Tech & Innovation : Artificial Intelligence
Facebook’s decision this month to open an Artificial Intelligence (AI) research laboratory in Paris is hardly surprising. The Director of its two existing research teams in Menlo Park (California) and New York is a Frenchman, Yann LeCun, who received his engineering degree and PhD in Computer Science in Paris. Now, a New York University professor in Computer Science, he is a pioneer in a particular field of AI called ‘Deep Learning’. This algorithm-based machine learning technique bears the promise of dramatic improvements in getting computers to do perceptual tasks like speech recognition, natural language processing, real-time translation and image recognition.
Paris was chosen for Facebook’s new research team as it offers a rare concentration of talents in Mathematics and Computer Science which are two essential disciplines in Deep Learning. A few months ago when the American company decided to reinforce its AI research team, it had already turned, partly, to some of the French-trained sharpest brains in AI: Léon Bottou who after his PhD in Paris did his pioneering work in the US on machine learning, stochastic optimization (stochastic processes involve probability and randomness used in predicting events), image compression etc ; Nicolas Usunier who was recently a University professor and obtained his PhD in machine learning in France; Gabriel Synnaève a postdoctoral fellow at École Normale Supérieure in Paris who received his PhD in Bayesian modeling (a subset in the field of statistics) applied to AI, from the University of Grenoble.
Yann LeCun himself working as a researcher for Bell Labs, showed how Deep Learning could be used to significantly improve handwriting recognition. Many of the cheques written in the US are now processed by the technology he developed.
AI is going through an exciting period and catching world attention with tech giants in the US competing for the leading minds in the field. The layperson name ‘Deep Learning’ is known in academia as Convolutional neural networks (a neural network is a system of programs and data structures which are close to the operation of the human brain). The technique is strongly associated with the name Yann LeCun. At New York University, he founded the University’s Center for Data Science. He says that other potential applications for AI and Deep Learning include self-driving cars, medical image analysis systems, Bayesian models of Cosmology, high-energy physics, computational models of the visual and motor cortex, machine-learning models of social behavior and economics. Mobile speech recognition applications (Apple, Android, Microsoft) already use Deep Learning.
Sources: newsroom.fb.com spectrum.ieee.org wsj.com
Video Source: https://www.youtube.com/watch?v=AbjVdBKfkO0
Online publishing, June 8th, 2015