for companies will be free, which should contribute to active development.
Many experts agree that cloud providers and IT companies will continue to move in the direction of AI-oriented services and projects. But how exactly this system will be built is still not entirely clear. Jesus Rodriguez, an investor, start-up consultant and founder of Tellago software development company, predicts several possible developments. For example, he is confident that cloud platforms will support the most popular deep learning frameworks, including TensorFlow, Caffe, Theano, Torch, and others, just as they support the implementation of web applications or background processes today.
Alphabet Chairman of the Board of Directors of Alphabet company Eric Schmidt (Eric Schmidt) shared his vision of the development of the industry at the “cloud” conference of Google in San Francisco: he is sure that in the next five years, most of the business decisions will be based on machine learning on a cloud platform . AI as a service can provide qualitatively new possibilities for processing big data and, undoubtedly, will bring many more surprises, as it was with the development of a mobile platform.
Services of this type will probably bring the main income to large corporations (see paragraph “The New Cloud”). And that is why they are actively involved in the race to develop solutions that will enable the use of the capabilities of AI systems. There are more and more startups in this area, so cloud services that provide sufficient capacity will be increasingly in demand.
A development boom will be possible, especially due to substantial savings. Agree with this and IDC, an analytical company in the field of IT. (See “Cognitive Computing Capabilities” at the source ). According to their forecasts, cloud infrastructure is democratizing the introduction of AI systems.
What other topics we consider in the blog IT-GRAD on Habré:
According to an IBM survey, 55% of respondents working with AI prefer it. In addition, the first users of the services note that the introduction of artificial intelligence in cloud platforms is a necessary element of their business strategy.
/ Flickr / jd hancock / cc
')
Situation in the market of cloud-intellectual services
Today, a number of advantages that companies get when switching to cloud platforms can safely add the ability to work with artificial intelligence services. For example, in Amazon cloud services, you can train AI models on the data used to predict customer demand and behavior in certain situations.
The development of this direction can be judged by personnel decisions of market leaders. For example, Fei Fei Li, director of artificial intelligence at Stanford University, joined the Google cloud computing team in November. The team’s task is not to conduct research, but to introduce artificial intelligence into Google’s products so that customers can use the software, for example, to forecast sales.
But not only corporations are engaged in these promising technologies. For example, StackStorm presented its solution for integrating the development and operation of software for IaaS. StackStorm CEO Evan Powell explains that the largest cloud service operators are about 10-100 times more productive than any average cloud provider due to internal automation services.
StackStorm will allow automation to be expanded to a wider market, and thanks to integrated machine learning, automation technology will respond faster to growing infrastructure needs. By the way, the cofounder of the startup is our compatriot, Dmitry Zimin, who previously worked at VMware.
There are other companies: some, like DataRobot, are developing as startups, offering software for creating predictive models based on machine learning and cloud computing. And the 2-year-old Nervana Solutions in August was bought by Intel. The startup also offers a cloud platform for deep learning, which is designed to create solutions in different areas of business and research.
A look into the future
If we talk about existing plans in this area, then in 2017, Japan plans to build a supercomputer with a capacity of 130 petaflops. The ABCI project (AI Bridging Cloud Infrastructure) is aimed, among other things, at deep learning of neural networks. Access for companies will be free, which should contribute to active development.
Many experts agree that cloud providers and IT companies will continue to move in the direction of AI-oriented services and projects. But how exactly this system will be built is still not entirely clear. Jesus Rodriguez, an investor, start-up consultant and founder of Tellago software development company, predicts several possible developments. For example, he is confident that cloud platforms will support the most popular deep learning frameworks, including TensorFlow, Caffe, Theano, Torch, and others, just as they support the implementation of web applications or background processes today.
Alphabet Chairman of the Board of Directors of Alphabet company Eric Schmidt (Eric Schmidt) shared his vision of the development of the industry at the “cloud” conference of Google in San Francisco: he is sure that in the next five years, most of the business decisions will be based on machine learning on a cloud platform . AI as a service can provide qualitatively new possibilities for processing big data and, undoubtedly, will bring many more surprises, as it was with the development of a mobile platform.
Services of this type will probably bring the main income to large corporations (see paragraph “The New Cloud”). And that is why they are actively involved in the race to develop solutions that will enable the use of the capabilities of AI systems. There are more and more startups in this area, so cloud services that provide sufficient capacity will be increasingly in demand.
A development boom will be possible, especially due to substantial savings. Agree with this and IDC, an analytical company in the field of IT. (See “Cognitive Computing Capabilities” at the source ). According to their forecasts, cloud infrastructure is democratizing the introduction of AI systems.
What other topics we consider in the blog IT-GRAD on Habré:
- NASA's Hyperwall System Now Uses Storage System vFlash
- IaaS Digest: Introducing Virtual Infrastructure
- “What is the benefit?”: Feasibility study of using IaaS
- SAP HANA in the VMware Cloud: Calculate Required Resources
- "Planning Issues": Cloud ERP Hosting
- How cloud technologies change the world of sports
- IaaS and GaaS: The Cloudy Future of Video Games
Source: https://habr.com/ru/post/316286/
All Articles