LeNet-5 used a convolutional neural network to recognize handwritten digits. The discussion connects that practical computer-vision task to automated check processing and the development of modern neural networks.
256 words · auto-generated from the episode video
1:52:01>> Last one in 1998, development of Lynette 5 conv uh convolutional neural networks which we hear a lot about these days. Yes, this is Yan Lun and his team at Bell Labs. They developed layet 5 which is the convolutional neural network architecture that revolutionized computer vision by automating handwritten digit recognition. This is the first real like um application industrial application of neural networks and convolutional neural networks specifically for the problem of computer vision because like when you write your checks and you know you like scan it and and and then it goes into your bank account. You don't have to
1:52:42like tell the teller and then the teller punches in something, right? All of that started with um this emnest data set of the handwritten digits and how they get classified into 0 1 2 3 4 5 and six. Yan Lun back in back in those days um he did not have the sense of fashion that he does. >> I was going to say is that a is that a young lun with the ponytail there? >> Yeah, >> that is a throwback photo. If you're listening, you might want to take a look at this cuz you see him nowadays. Very fashionable. very fashionable like got his glasses are not like whatever those glasses are right >> definitely made that major upgrade another one of the uh godfathers of the modern AI world we live in today many of
1:53:25us will remember this as we get into the