Every big company in Silicon Valley is chasing artificial intelligence, the increasingly powerful computer software that can learn from big sets of data almost like a human does.
And Apple just hired one of the big guns of AI research, as the company continues to try to prove that it is not falling behind Google and its other rivals in the hot field.
Ruslan Salakhutdinov announced on Twitter on Monday that he’s joining Apple as a director of AI research. Salakhutdinov will continue to do work at Carnegie Mellon University, where he advises and does research on deep learning, a key AI technique.
His hiring raises the question of whether Apple’s growing AI team will take a more academic approach to AI research going forward. One of the big reasons why people believe that Apple is falling behind in the field is that Apple’s researchers don’t publish their research, because Apple has traditionally kept its methods secret until a new product is launched. That secretive approach may be hampering Apple’s recruiting efforts in the AI field, where many of the top researchers want to see their work published.
By contrast, Google and its AI subsidiary, DeepMind, publish their research all the time. Facebook’s head of AI research, Yann LeCun, still puts his name on papers, including one (Recurrent Orthogonal Networks and Long-Memory Tasks) from June of this year, that are frequently published on Facebook’s research page.
Apple did not immediately return a request for comment about what Salakhutdinov will be working on in his new role, though the company has numerous products with obvious artificial intelligence needs, from its Siri virtual assistant to self-driving cars.
According to Salakhutdinov’s faculty page on the CMU website, he’s interested in deep learning, probabilistic graphical models, and large-scale optimization.
One of his recent talks discusses deep learning model that could be used to, say, write captions for various photos, a similar task to a new photo feature Apple introduced in the latest version of the iPhone software. His most recent paper discusses ways that machines can understand text and other documents.