Yann LeCun has become mainstream during the recent AI revolution in his role as the Chief AI Scientist at Meta, but he’s been around in the AI world for decades.
A video from 1993 has surfaced which shows a young Yann LeCun demonstrating his number-recognition AI system. LeCun and his team had created the world’s first world’s first convolutional network for text recognition which could read hand-written numbers. The video shows LeCun showing off their AI system on old CRT monitors.
The video shows a youthful looking Yann LeCun sitting in front of an old-fashioned computer and a keyboard. LeCun places a piece of paper with hand-written numbers under some sort of a camera, which shows up on the monitor. He then presses a button the keyboard, and the computer recognizes the numbers, and instantly prints them on the monitor.
This might seem trivial today with the advent of OCR software and AI models, but reading numbers was quite a breakthrough 3 decades ago. LeCun received a Diplôme d’Ingénieur from ESIEE Paris in 1983, and had completed his PhD in Computer Science from Université Pierre et Marie Curie (now Sorbonne University) in 1987. His PhD thesis introduced an early form of the back-propagation learning algorithm for neural networks and focused on machine learning.
As a post-doc, LeCun joined the research group of Geoffrey Hinton, who would become one of the godfathers of AI, at the University of Toronto in 1987, which deepened his interests in neural networks. In 1988, he started at AT&T Bell Laboratories, where he developed foundational methods including convolutional neural networks (LeNet), the “Optimal Brain Damage” regularizer, and graph transformer networks for handwriting and character recognition. LeCun headed the Image Processing Research Department at AT&T Labs-Research (1996–2002).
LeCun joined New York University as a professor in 2003, later becoming Silver Professor at the Courant Institute and founding director of NYU’s Center for Data Science. In 2013, he became the Director of AI Research at Facebook and is now the Chief AI Scientist at Meta.
LeCun is recognized as the pioneer of CNNs, which revolutionized image and speech recognition. His check recognition system, which used the same technology as in the video from 1992, was deployed commercially and read over 10% of all checks in the US during the late 1990s and early 2000s. LeCun developed the widely used DjVu image compression technology and the Lush programming language. He was the co-recipient of the 2018 ACM Turing Award, considered the highest honor in computer science, alongside Geoffrey Hinton and Yoshua Bengio, collectively called the “Godfathers of AI”. And this video isn’t only a peek into the early world of AI, but also a reminder of how far things have come in just a few decades.