NVIDIA CEO Jensen Huang Kicks Off Berkeley AI Event

Jensen-Huang-BAIR-Berkeley.jpg

Highlighting the growing ties between NVIDIA and researchers at elite universities, NVIDIA founder and CEO Jensen Huang spoke to a gathering of AI researchers at the University of California, Berkeley, Wednesday.

The talk kicked off BAIR NVIDIA AI Day at UC Berkeley’s Memorial Stadium, which brought together researchers from the university and the company for talks and demos.

“You here at UC Berkeley are at the intersection of artificial intelligence, computer science and autonomous machines,” Jensen, clad in his trademark black leather jacket, told more than 400 students, faculty and researchers. “Really, it can’t get any better than that.”

The event is the latest example of the close ties between NVIDIA and elite researchers who are using deep learning to advance robotics, autonomous vehicles and computer vision. (See “Deep Learning Pioneers Boost AI Research at NVIDIA AI Labs Around the World.”)

From 3D Graphics to AI

In his talk, Jensen explained how NVIDIA found its way to the center of the AI revolution that has upended computing.

NVIDIA, Jensen explained, began as a bet that on demand for 3D graphics and gaming. To whet the appetite for more sophisticated 3D experiences, NVIDIA invented the programmable shader. And to unlock the computing power of the shaders built into every GPU, NVIDIA invented CUDA.


NVIDIA founder and CEO Jensen Huang spoke to a gathering of AI researchers at the University of California, Berkeley, Wednesday.

CUDA, in turn, gave researchers the tool they needed to unleash the parallel computing power of GPUs, turning the vast quantities of data generated by the internet, and a new generation of neural network models, into the deep learning technology that powers services now relied on by hundreds of millions of people every day.

And — thanks to a new generation of researchers armed with AI computing platforms from NVIDIA — more is coming.

Work that used to take a decade to commercialize now finds it way to market in six months, explained Trevor Darrell, a professor in UC Berkeley’s Department of Electrical Engineering and Computer Science.

“The time between research innovation and the impact on the marketplace has been reduced by an order of magnitude,” Darrell said. “We’re seeing industry partners more excited to support fundamental research.”

Self-Driving Cars, Robots that Learn, and More

The talks from UC Berkeley and NVIDIA researchers highlighted the close ties between researchers in academia and at NVIDIA.

One of day’s highlights: UC Berkeley Assistant Professor Sergey Levine, who explained how he and his team of researchers are developing new deep learning techniques that are giving robots the ability to observe an action, imagine how they can duplicate that action, and then judge how their own actions compare to the mental model they’ve created.

Complementing Levine’s talk, Bryan Catanzaro, NVIDIA’s vice president of applied deep learning research, spoke about how his team is putting deep learning techniques pioneered in academia to work in areas at NVIDIA as diverse as semiconductor design and virtual world design.

“NVIDIA uses AI in every aspect of its work,” said Catanzaro, who earned his Ph.D. at UC Berkeley. “AI is proliferating in new and unexpected places.”

Other speakers included Darrell, who spoke on using deep learning to help machines reason and explore; UC Berkeley Professor Pieter Abbeel, who talked about deep reinforcement learning and meta learning; and NVIDIA’s Larry Jackel, a veteran of Bell Labs who explained how NVIDIA’s work on self-driving vehicles builds on decades of neural network research.

I Am AI, Literally

After the talks, students and researchers gathered to network over wine and appetizers, gawk at demos of how deep learning is being put to work, and talk about what they’d heard.

Computer science student Humza Iqbal says he thinks deep learning will continue to surprise the wider world, particularly in computer security.

Guy Isely, a Ph.D. student in neuroscience, said he thinks people will be surprised by how common robots could become in five to 10 years.

Others, such as Esmond Ai, a student at Berkeley’s Haas school of business — just across the street from the event’s venue — see deep learning opening up vast new opportunities for entrepreneurs.

Ai said he was inspired by an entrepreneurial tale Jensen shared with students — and the “I am AI” t-shirts worn by NVIDIANs at the event.

“‘I am AI’ makes a lot of sense to me,” Ai said with a grin. “I guess you could say I saw those shirts and was drawn in.”