Nvidia Unveils new AI Architecture at GTC 2016

Processing tech giant Intel and graphics hardware designer Nvidia are taking bold steps to tie AI Architecture into their products and businesses. At the same time, both of these companies are working to fill the holes in their hardware and software capabilities with AI. 

In the past, Nvidia and Advanced Micro Devices (AMD) have lead the charge in the quest to ramp up graphic processing power for gaming systems. Interestingly, the gaming world’s insatiable desire for more life-like gameplay directly led to the development of more powerful graphics processors. If companies can lay out effective AI architecture that improves the consumer experience, will consumers then drive developers to improve the new technology?

“Not only is the DGX SaturnV the most energy-efficient supercomputer ever built, it is available to use in your home.”

What Could AI do for us?

One could argue that the cell phones we carry in our pockets are more powerful than the supercomputer that helped put a man on the Moon. In the past, the emphasis for development had been focused on computing power. The power to compute more created lag time. Now, the emphasis has turned towards the use of AI Architecture to make computing speed more power efficient.

Increasing speed for better power efficiency enables the use of AI systems for everyday tasks. For example, take our modern day appliances and conveniences such as dishwashers, washer/dryers, home energy consumption, and maintenance. The guesswork and waste involved in watering the lawn could be totally removed by an AI platform.

Artificial Intelligence would make arbitrary power cycles a thing of the past. With an AI controlled lawn management system, AI architecture would enable the system to monitor the water level of the grass for over-watering, drought conditions, and soil levels. This would be a revolutionary development, especially in the case of drought-ridden regions where water waste is basically sacrilege.

AI would reduce our carbon footprint by eliminating wasteful practices, not just by watering the grass, but also by improving energy efficiency in our homes.

But before we go too far, AI tech is not quite practical for every home. That is a vision for the future. AI architecture would first be implemented into higher level systems such as public planning or the coordination of air and ground traffic. After these systems become a successful proving ground we might begin to see AI help manage our personal lives.

Developers Interested in AI Architecture; Nvidia Takes a Leap Forward

In late October, at the Graphics Technology Conference (GTC) in Washington D.C., Karl Freund, Moor Insights & Strategy analyst for machine learning reported that “AI technology was the focus of every keynote address.” Furthermore, Mr. Freund wrote, “companies such as Google, Nvidia, Apple, and AMD were interested in the unique aspects of Voice Recognition, Search Advance, and Image Recognition.” Tech giants clamoring to harness the potential of AI technology bodes well for the interested consumer.

Nvidia stole the show, however, when they unveiled the DGX SaturnV, an ultra-efficient in-house supercomputer that now ranks among the World’s 500 most powerful computers. Boasting 63,488GB of RAM and 60,512 Intel Xeon E5-2698v4 CPU cores, the system’s AI architecture is built around Tesla P100 GPUs and several Nvidia 125 DGX-1s, which are regarded as their own AI deep learning supercomputers.

Not only is the DGX SaturnV the most energy-efficient supercomputer ever built, it is available to use in your home. Typically supercomputers are reserved for government or research center use. However, Nvidia will most likely market their supercomputer towards the private sector. Software companies and AI research groups alike are lining up to deal with Nvidia.

Intel’s AI Development

Intel has acquired Nervana Systems (Deep learning AI startup based in San Diego and Silicon Valley) according to Karl Freund. Intel should be unveiling their new platform in 2017. Hopefully, this will serve as an answer to Intel’s AI architecture inadequacies described by Mr. Freund when he mentioned that, “Intel has 1 or more of every architecture flavor except GPUs.”

Intel, in its forward thinking ways, has already expressed interest in creating an AI Education Program. Furthermore, they claim to have the “Best GPU” by the year 2020.

When Will This Technology Affect our Lives?

Nvidia and Intel have made noble strides towards improving their AI architecture, which, in time, will translate into practical uses for the average consumer. Who wouldn’t want a neural network managing the lawn care and security of their home? Consider even an AI system capable of allocating and distributing power generated by rooftop solar.

Let us at Edgy Labs know how you feel about these new advancements in the AI field. Will it help you or hurt you? Your job? Do you think we’ll see revolutionary AI applications in our lifetime?