Abhiram Follow Abhiram - Co-Organiser

Talks – July, 2017

For July’s session, we started with a “Machine Learning Basics” theme. The venue was Treebo Hotels, and there were 4 speakers. The talks were each of 40-minutes.

The first talk was by Shrishty Chandra and her talk was “An Introduction to Linear Regression”. She started off with fitting sets of points into curves, taking up a usecase to predict the prices of houses, given a bunch of features to describe the house with and some amount of test data with which to train the model. Then she moved on to explaining the derivative and gradient descent ways of calculating the minimum RSS value - which is the sum of squares of the errors in predictions. This was followed by some demo-ed code for the same.

YouTube video for the talk -

The second talk was by Anand , titled “How to Teach Neural Networks to Your Grandma”.He mentioned and demonstrated systems like solving linear equations and identifying numbers that utilize existing training data to create models and then use them on test data.

YouTube videos for the talk -

The third talk was presented by Suresh Saini and he spoke on the topic “Why Deep Learning” . He showed a video on how Boston Dynamics utilized Deep Learning in robots and the origins of Deep Learning, and some reasons for why Deep Learning emerged now and discussed about why it wasn’t required in the past.

YouTube video for the talk -

In the final talk, Kenso spoke about the “Basics of Neural Nets” - what they are ,comparing neural nets to the physical anatomy of the human brain, what they use (supervised regression) and a usecase involving the hours of sleep and hours of study put in by a student and how that affects the marks he gets. He proceeded to speak about the importance of normalizing data and concluded with some references for the same.

YouTube video for the talk -