ES 654 IIT Gandhinagar

Home Lectures Assessment FAQ and grading Project ideas

Lectures

Every class of the January–April 2022 offering, numbered as in class, with its slides, recordings, notes and notebooks.

NotesSlidesPDF handoutFull recordingNotebookInteractive labReading

Introduction and metrics1–4Decision trees and ensembles5–12Linear regression and optimisation13–25Bayesian learning and logistic regression26–31Neural networks32–35Naive Bayes, SVMs and unsupervised learning36–39

All recordings (Google Drive)Class spreadsheets (Google Drive)

Introduction and metrics

Lectures 1–4

Lecture 01	Introduction	Slides Recording
Lecture 02	Introduction (continued)	Slides Recording
Lecture 03	Conventions, metrics, classification and regression Lectures 3 and 4 were a double class.	Slides Recording
Lecture 04	Conventions, metrics, classification and regression (continued) Second half of the double class; slides and recording as in lecture 3.	Slides Recording

Decision trees and ensembles

Lectures 5–12

Lecture 05	Decision trees I	Slides Recording
Lecture 06	Decision trees II	Slides Recording
Lecture 07	Quiz 1 Held in class.	Quiz paper
Lecture 08	Decision trees II (continued)	Slides: Decision trees II, bias and variance Recording
Lecture 09	Cross-validation	Slides: Decision trees II, bias and variance Recording
Lecture 10	Ensemble learning I	Slides Recording
Lecture 11	Ensemble learning II	Slides Recording
Lecture 12	Ensemble learning II (continued)	Slides Recording

Linear regression and optimisation

Lectures 13–25

Lecture 13	Maths for ML	Slides Recording
Lecture 14	Linear regression	Slides Recording
Lecture 15	Geometric interpretation of least squares; contour plots	Slides: Geometric interpretation of least squaresSlides: Contour plots Recording
Lecture 16	Linear regression II	Slides Recording
Lecture 17	Convexity	Slides Recording
Lecture 18	Gradient descent	Slides Recording
Lecture 19	Linear regression: time complexity	Slides Recording
Lecture 20	Constrained optimisation	Slides Recording
Lecture 21	Ridge regression	Slides Recording Notebook: Ridge regressionHugging Face demo: Ridge regression
Lecture 22	Ridge regression (continued)	Slides Recording Notebook: Ridge regression IIHugging Face demo: Ridge regression (updated)
Lecture 23	Ridge regression (continued)	Slides Recording
Lecture 24	Lasso regression I	Slides Recording
Lecture 25	Lasso regression II	Slides Recording

Bayesian learning and logistic regression

Lectures 26–31

Lecture 26	Bayesian learning	Slides
Lecture 27	Bayesian learning (continued)	Slides Recording
Lecture 28	Bayesian linear regression	Slides
Lecture 29	Bayesian linear regression (continued)	Slides
Lecture 30	Logistic regression I	Slides
Lecture 31	Logistic regression II	Notes: Logistic regression cost derivation, part 1Notes: Part 2Reading: From Taylor series to gradient descent Video: Taylor series (3Blue1Brown)

Neural networks

Lectures 32–35

Lecture 32	Multilayer perceptrons	Slides
Lecture 33	Multilayer perceptrons (continued)	Slides
Lecture 34	Convolutional neural networks	Slides
Lecture 35	Convolutional neural networks (continued)	Slides

Naive Bayes, SVMs and unsupervised learning

Lectures 36–39

Lecture 36	Naive Bayes	Slides
Lecture 37	Support vector machines I	Slides
Lecture 38	SVM kernels and soft margin	Slides: KernelsSlides: Soft marginReading: Jake VanderPlas, Support vector machines Video: SVM kernelsVideo: SVM soft margin
Lecture 39	Unsupervised learning	Slides Video: Unsupervised learning

Slides are the PDFs used in class; recordings are on Google Drive. Lecture 7 was Quiz 1, held in class.