Pramit is a 1st Year PhD student at the University of Michigan. He achieved AIR 1 in GATE 2020 in Statistics. He is an ISI Calcutta graduate and an INMO Awardee.
Abstract
We plan to learn the fundamental uses of Eigenvalues, matrix factorization, and how we use those notions in ML. In the second part, we intend to go over some basic statistical distributions and a very simple yet useful technique called Naive Bayes Classification.
Aman Bhardwaj
Topic: Linear Regression and Multi-Variable Calculus
Speaker Bio
Aman is a 2nd Year MS(Research) student in the CSE Department at IIT Delhi. He was the winner of the ICVGIP 2020 Visual Data Challenge.
Abstract
Aman will broadly be covering the following topics:
Introduction to ML, its types and important general concepts such as overfitting
A tour of Multi-Variable Calculus, starting from scalars through Gradients, Convexity and ending with applications
Linear Regression - Basics, Algorithm and hands-on Demo.
Ananye Agarwal
Topic: Introduction to Deep Learning
Speaker Bio
Ananye is a 4th Year B.Tech student in the CSE Department at IIT Delhi. He recently authored a paper that was accepted at ICML 2021. He has also interned at Microsoft Research India, and was AIR-3 in the IIT-JEE 2017.
Abstract
Ananye will be starting with a brief tutorial on Logistic Regression, and will then move on to Deep Learning. He will talk about the motivation behind it and fundamental concepts of Neural Nets, followed by a hands-on introduction to Pytorch.
Jay Paranjape
Topic: Introduction to Convolutional and Recurrent Neural Nets
Speaker Bio
Jay will graduate this year from the Computer Science Dept IIT Delhi. He has been associated with ML right from his first year and has experience working on problems with industries like Bosch and Microsoft. He will now be joining Microsoft as an ML engineer. He is currently working with Prof Chetan Arora in collaboration with AIIMS.
Abstract
Computer Vision and Natural Language Processing are two of the most studied areas of Deep Learning. Computer Vision is used in various fields like Medical imaging, automatic driving and so on. Similarly, NLP is used in chatbots, recommendations, world understanding and so on. This talk will take you a bit deeper into what makes the computer understand images or text - CNNs and RNNs. Get to know more about these building blocks how they have been crucial to Machine Learning over the years and we hope you will be using them in your own projects soon enough.
Anshul Mittal
Topic: More Methods in Machine Learning
Speaker Bio
Anshul is a Google PhD Fellow at IIT Delhi, and is in his 2nd year. He has authored 6 Publications in top ML Conferences and Journals.
Abstract
Anshul will be covering some interesting traditional ML Methods. The algorithms to be discussed include Gaussian Discriminant Analysis, K-Means and Expectation-Maximization. This will serve to give an idea of the diversity in ML Methods, beyond Regression and Deep Learning.
