Get an idea of the learning framework Probably approximately correct (PAC) learning Understand why machine can learn and with what constrains Hoeffding's inequality No free lunch theorem Vapnik–Chervonenkis dimension and Shatter Set Arm some mathematic knowledge with you, in order to understand and design algorithms and optimize the cost functions Some Linear Algebra Normed vector space Positive-definite matrix Eigenvalues and eigenvectors Determinant of square matrix Some Mathematical Analysis Derivative and Differentiation rules Taylor series Some Mathematical statistics i.i.d. Standard deviation and Expected value Probability density function Receiver operating characteristic (ROC) Skewness Some Mathematical Optimization L1 and L2 Regularization Lagrange multiplier Quadratic programming
Comments !