Rochan Avlur Venkat

I'm a senior-year undergraduate student pursuing my B.Tech in Computer Science and Engineering at Mahindra Ecole Centrale.

Currently, I am working under Prof. Chandrajit Bajaj at the CVC lab, University of Texas, Austin on using Mulit-Task Reinforcement Learning and Variational Auto-Encoders to attain efficient cancer prognosis. During the summer of 2019, I worked at the Web Science Lab at the IIIT - Bangalore under Prof. Srinath Srinivasa. Prior to that, I worked under Anupam Mediratta on developing a Python Library for Sentence Prediction using Topic Modeling.

Email / GitHub / Google Scholar / Twitter / CV

Research

I'm interested to explore and understand the most mesmerizing aspect of human intelligence: our ability to extract and abstract information so rapidly & flexibly by perceiving & interacting with our environment. My burning question is: How do we infer so much from so little? In pursuit of this question, I am exploring and learning about ideas in deep learning, programming languages and logic.

News

Mar '21	Accepted UT Austin MSCS offer! Will join Fall 2021.
Mar '21	Our paper on training NN using DE has been accepted at IEEE CEC 2021 for Oral presentation!
Mar '21	Our paper on comparitive study of training NN has been accepted at IEEE CEC 2021 for Oral presentation!

Publications

Projects

Internships and other professional experience.

Exploring Feature Extraction of Atari 2600 Games with Deep Reinforcement Learning

Mentor: Prof. Achal Agarwal
Mahindra Ecole Centrale, 2020
code /

Exploring feature extraction in Reinforcement Learning problems, specifically tackling Atari 2600 game environment. Working with unsupervised image segmentation DL models, computer vision techniques and popular RL algorithms such as DQN, A2C and PPO. Conducting a study on interpretability and transferability of RL algorithms.

Topic2Document

Mentor: Prof. Srinath Srinivasa
Web Science Lab, International Institute of Information Technology - Bangalore, 2019
code /

Convert a given input bag of words vector into a natural language virtual document. This library was developed for the Narrative Arc project.

Sentence Prediction using Topic Modeling

Mentor: Anumpam Mediratta
ShowUpHotels, 2018
code /

A Python library to train topic models for sentence prediction using gensim and MALLET. It was developed as a summer internship project under the guidance of Anupam Mediratta.

Other Projects

These include side projects and hackathon stuff.

	Predict future jobs using historic job data and news data Hackathon Winner, Smart India Hackathon, 2020 website / Our solution basically uses historical job data for information about the location, salary and requirements and News articles that provide information about the future (ie. predictions).
	Distributed Compute Fabric using Mobile Devices Hackathon 3rd Place, Start-up Sprint, E-Summit @ MEC, 2020 Platform to make computation on smartphones accessible and profitable. Designed and implemented efficient data and compute traffic distribution algorithms.
	AI/ML for Cart Conversion Hackathon Winner, Dell EMC Hack-to-Hire, 2019 Multifaceted solution to solve the cart conversion problem faced by e-commerce companies. Contribution included developing a recommender system using Bayesian Networks and ensemble learning (Classification and Regression methods).
	DinoEnv Gym Environment Project College Dorm, 2019 code / OpenAI Gym environment based on the Google Chrome Dino game.
	oWatcher Project College Dorm, 2017 code / Discord bot to display detailed in-game performance statistics of a players in Overwatch by Blizzard Entertainment.