Location Prediction using Language Variation.
TwiLoc investigates the feasibility of geographically locating Twitter users based solely on tweet content. We are trying to locate a user using their tweet content by understanding the dialect differences across geographies through deep learning techniques. We are not using any other external information to locate the user. This project provides an approach to augment existing systems that locate users.
Click Here for GitHub Repo...
The focus of the PhysioNet/CinC Challenge 2012 is to develop methods for patient-specific prediction of in-hospital mortality. Participants will use information collected during the first two days of an ICU stay to predict which patients survive their hospitalizations, and which patients do not. We have performed exploratory analysis initially to design the feature matrices. Later we tried different regression and classification techniques using the designed feature matrices to predict the in-hospital mortality and length of stay.
Click Here for GitHub Repo...
Link to Tableau Dashboard used for exploration.
During my period at NUS, developed a reinforcement learning agent for side-scrolling games like FlappyBird, PixelCopter. I worked on this project for CS4246 Module - AI Planning and Decision Making. This project is an implementation in Keras and OpenAI Gym of the Double Deep Q-Learning algorithm (DDQN) by Mnih et al.
Click Here for Project Report...
Year:2017 - Present
I am developing a personal bot for me. It is basically a digital assistant that runs on my Android phone. Saving my time from social media like a Digital Dietitian is what my bot is doing for me now. It notifies me with motivational quotes to make me close the social media apps. If I am too adamant to come out of social media, it will even create fake emails, WhatsApp messages to distract me from social media. My bot saves my time not only from FB, Instagram but also from apps like COC, 8 Ball Pool etc. Soon my bot will keep me posted about serials, movies that interest me the most. I have plans to open source this project.
Under the initiative named SmartDroid, I developed an Augmented Reality application for Amrita School of Engineering TechFest called Anokha in the year 2015 during my undergraduate program. Anokha AR is a Location-based Augmented Reality application that will help TechFest visitors to find the location of events happening inside the campus during the TechFest.
During my period at NUS, developed a distributed fault-tolerant peer-to-peer maze game using Java RMI. I worked on this project for CS5223 Module - Distributed systems. The primary goal of this project is to design and implement a fault-tolerant distributed system. Our program successfully passed all the checkpoints during the stress test containing 200 test cases.
Under the initiative named SmartDroid, I developed Android applications for Amrita School of Engineering students during my undergraduate program. ASE AUMS app will help the student access student portal called AUMS easier. ASE apps will help the students access faculty details and latest news inside the college etc.
I developed an Android application having information about hotels, restaurants, movie theaters etc located in Tamilnadu, India. All the data will be downloaded from the server for the first time and made available offline. So that users can access the information whenever needed without the Internet. It will be useful for someone who is new to Tamilnadu. I developed and published this app during my undergraduate program out of curiosity for actual people using my application.
Live Wallpaper used to catch my attention easily. I always wanted to try out new Live wallpapers available in the play store. One day I decided to create one for me. By animating a sequence of images I created one live wallpaper for my Android phone. Published the same out of curiosity.