Facebook launched a new artificial intelligence lab in September, and now we know the person who will run it — New York University deep learning expert Yann LeCun. He has been an instrumental figure in the space, doing some seminal handwriting-analysis work in the 1990s, and now focuses generally on applying deep learning techniques to the field of computer vision and image recognition.
As I explained recently in a post laying out the landscape of deep learning projects, there’s a battle brewing among companies such as Google, Facebook, Microsoft and Yahoo to bring in the best researchers. The better they can teach machines to understand what’s in the images users are uploading and the sentences they’re typing, the more intuitive and sticky they can make their services. We’ll be talking more about the business value, and overall promise, of deep learning at our Structure Data conference in March.
Facebook’s hire today is particularly bold, if only because it seems like a direct shot across the bow at Google. Last year, Google bought a deep learning startup created by University of Toronto professor Geoffrey Hinton and brought him on board part-time as a Distinguised Researcher. Now, Facebook has hired someone with comparable name-recognition among deep learning researchers to lead its team.
Like many research fields, though, the deep learning and computer vision fields are pretty interconnected. One of Facebook’s first hires for its AI lab was Marc’Aurelio Ranzato — a former Google employee and member of Hinton’s lab in Toronto who earned his Ph.D. from NYU under LeCun.
No less than Facebook CEO Mark Zuckerberg will formally announce LeCun’s hiring — as well as a research partnership between Facebook and LeCun’s Center for Data Science at NYU (where he will still teach) — at the NIPS Deep Learning Workshop in Lake Tahoe on Monday.