
Simple introduction.
Perceptron is first ancestor of neural network!
Perceptron showed its problem. We are going to know why perceptron has weakness
We are going to know why nonlinearity is so important.
Activation functions is what gives non linearity to neural network. We are going to talk about activation function.
We want make neural network do tasks instead of us. How can we do that? Loss function is answer.
Target is reducing loss functions. But how can we reduce loss function? Gradient decent is key.
Sigmoid function and tanh function has weakness! You have to know what gradient vanishing and gradient exploding is!
There's better activation functions! i'm going to introduce famous activation functions!
Gradient decent has its own weakness. We are going to know how can we update weights of neural network by using better methods.
We cannot make perfect neural network. When we train neural network, we have to face underfitting or overfitting. We are going to know what is underfitting and overfitting is and how to deal with this problem (weight regularization , dropout)
There's few things you have to know before starting training neural network. We are going to know what it is.
Welcome to my lecture! Nowadays, artificial intelligence is everywhere! You can see artificial intelligence, especially deep learning everywhere. Deep learning can draw beautiful art image , deep learning can make astonishing music, deep learning can even drive! I'm pretty sure that it's okay to say nowadays are era of deep learning. And what if you can also dive into world of deep learning? As a graduate student, i always wanted to summarize what i learned and introduce it to people to help them understand deep learning more easily. Therefore, in this lecture, we are going to learn basic and really important knowledge that you have to know if you are beginner of deep learning. Before starting this lecture, you should know basic linear algebra, basic knowledge about probability and statistics and basic python programming in advance if you want more smooth understanding. After this lecture, you'll be ready to dive into deep learning world that change your computer from just computer to tool for deep learning. I hope you to ask anything you want to ask me about lecture, or if there's something missing or you want to know more, let me know without any hesitation. Again, welcome to my lecture everyone!