Projects

Here are some projects I have worked on.

Automatic Essay and Grammar Scoring

Developed a biLSTM-based model for grammar and essay scoring. Implemented augmented C&W word embeddings which would treat grammatical errors as informative.

Key Information Guide Network

An abstractive text summarizer model which has a encoder for keyword representation, an only attention-based pointer-generator network to generate summaries, and a value estimator network.

Sarcasm Detection

A neural network and a SVM model predicting if a tweet is sarcastic based on lexical, pragmatic, lexical incongruity and context incongruity features.

Traffic Sign Classification

Implemented a three-layered CNN architecture which would classify a given traffic-sign into one of the 47 traffic-sign categories with 97.75% accuracy on the test dataset.

OSAVA

OSAVA is an Operating Systems Algorithms Visualization app. It visualizes standard versions of the OS algorithms as given in Silberschatz et al. (2012).