Author(s): Raghu Gopal
As companies embark on a digital transformation, many technology players have begun to turn to artificial intelligence to deliver the next generation of IT services to their enterprise customers.
Last week at the SC16 supercomputing conference in Salt Lake City, Utah, Nvidia announced a collaboration with Microsoft to improve its upcoming Tesla server graphics processing units (GPUs) with an artificial intelligence framework. Microsoft and Nvidia plan to develop methods of applying the technology to boost their enterprise customers' efficiencies by using graphics cards that support deep learning. Nvidia will optimize its enterprise artificial intelligence framework to run on Tesla GPUs for both the Microsoft Azure cloud and its own servers.
Artificial intelligence has become synonymous with artificial assistants, autonomous cars and chat bots. Microsoft's Cognitive Toolkit, an open-source deep-learning solution that mimics how the human brain processes information, is one of many platforms that trains computers to learn, and it's getting an upgrade. The solution, used for Bing, Cortana, Skype and Xbox, trains and adjusts algorithms to scale according to inputs, making GPUs more efficient with a hybrid cloud platform. Based on input, researchers continuously modify models and tweak parameters.
Nvidia GPUs already work on the Cognitive Toolkit with existing libraries, but the company claims the new framework enhances the performance by seven times. The training of computer models can run for days and requires intense computing horsepower. The power offered by GPUs is used by companies such as Facebook and Google. Microsoft will enable customers to test GPUs as part of its Azure cloud service. The updated Cognitive Toolkit will run faster on Microsoft's Azure N-Series cloud offerings.
The collaboration to promote artificial intelligence in enterprises will allow companies to harness the technology for decision-making. Nvidia and Microsoft plan to continue their partnership to help optimize the Cognitive Toolkit for Nvidia GPUs in Azure and as part of a hybrid cloud platform. The gold rush for artificial intelligence solutions is underway among semiconductor makers. It's little wonder that other players such as Qualcomm and Intel are also investing heavily in this area.