Abhinav Moudgil

Deep Learning, AI and Open Source lover.

Humor Generation with Recurrent Neural Networks
This blog post discusses an attempt to generate funny text, using Long Short Term Memory (LSTMs), trained on a large corpus of jokes, which I compiled by crawling various websites.
TMVA Project in Machine Learning
This is my concluding blog post for Google Summer of Code 2016 with CERN SFT. I worked on the feature engineering module of the Toolkit for Multivariate Analysis (TMVA). Details like the code I worked on, and the challenges I faced during the course of this project, are discussed in this post.
Autoencoders Implementation Design II
I have developed the first working prototype of Deep Autoencoders in TMVA. I have attempted to keep the implementation simple and scalable. This blog post showcases the user interface and implementation design of Deep Autoencoders in TMVA.
Autoencoders Implementation Design I
One of the major tasks of my Google Summer of Code project was to implement Deep Autoencoders in TMVA. In order to do this efficiently, the current TMVA design requires some major design changes. This blog post explores the codeflow and issues in the current design.
Progress Update
After a few discussions over Skype with my Google Summer of Code mentors, we came to the conclusion that we should showcase the implemented methods using graphical plots for better understanding. I took a step ahead in this direction, along with some TMVA C++ classes updates, which I have talked about in this blog post.
Getting started with Variable Transformation
The first week of Google Summer of Code has passed, and I have been able to complete the first task of my project proposal quite smoothly! A great start indeed! In this blog post, I have enlisted the conclusions of my study of TMVA codebase, and details of the method implemented.
Google Summer of Code 2016 at CERN
The results of Google Summer of Code 2016 have been announced. I am so excited to work with CERN SFT on the TMVA Project in Machine Learning! I will blog about my progress once every two weeks. All the details of my project are covered in this introductory blog post.