Under Recovery, Please wait.
AI/ML : The Essentials of Machine Learning : 3-0-0 Course
General Drive Link : AI20001 Spring 2026 by ASR
This includes material from Autumn 2025 & Spring 2026 Semester
20 M - Teachers Assessment
15 M divided into 2 class tests (CT1 on 16th Feb 2025)
5 M for attendance roughly.
30 M - MidSems
50 M - EndSems
The old mark distribution is used for parity with students of previous semester apparently.
Course Instructors for this sem :
Prof Mahesh Mohan M R - Classification, Supervised (Module 1 & 2)
Prof Srinivas Setty Kota - Unsupervised Learning, Neural Networks (Module 3 & 4) - ksreddy@ai.iitkgp.ac.in
---
AI20001 Autumn 2025 Syllabus - Detailed (Dossier)
Overview :
Supervised (Regression, Classifiers, Extensions).
Optimization (Grad Descent)
Unsupervised (Dim Red, Clustering)
Neural Networks (Internals, CNN, RNN & beyond)
Till Ensemble Learning in my Course Outline
Till Ensemble Learning in my Course Outline too.
Our Section only had MCQ-1 (20Q) + 2 Essay type PSets - This is what I have solved
Got a collection of MCQs (300+) from Chem - with answer key
Got MidSem prep q's from Geo - No answer/solns yet.
Tutorials + Student Solution (mine-ASR). Can contain mistakes. SEE THIS
So my madlad friend decides to write a 500+ page book on ML after this course, and goddamn is it good. Since it is inspired by the course, it is one of the best materials for the course, though it is very in-depth. If you get time, I definitely recommend this. Link to the book in my drive. I have also linked to his github repo for this book.
(My list is way smaller, but may miss some crucial points mentioned by the above book)
I have attached Musaib's list of resources here, made during this course, that might be helpful for you too
The Best Resource Overall in my opinion - The Elements of Statistical Learning by Trevor Hastie, Robert Tibshirani, Jerome Friedman. Genuinely has everything. Use the course outline I made to find necessary topics, or use AI to find which chapters to read.
Ideally each concept should take an hour, but some might be longer than others.
Finishing any one bullet point should be sufficient for that topic unless said otherwise. Each bullet point corresponds to a major section in the syllabus I compiled
Pre-Requisite Maths (Linear Algebra, Vector/MV Calc, PnS)
Maths for ML Book (Deisenroth, Faisal, Ong) - Chapters 2 to 6. Sufficient but a lot.
Free Code Camp - Maths behind AI - Chapters 4 to 6. Not the best in my opinion.
Intro to ML
All 3 together:
Geek for Geeks, very high overview - Only for an idea about ML - Not for Loss & Hyper-parameters
Data Camp on Loss fns - Pretty good, might be a bit too deep
Train in Data Blog on Hyperparameters - Some concepts are for a later time, no probs if you don't understand.
Prof Mohan's Slides - Honestly not bad at all for an intro.
Optimization & Gradient Descent
IITM Maths Slides on Grad Descent - Sufficient, a lot of maths. Only do till Adagrad/Adam
Abhay Singh Blog - Overview of Gradient Descent, good for an idea but not more.
Ruder's Paper on Grad Descent - Sufficient.
Linear Regression
All 3 together
Google Crash Course on Linear Regression. Good
Explained.ai by Terrence Par on Regularization - Extremely In-depth for Regularization. Scroll down the page to the links for the full overview of Regularization.
(Just memorize the ridge regression formula for Matrices tho, not available here.)
Geek by Geeks on LWR - Locally Weighted regression. A bit complex.
Logistic Regression
Google Crash Course - Not sufficient imo, but good for an Idea
IBM Course - Should be sufficient.
Linear Classification
All 2 together
Cornell on Perceptron - Complex, watch the video there maybe.
MIT on SVM - A bit complex, i.e. might take time, but worth it. Lagrangian & Dual Spaces MIGHT not be needed.
KNN Classifiers - Honestly couldn't find a good resource for this, the bottom ones are AI suggested but I'm not happy with any of them. My reco is use the slides and research. Definitely use my course outline or slides to make sure you didn't miss anything because the below is not sufficient
Not sufficient but all 3:
Example based idea of KNN - Voronoi Diagrams, Decision Boundaries, and Computational Complexity
Dist Metrics for KNN by Kdnuggets - Distance Metrics
RPI on CNN (Advanced) - Use only if you understoof previous topics. CNN (Condensed Nearest Neighbor), Storage Reduction, and Minimum Consistent Sets
Naive Bayes
The slides should be sufficient.
Applied AI Course on Naive Bayes - Sufficient if you also do the problems
Decision Trees & Forests
Google Course on Decision Tree & Forests - Good & Should be sufficient.
Bias Variance Tradeoff & Ensemble Methods
Again, the slides should be sufficient.
UT Dallas on BVT & Ensemble - Pretty good, sufficient.
Formula Sheet by Musaib Bin Bashir. Very sufficient, but can be advanced.
https://youtube.com/playlist?list=PLcz-75fILqjCWKwGdB8JHnz4VRmEwZz6-
These are my EML Lectures (covers about 50% right now), i.e. made by me, ASR.
It has Mistakes, and please do questions after these. Definitely read the description in case I had to make changes/give suggestions.
Will try to upload rest by 25th Night, but no promises.
Collection of many ML books (pirated)
My recommendation
The Elements of Statistical Learning by Trevor Hastie, Robert Tibshirani, Jerome Friedman. Genuinely has everything. Use the course outline I made to find necessary topics, or use AI to find which chapters to read.
Course Recommendations
Introduction to Machine Learning, Ethem Alpaydin, MIT Press, 4th Edition, 2020 - Available on Matlabyar, a Persian resource sharing website.
Honestly surprised it is still there given the Iranian government murders (IIRC 30,000 in a week) & blackout, but I guess that's what happens if every country does nothing about it.
Pattern Recognition for Machine Learning, Christopher M. Bishop, Springer, 2006
Neural Networks and Deep Learning: A Textbook, Charu C Aggarwal, Springer, 2018
Deep Learning, Aaron Courville, Ian Goodfellow, Yoshua Bengio, MIT Press, 2016.
So, this is getting the PnS treatment. What does that mean? Guides to study this, solutions & materials made by me.
PS: Many of the materials here (PYQ Soln, List of Resources 2, Formula Sheet) were provided by Musaib Bin Bashir, go show him some love too.
Go to Syllabus for this information.
The Dossier/Syllabus provided by institute isn't comprehensive enough to what was thought though. If you have time, go through your prof's slides to know what to learn. If you don't have time, see the outline I have made (available under Syllabus).
Portions taught by midsems were different last year, so don't use that as a guide.
Finishing one of these paths completely is enough to do well in MidSems, in my opinion. I obviously don't know the paper, guessing based on Prob Sets.
Solving the Problem Set or seeing my Solutions is mandatory - Available here
This is a topic wise list of reading resources, curated by me (Actually curated by me spending hours, not by a LLM). It's predominantly blogs/websites/books, and in my opinion, gives the most comprehensive look, barring reference books. Will update if needed.
Available by scrolling down, or clicking here.
I have heard some good things from some friends about these, and that they are sufficient for problem set. Dunno myself how good they are.
Available by scrolling down, or clicking here.
No reviews on this yet.
Available by scrolling down, or clicking here.
Enough for me, but not that explanatory to most people. Use if you prefer googling/revising.
Available by scrolling down, or clicking here.
https://youtube.com/playlist?list=PLcz-75fILqjCWKwGdB8JHnz4VRmEwZz6-
These are my EML Lectures (covers about 50% right now), i.e. made by me, ASR.
It has Mistakes, and please do questions after these. Definitely read the description in case I had to make changes/give suggestions.
Will try to upload rest by 25th Night, but no promises.
Most comprehensive, and if you willing to spend time/want to learn the maths, use this.
Available by scrolling down, or clicking here.