Spring 2021
Prof. Biplab Banerjee
Author: Hrushikesh Deshpande
Pre-requisite courses: CS101
Pre-requisite skills: Prior programming experience in Python/ Julia is useful for the assignments. However, this is not a hard requirement as demo sample codes were provided for models based on the algorithms covered. The TAs also held coding sessions regularly aside of the lectures.
Course Content:
-
Introduction to ML
-
Idea of supervised, unsupervised, semi-supervised, reinforcement learning
-
Linear regression
-
Idea of model complexity, generalization, bias-variance trade-off, regularization
-
Cross validation, VC dimension
-
Supervised classification algorithms: K nearest neighbor, LDA, Decision Tree, SVM and kernel methods, Neural Network, Naive Bayes’, Gaussian discriminant analysis, Ensemble methods etc.
-
Probabilistic learning models: Parameter estimation using MLE, MAP, GMM, EM algorithm
-
Unsupervised learning: Clustering and kernel density estimation, K-means, DBSCAN, Parzen window technique etc.
-
Dimensionality reduction using PCA and kernel PCA
-
Intro to reinforcement learning
-
Intro to deep learning and convolutional networks, recurrent networks
Motivation to take up the course: One of the two mandatory courses required for a minor in AI and Data Science.
Quizzes/Midsem/Endsem papers Difficulty: 3/5
Difficulty level of Projects: 2/5
Evaluation Structure:
- Assignments - 20%
- 3 Quizzes - 60%
- Group Project - 20%
Lecture Style and Delivery: The professor uses slides for teaching and adds annotations on those during the lectures, which can get a bit disorganised at times. He covers the material at a brisk pace during the lectures so it is important to be attentive.
Attendance Policy: No DX grade given.
General funda: The instructor covers some extra details in the lectures which are not present in the slides uploaded. Hence, it is better to attend the lectures rather than just relying on the slides. The quizzes though objective in nature, required the students to separately write justifications for credit. Being thorough with the concepts is advised, as the quizzes were focused on the understanding of the topics, rather than pure mathematical rigour.
References:
-
Pattern Recognition and Machine Learning, by Christopher Bishop, Springer 2011
-
The Elements of Statistical Learning: Data Mining, Inference, and Prediction, Second Edition, by Trevor Hastie and Robert Tibshirani (Springer Series in Statistics) 2016
-
Supplementary material available online. E.g. Dive into Deep Learning by Aston Zhang, Zack C. Lipton, Mu Li and Alexander Smola, 2020 (https://d2l.ai)