- Daily Zen
Apple is observing serious measures to lead the impending artificial intelligence boom many sectors are looking up to. It’s no surprise. The premium phone maker has recently shown off aggressive business strategies; exploring into more areas to bloat its fortune. Apple launched its content offering and the highly-praised Apple Card last week to mark new entries and expand its horizon in the entertainment business and financial services respectively. The latest move is to poach one of Google’s top AI researchers, Ian Goodfellow, as the fight for top AI talent between tech giants continues to grow stronger.
Ian Goodfellow (34) is undoubtedly one of the most prominent AI experts out there. He is the father of GANs, an AI approach known as general adversarial networks, and Goodfellow’s works are widely cited in machine learning literature. He has worked as a research scientist at Elon Musk’s OpenAI and served as a senior staff research scientist at Google for over 2 years. Currently, Goodfellow’s LinkedIn profile, first reported by CNBC, indicates he is now working at Apple’s Special Project Group as a director of machine learning.
Goodfellow is not the first Google’s AI talent to join Apple. A year ago, the Phone maker hired Goodfellow’s former boss, John Giannandrea, who was heading Google’s AI division.
While artificial intelligence remains a key to the future, its development is still in the early stage and yet to convincingly move into the mainstream. Hence, tech companies struggle to find AI researchers that can move the technology forward. Google may be recognized as the world’s top AI company, the technology exists in most Apple products; from photography to facial recognition. However, Apple is also developing self-driving cars through its Project Titan autonomous vehicle group. Top AI engineers are required to make the latter a reality but Apple reportedly laid off over 200 employees from the secret division.
Goodfellow’s most prominent work, GANs are made up of two parts: a generator and a discriminator. The generator network is trained on a set of data with plans to replicate it and the discriminator, which receives works from the generator, is designed to distinguish the trained dataset and the replicated version – new output. The discriminator would send the generator back to try again if it’s able to tell the difference between the new output and the trained dataset. In practical, GANs is the algorithm behind AI-generated portraits from real images and the software that generates photorealistic paintings from doodles.
Apple may have no interest in generating fake images. The company simply recognizes Goodfellow’s contribution to AI development and what he could offer in its AI programs. Still a very young talent, Apple can deploy his experience and influence in various future projects which include improving Siri.