Diganta Misra

			Diganta Misra
		
			Founder & Research Student
		
			Landskape
		
			Email: mishradiganta91[at]gmail[dot]com
		
			Please mention Landskape in the subject while sending mails.

Publications Research Code Talks

Introduction

Greetings! I recently completed my undergraduate in Electronics Engineering with a minor in Electrical Engineering under the supervision of Bhargav Appasani at Kalinga Institute of Industrial Technology.
My research interests include Abstract Algebra, Algebraic Topology, Lie Groups, Computer Vision, Deep Learning Theory (more specifically Mean Field Theory, Convex Optimization and Non-Linear Dynamics).
I also hold a position as a research affiliate at the Laboratory of Space Research, Hong Kong University (LSR-HKU).

News

[08/2020] One paper accepted to BMVC 2020.

Education

B.Tech. in Electronics and Electrical Engineering
2016 - 2020

School of Electronics, Kalinga Institute of Industrial Technology, India

Advisor: Bhargav Appasani

Selected Publications [Google Scholar]

Mish: A Self Regularized Non-Monotonic Activation Function
Diganta Misra

In Proceedings of the 31^st British Machine Vision Conference (BMVC), 2020.

[Paper] [BibTex] [Code] [BMVC 2020 Paper page]
[Loss Landscape video (in collaboration with Javier Ideami)]

Research and Experience

Research Affiliate at Laboratory of Space Research, Hong Kong University (LSR-HKU)
January 2020 - Present

Working on Planetary Nebulae analysis using deep learning and computer vision based approaches. Mentored by Kris Stern and Prof. Quentin A. Parker.
Research Intern at Bennett University
June 2018 - Aug 2018

Worked on large scale visual recognition and high performance computing for deep learning based vision applications. Advised by Prof Dr. Deepak Garg and Dr. Suneet Gupta.
Intern at Indian Institute of Technology, Kharagpur (IIT-KGP)
May 2018 - June 2018

Studied basic algorithmic tecbniques using functional programming languagues - Lisp and Prolog under the guidance of Pawan Kumar .
Exchange Student at Bangkok University
June 2017 - Aug 2017

Served as a primary instructor for cultural engagements along with teaching basic english and computer science to primary grade students at RangsonWittaya School, Nakhon Sawan under the AIESEC SDG #4 programme.

Talks

Podcast - Mish: A Self Regularized Non-Monotonic Activation Function [Talk]
2020

Podcast episode 7 with Miklos Toth on the Machine Learning Cafe podcast.
Research Presentation - Mish: A Self Regularized Non-Monotonic Activation Function [Talk]
2020

Presented internally at the Sicara weekly deep learning club.
Contributed Talk: Non-Linear Dynamics in Neural Networks
2020

Presented at the Deep Learning colloquium at the University of Athens.
Invited Talk - Mish: A Self Regularized Non-Monotonic Activation Function [Talk]
2020

Presented at the Computer Vision Talks.
Invited Corporate Talk - Mish: A Self Regularized Non-Monotonic Activation Function
2020

Internal (virtual) presentation at the Bangalore Robert Bosch office.