Welcome

CS771: Introduction to Machine Learning

Course Details
Number: CS771
Name/Title: Introduction to Machine Learning
Admin Team: wait for next slide
Website: https://www.cse.iitk.ac.in/users/purushot/courses/ml/2023-24-w/
Videos (YouTube): https://www.youtube.com/c/MLConversations
Discussion (Piazza): https://piazza.com/iitk.ac.in/secondsemester2024/cs771
Slides, code, notes (GitHub): https://github.com/purushottamkar/cs771
Admin Team
Email
SNo Admin Name Prog Interests
(@cse.iitk.ac.in)
1 Subhajyoti Saha PhD subhajyoti Machine Learning
2 Jegan Balaji B S MS jeganb Graph mining, Recommender systems
3 Avnish Tripathi MTech avnish ML in healthcare
4 Divyesh Devangkumar Tripathi MTech bittu Computer Vision
5 Hammad Ansari MTech hammad Systems
6 Kartick Verma MTech kartickv NLP and applied ML
7 Kush Shah MTech kushshah Bilevel optimization
8 M N V Ramachandra Ananthapalli MTech mnvrama Full stack development
9 Manik Shekhar MTech manik Virtualization
10 Nihal Thukarama Rao MTech nihal Computer Vision, NLP, GPGPU computing
11 Pranjal Kumar Srivastava MTech pranjalks Machine Learning, Deep Learning
12 Samarth Joshi MTech samarth Algorithms
Admin Team
Email
SNo Admin Name Prog Interests
(@cse.iitk.ac.in)
13 Sanket Kale MTech sanket Machine Learning Theory
14 Shubhashish Moitra MTech smoitra Machine Translation and Model Exfiltration
15 Subhajit Panday MTech subhajitpanday Machine Learning
16 Kakumanu Vamsee Side Channel Attack and Analysis, Image Reconstruction from
Krishna MTech vamsee EM Leakages, Deep Learning - Image Processing & Classification
17 Yash Uttamchandani MTech yashlu Machine Learning
18 Aditya Kankriya MTech adityakankriya Computer Vision, Software Development
19 MitrajSinh Chavda MTech mbchavda Machine Learning, Encryption, Blockchain
20 Yogesh Shrivastava MTech yogesh Cybersecurity
21 Anubhav Dixit PhD anubhavdixit Computer Vision and Deep Learning for Healthcare
22 Optimization techniques in ML, Probabilistic ML, Computational
Debkanta Chakraborty PhD debkanta Genomics, Bioinformatics
23 Shivam Pal PhD pshivam Machine Learning
24 Kuruma Abhinav MS abhinav Linux
Auditors
Please email the instructor [email protected] to get enrolled

Auditors will have access to

Lecture videos, slides, code, notes
Assignment, quiz and exam questions and solutions

We regret our inability to extend the following services to auditors

Submit assignments and receive graded submissions
Appear for quizzes, examinations and receive graded answer scripts
Grading Scheme
20%: Quizzes
30%: Assignments
20%: Mid-sem Exam
30%: End-sem Exam
Quizzes
Two quizzes to be conducted – each worth 10%
Tentatively scheduled on Jan 24, 2024 (Wed), Mar 20, 2024 (Wed)

Only registered students can appear for quizzes (no auditors)

Syllabus for each quiz will cover roughly 3-4 weeks – short quiz

If you are regular with lectures, discussion hours, practice questions,

then you should not have to prepare too much for these quizzes
Assignments – 30%
Three mini projects (weightage TBA)
Replaces the single semester-long project in previous offerings of CS771
To be done in groups of five (5) students each – 2-3 weeks for each project
Start forming your group-of-5 today
Will ask you to submit group details once add-drop is over
Groups can only contain registered students (no auditors)
Create a homepage on CC/CSE home servers
Submission will include code + report
Code should be in Python – start learning Python today
Report should be in LaTeX – start learning LaTeX today
Extra Classes
Two institute holidays clashing with course schedule
26 January 2024: Republic Day
08 March 2024: Maha Shivratri

Two extra classes to be held to compensate

07 February 2024 (Wednesday), 6PM L20
10 April 2024 (Wednesday), 6PM L20

Note that all extra classes are on Wednesdays

Reference Material
No single textbook for the course

List of reference material is up on course website

Python Resources: several available – choose your favourite

www.geeksforgeeks.org/python-programming-language/

LaTeX resources: several available – choose your favourite

www.overleaf.com/learn/latex/Tutorials

Thanks to Amit Chandak and Gourav Takhar for the helpful links!
Course Website
Detailed syllabus for this
course
Course calendar: schedule
for holidays, exams,
quizzes
Course policy: assessment,
course drop, make-up
Use of unfair means,
penalties and safeguards
Course etiquettes
A Summary of To-Dos for You
Everybody
Refresh your calculus, probability theory, linear algebra basics
Start learning/refreshing Python and LaTeX skills
Create a homepage on CC/CSE home servers
Students who are already registered
Start forming groups of five (5) students – do not wait long
Students who wish to audit
Send an email to the instructor if not already done so
Students who wish to credit
Apply during add-drop – will use published policy to accept
Policy: https://www.cse.iitk.ac.in/users/purushot/courses/ml/faq.html
A Teaser
 What is the point of machine learning?
 A few cool ML apps developed by your peers
What is the point of ML anyway? “
“ TheThe
art art
andand
science of designing
science of dealingadaptive algorithms
with uncertainty

ML is a way to uncover hidden patterns in data

ML is a way to automate tedious and repetitive tasks
ML is a way to predict the future by looking at the past

At a high-level ML does this by

Looking at lots of data to examine input-output behaviour
Replicate that behaviour by writing a program
Machine Learning “ 15
“ TheThe
art art
andand
science
A Non-adaptive Algorithm
of designing
science of dealingadaptive

Sorting: given numbers, sort them in

decreasing order of their value
algorithms
with uncertainty
An Adaptive Algorithm
Recommendation: given a person
John and items, sort items in
INPUT OUTPUT INPUT OUTPUT decreasing order of how much John
4 9 5 5 likes them
1 7 -6 4
5 5 4 1
9 4 -3 0
ML can help you learn
3 3 -2 -2 patterns that allow you to
7 2 1 -3 sort the same set of items
2 1 0 -6 differently for each person
according to their taste
Machine Learning 16
“
“ TheThe
art art
andand
science of designing
science of dealingadaptive algorithms
with uncertainty
Machine Learning “
“ TheThe
art art
andand
science of designing
science of dealingadaptive algorithms
with uncertainty
When to apply ML 18
Complexity: no “closed form” solutions macine
Humans cannot specify simple rules to get solution
Detecting spelling mistakes not a good ML problem
A simple dictionary lookup (binary search) is enough macine
Presence of immense variety
Too many variants to be solved independently
machine
Correcting spelling mistakes a very good ML problem
Need for automation
Scalability and speed are main criterion
Do we need to automate medicine, driving?
ML for e-Governance
A novel search engine for public
grievances at the MoD
Emergency warning / dispatch systems
Accountability
Estimate income of citizens
Determine eligibility for public benefits
services
Convert existing pull model to a push model
Automatically recognize land parcel
boundaries in rural areas
Identify infrastructure damage on
national highways
Assess volume of material moved at mining
sites
Figure courtesy Prof. Nisheeth Srivastava
Understanding Cancer Progression with
ML Understand intratumor heterogeneity
Better diagnosis and treatment
Reduce relapse and drug resistance

Figure courtesy Prof. Hamim Zafar https://hamimzafar.wixsite.com/home/tumor-heterogeneity-and-evolution

Efficient Policing with ML
Suggest reasonable patrol locations
and routes based on call records
Promote fuel economy, reduce
response time

Avijit Roy and Nisheeth Srivastava, Decentralized reinforcement learning for multi-agent patrol routing, AASG@AAMAS, 2021
Reveal Cybersecurity Vulnerabilities
Clean Air with ML
Vehicular Pollution
Urban Organic Aerosols
Industrial Emissions
Transported Organic Aerosols
Biomass Combustion
Dust Related

An example of Hydrocarbon-like Organic Aerosol (HOA) levels

predicted by a machine learning algorithm for a period of two
days. The algorithm tracks well the peaks and dips in HOA levels

Deeksha Arora, Akash Halayyanavar, P. K., Sachichidanand Tripathi, Low-Cost Real-time Source Apportionment, manuscript, 2024
Anomaly Detection in Satellites
Fully autonomous missions a lofty goal in space exploration
Space missions still rely on ground staff support
Vast amounts of telemetry sent to mission control to be manually inspected
Expert ground staff provide remediation telecommand in case of anomaly
Early detection key to success of this setup
Enable ground staff to quickly take curative action

TE
LE ND
Critical to preserving spacecraft and mission health

CO
M
M
TE

A
LE
M
ET
RY
Prajjwal Yash, Sharvari Gundawar, Nitish Kumar, Uma B R, Krishna Priya G, Purushottam Kar,
"Multiforecast-based Early Anomaly Detection for Spacecraft Health Monitoring", CODS 2024
Sign Language Understanding

Figure courtesy Prof. Ashutosh Modi

Program Correction with ML 26
IITK offers a C programming course to ~1000 students each year
#include <stdio.h> #include <stdio.h>
int main(){ int main(){
int a; int a;
scanf("%d", a); scanf("%d", &a);
printf("ans=%d",a+10); printf("ans=%d",a+10);
return 0; return 0;
} }

Line-4, Column-9: warning: format ‘%d’ expects argument of type ‘int *’, but argument 2 has type ‘int’

Ahmed et al. Compilation Error Repair: For the Student Programs, From the Student Programs, ICSE-SEET 2018.
Program Correction with ML 27
IITK offers a C programming course to ~1000 students each year
#include <stdio.h> #include <stdio.h>
int main(){ int main(){
int x, x1 ,d; int x, x1 ,d;
d=(x-x1)(x-x1); d=(x-x1)*(x-x1);
return d; return d;
} }

Line-4, Column-11: error: called object type ‘int’ is not a function or function pointer

Ahmed et al. Compilation Error Repair: For the Student Programs, From the Student Programs, ICSE-SEET 2018.
ML for Program Retrieval

Figure courtesy Prof. Amey Karkare

Lifelong Learning

Lifelong Deep Generative Models (without forgetting)

Figure courtesy Prof. Piyush Rai

Image Denoising
Original
30

Poisoned
Image De-watermarking
Original
31

Poisoned
Stay Awesome!
See you in the next one