Short explainer video where Károly Zsolnai-Fehér introduces artificial neural networks and deep learning. He describes neural networks as computer simulations loosely inspired by the brain (inputs→hidden layers→output), explains hierarchical feature learning (edges → parts → objects), and gives examples in image and audio recognition, noting a neural network outperforming humans in a traffic sign recognition contest.
Chapters
I amaiher and this is two minute papers where I explain awesome research in simple words. First of all, I am very happy to see that you liked the series. Also, thanks for sharing it on the social media sites and please keep them coming. This episode is going to be about artificial neural networks. I will quickly explain what the huge deep learning rage is all about. This graph depicts a neural network that we build and simulate on a computer. It is a very crude approximation of the human brain. The leftmost layer denotes inputs, which can be, for instance, the pixels of an input image. The rightmost layer is the output, which can be, for instance, a decision whether the image depicts a horse or not. After we have given many inputs to the neural network in its hidden layers, it will learn to figure out a way to recognize different classes of inputs, such as horses, people, or school buses. What is really surprising is that it's quite faithful to the way the brain does represent objects. On a lower level, it has a very similar edge detector and it also works for audio. Here you can find the difference between the neurons in the hearing system of a cat versus a simulated neural network on the same audio signals. I mean, come on, this is amazing. What is the deep learning part all about? Well, it means that our neural network has multiple hidden layers on top of each other. The first layer for an image consists of edges, and as we go up, a combination of edges gives us object parts, a combination of object parts yield object models, and so on. This kind of hierarchy provides us very powerful capabilities. For instance, in this traffic sign recognition contest, the second place was taken by humans. But what's more interesting is that the first place was not taken by humans. It was taken by a neural network algorithm. Think about that. And if you find these topics interesting, you feel you would like to hear about the newest research discoveries in an understandable way, please become a fellow scholar and hit that subscribe button. And for now, thanks for watching and I'll see you next time.
