[DL개념] MLP (multi layer perceptron)

=>입력받은 신호를 처리하고(weighted sum) activation function을 통과시켜 다음 뉴런으로 전달되는 모델

 

solving XOR with MLP 

=> let's assign W and b randomly and see if it works

 

How can we train weight and bias? => Backpropagation with computation graph 

=> we can easily compute any partial derivaitives by chain rule :))

 

 

 

추천영상 : https://www.youtube.com/watch?v=oOQCrm4Vemo&list=PLSAJwo7mw8jn8iaXwT4MqLbZnS-LJwnBd&index=7