개발로 하는 개발

Softmax Another popular classifier (like SVM) Generalized version of binary Logistic Regression classifier - Softmax function : - Loss function : cross-entropy function - Numerical Stability : exponential -> very large number -> normalize values - SVM vs Softmax Same score function Wx = b, different loss function. Softmax : Calculate probabilities for each classes. Easier to interpret. Softmax.p..

KNN - space inefficient : have to remember all the data in the training set - classifying is expensive : must calculate all the distances to all of the training set -> Use SVM SVM Linear Classification - Score function, Loss function 사용 : minimize the loss function with respect to the parameters of the score function. CIFAR-10 we have a training set of N = 50,000 images, each with D = 32 x 32 x ..