I am a fourth year Ph.D. student in the School of Interactive Computing at Georgia Tech. My research interests are natural language processing, machine learning and social media . I am currently working in the field of text simplification and stylistics with Professor Wei Xu as my advisor.
I received my Master's degree in Computer and Information Science from University of Pennsylvania (UPENN) and my Bachelor's degree in Computer Science and Engineering from International Institute of Information Technology Hyderabad (IIIT). At UPENN, I worked in the field of social media text analytics and at IIIT, I worked in data visualization. Prior to joining OSU, I also worked as a Software Development Engineer at Big Data Technologies, Amazon .
If you want to contact me, please drop me an email at firstname.lastname@example.org .
Designed a new hybrid model for sentence simplification task. Our approach combines linguistic rules with a data-driven Transformer model to generate a simplified version of the input complex sentence. Our model outperformed the state-of-the-art system in terms of both automatic metrics and human evaluation.
Developed a novel neural model to break a hashtag into its constituent words. Our approach addresses the diverse language style in social media and also adapts to the type of hashtag. Our model outperformed the state-of-the-art by 1.8 points in F1 and also improved the performance of the downstream sentiment analysis task by 2.4 points in F1.
Developed a classifier that captures word spelling patterns to predict how likely the input word can be a code token without any sentential context. When combined with a BERT-based Named Entity Recognizer, the classifier has shown to improve the recognizer performance on StackOverflow posts by 2.7 points in F1.
The project was a part of World Well Being Project (WWBP). We captured the different sources and interpretations of well-being across various cultures or countries. using the context of sentiment words and their distribution across countries.