Speakers: Dipam Paul, Alankrita Tewari
Summary
In this talk, our mission is to highlight and try to solve one of the most pressing problems that exist in the world of training and deploying Neural Networks. This problem is called the 'Minority Class Imbalance Problem'. Our proposed technique comprising Deep Generative Models would not just solve the problem but would also show a way to how one can seamlessly attain state-of-the-art accuracy!
Dipam Paul's Bio
Commonly referred to as ‘The Boy from Kolkata’ - Dipam is a recent Engineering graduate with a major in Electronics and Telecommunication from KIIT University, India.
Currently, he has joined Accenture, India as a Software Engineer. Prior to this, he was working as a Research Assistant @ School of Computer Science, Carnegie Mellon University. He primarily worked on Object Detection problems concerning 3D-Cryo ET data and areas of Deep Active Learning with focus on Dynamically Expandable Networks.
He has previously worked in and collaborated with labs at Georgia Institute of Technology, Universidade Federal de Sao Paulo, and IIT-Bombay in various roles and capacities.
Having always been fascinated by the wonders one can do using Python and Data, his periphery of interest largely lies in Deep Learning, Computer Vision, and Intelligence Theory related problems.
In the pandemic, he spends his days toying around with Deep Generative Models and fine-tuning Neural Nets when he is not eating, raconteuring, or engaging in a lively debate about Geopolitics or Football!
- GitHub: https://github.com/dipampaul17//
- Want to help add timestamps to our YouTube videos to help with discoverability? Find out more here: https://github.com/numfocus/YouTubeVideoTimestamps
