Intel has big AI plans – but it’s keeping some quiet…
“Unfortunately I can’t share too many details with you today, they won’t let me,” said Naveen G. Rao, corporate vice president and general manager of the Artificial Intelligence Products Group at Intel Corporation .
He was speaking today at the opening of Intel’s AI developer conference (AIDC2018) in San Francisco.
What he can’t tell you about is a discrete accelerator Intel are designing for their deep learning inference engine, but this secrecy was the anomaly in a rapid paced hour and half keynote speech.
“AI is really a mix of techniques requiring diverse computing platforms and different kinds of tools,” emphasised Rao, as Intel brought a range of people from a diverse array of backgrounds on stage to drive home the point – and the potential.
From Farms to Pharma
First up was Jason Knight, head of software at Intel, he began by pointing to how “Researchers in Penn state worked with farmers in Tanzania,” to help “identified disease in plants to help improve the lives of farmers there.”
Community and working together was a recurrent theme throughout the presentation. One such partnership is with Dr Rachael Callcut from the University of California. “When trauma patients come in they are at the worst moments of their lives, for many of them their life is truly on the line.”
“Medical data is incredibly complicated,” she points out and so it is a perfect example to use AI. She had found her self “particularly frustrated by the traditional statistical techniques that were available.”
They identified a space where AI and ML could work but “lacked the hardware and software” so they entered into partnership with Intel over a year ago.
Progress has been made, but the next step is moving from making workflow efficiency tools for radiologists to “tools that actually make it to the bedsides of patients,” helping to bring about real outcomes.
Back in steps Naveen G. Rao to reinforce that; “software really is the key to making all these different things happen.”
It was a bit of a whirlwind presentation, but slickly produced with moving graphic slides and intro and out through music for each additional guest on stage.
It really showcased the different aspects of AI and ML in use today.
A big hitter from the film industry was brought out James Jacobs. Jacobs Co Founded Ziva and was recognised by the academy for his work on Avatar and the Hobbit, mainly for his work on tissue physically–based character simulation framework
Using this technique and using AI and ML he works on bringing to life “all the different biomechanical effects that you expect to see, like skin sliding over the underlying structures.”
The animal models look real and you can see the skin slide over the defined muscles underneath. As AR and VR become more common this authentic look is what will sell the products of the future, he emphasised.
The keynote was over, much more was to come.