Word Prominence Detection
Experimental results on a corpus of spontaneous speech indicate that prominence detection accuracy using only the new prosodic features is better than using both lexical and syntactic features.
Hi, I'm Harshvardhan Chavan.
I’m an undergraduate student pursuing dual degrees—one at IIT Madras and the other at Pune University. My heart beats for Computer Science, and I’m passionate about software design, artificial intelligence, machine learning, computer vision, and natural language processing. Currently, I’m working towards a Bachelor of Technology (BTech) in Computer Engineering from Pune University and a Bachelor of Science in Data Science and Applications from IIT Madras.
Recent Work
Card Vision
Implemention of a software that can detect playing cards using images. Experimented on a wide number of images containing multiple cards and some, even with a highly skewed image, with a high success rate.
Quora Duplicate Question Detection
Explored Natural Language Understanding (NLU) by building various classification models - Linear Models, Tree Based Models and Neural Networks and compared the performance of each of these.
Micrograd
A tiny scalar-valued autograd engine and a neural net library on top of it with PyTorch-like API.
Pong-AI
An AI teaching itself how to play the game of Pong. It's doing it by using an algorithm called NEAT which stands for Neuro Evolution of Augmented Topologies.
makemore
From scratch implementation of a sparse mixture of experts language model inspired by Andrej Karpathy's makemore.