EE269 - Signal Processing for Machine Learning

Lecture Slides

Lecture 1: Introduction, signal processing and machine learning systems

Lecture 2: Discrete signals, change of basis

Lecture 3: Discrete Fourier Transform

Lecture 4: Distance based signal classification, nearest neighbor classifier, Hilbert space

Lecture 5: Linear systems, circulant matrices, convolution, eigenvector decomposition

Lecture 6: Bayes classifiers, Bayes risk, signal detection

Lecture 7: Stationary signals, autocorrelation, linear and quadratic discriminant analysis

Lecture 8: Fisher's linear discriminant, simultaneous diagonalization

Lecture 9: Multi-class discriminant, separating hyperplanes, Support Vector Machine (SVM)

Lecture 10: Constrained optimization, convex duality, dual SVM

Lecture 11: Nonlinear features, kernels, kernel machines

Lecture 12: Least-squares regression, autoregressive models

Lecture 13: Reproducing Kernel Hilbert Spaces and functional regularization

Lecture 14: Adaptive filters, Least Mean Squares algorithm

Lecture 15: Neural networks

Lecture 16: Deep learning, convolutional networks and spectrograms

Lecture 16: Convex optimization for neural networks and overparameterized models

Lecture 17: Wavelets, Discrete Wavelet Transform and Short-Time Fourier Transform

Lecture 18: Autoencoders, Robust Principal Component Analysis and nuclear norm

Lecture 19: Nonnegative Matrix Factorization, clustering and deep matrix factorizations

Lecture 20: Dictionary learning and matching pursuit