GPU Visionaries: Deep Learning Experts Make MIT’s ‘Innovators Under 35’ List

by Brad Nemire

They’re young. They’re visionary. And they’re using GPUs to change the world.

MIT Technology Review revealed this week its annual “35 Innovators Under 35,” which lists young technologists using today’s emerging technologies to transform tomorrow’s world.

Two of the five winners in the Visionary category are harnessing the computing power of NVIDIA GPUs to drive their artificial intelligence applications. (The other categories are Entrepreneurs, Humanitarians, Inventors and Pioneers.)

Advancing Google Brain

Ilya Sutskever, 29, left high school early to begin university, and ultimately became the protégé of the godfather of deep learning, Geoffrey Hinton. They invented Supervision – a deep neural network that leapfrogged previous systems and was used to win the 2012 ImageNet Visual Recognition Challenge.

Sutskever is now a key member of the Google Brain research team, focused on developing deep learning algorithms that could mimic human brains and vision by training artificial neural networks to recognize objects – similar to how the human brain does.

“When you look at something, you know what it is in a fraction of a second,” Sutskever says. “And yet our neurons operate extremely slowly. That means your brain must only need a modest number of parallel computations. An artificial neural network is nothing but a fairly small number of very parallel, simple computations.”

In a talk last month, Sutskever mentioned there have been 47 deep learning-related product launches in the last two years by Google, such as photo search, Android speech, StreetView and ads placement.

Multilingual Speech Recognition

The other visionary is Adam Coates, 33, who’s the director of Baidu’s Silicon Valley Artificial Intelligence Lab. Its mission is to create technology that will have an impact on at least 100 million people. Coates was one of the early proponents of applying powerful high-performance computing techniques to deep learning. Today, deep learning is the foundation of their research.

“I believe we can use deep learning and HPC to teach computers, devices and appliances to interact with people far better than they do today.” — Adam Coates, AI researcher at Baidu 

His lab, which focuses on speech recognition, recently presented a new model for handling voice queries in Mandarin. Deep Speech, which runs on a powerful GPU-based system, was publicly announced late last year – initially focused on English speech recognition. Today, it has “multi-lingual” capabilities and boasts an accuracy rate of 94 percent.

The complete list of honorees is on and will be featured in the September/October issue print magazine that hits newsstands worldwide on Sept. 1.

For more on how GPUs are fueling advances in deep learning, check out the keynote address by Andrew Ng, chief scientist at Baidu, at our GPU Technology Conference in March.