CS231N Spring Lecture1 Lecture Notes

1. Biologists have experimented with finding that the brain cortex responds to simple structures such as horns and edges, and through complex neurons, these simple structures ultimately help organisms to have more complex visual systems. 1970 David Marr's vision processing process follows the principle that after getting the image, it extracts simple geometric elements such as angles, edges, curves, and so on, and then uses more sophisticated information, such as depth information, surface information, and so on, and finally the more abstract expression at higher levels. Deep learning also follows this basic idea, starting from the simplest features, passing through multi-layer functions to achieve complex functions.

2. Image-net competition, 2012 breakthrough changes, alexnet with convolutional neural network greatly improved the accuracy rate, after this method became mainstream, the number of layers, the 2015 Microsoft used more than 100 layers of network. After that, this kind of game doesn't make much sense, because it does have better results with more layers (such as layer 200), but requires hardware such as the GPU, and the recognition rate has surpassed that of humans.

3. Deep learning did not suddenly fire, the 2012 alexnet was essentially the same as the 1998 LeCun letter-recognition paper. The most important reason for the deep learning in recent years is the progress of hardware and the increase of data volume.

CS231N Spring Lecture1 Lecture Notes