Here is the planned syllabus:
Lectures Calendar
Tutorials Calendar
Lectures Calendar
| Subject | |
|---|---|
| 1 | Introduction, Gentle start: online |
| 2 | Probability Review, Uniform Convergence Principle |
| 3 | PAC learning and VC theory |
| 4 | Nonuniform Learnability |
| 5 | Computational Complexity of Learning |
| 6 | SVM (support vector machine): geometric margin, norm regularization, representer theorem, kernel trick, sGD (sub gradient descent) and stochastic sGD for SVM |
| 7 | Kernels, Regularized Loss Minimization, Validation |
| 8 | learning guarantees for SGD, early stopping as regularization, online learning, online vs stochastic optimization |
| 9 | Feature selection |
| 10 | multilayer networks, learning input representations |
| 11 | Probabilistic models, Naive Bayes, generative vs discriminative models, MAP (maximum a posteriori) vs SRM (structural risk minimization) |
| 12 | Advanced topics |
| 13 | Advanced topics |
Tutorials Calendar
| Subject | |
|---|---|
| 0 | Intro |
| 1 | Offline Learning, Hoeffding Bound |
| 2 | VC dimension tutorial |
| 3 | Least Squares, Linear Programming, Perceptron |
| 4 | convex sets and functions, convex problems, GD (gradient descent), smoothness, lipschitz, newton, sub-gradient and sGD (sub gradient descent) |
| 5 | sub-gradients, sub-gradient descent for SVM |
| 6 | Not decided yet |
| 7 | Not decided yet |
| 8 | Not decided yet |
| 9 | Back-propagation |
| 10 | ML (maximum likelihood), MAP (maximum a posteriori), Bayesian estimates |
| 11 | Linear Discriminant Analysis (LDA) and other probabilistic models |
| 12 | Not decided yet |
