Welcome!

I am developing AI models for 3D medical images at ImmersiveTouch. During my graduate studies at Rensselaer Polytechnic Institute, I focused on Natural Language Processing and Machine Learning. I have experience training and fine-tuning large language models like BERT and GPT for domain adaptation and information extraction.

Overview
My graduate research was on domain adaptation of neural models for causal information extraction. I experimented with unsupervised domain adaptation methods for extraction of cause and effect spans from text. Integrating domain independent linguistic information in neural models led to improvement in adversarial domain adaptation methods. I have also worked on multi-sense embeddings to address the meaning conflation problem in word embeddings since they encode different meanings of a word into a single vector. I implemented a knowledge distillation method to transfer the contextual information from pre-trained language models into multi-sense embeddings. I have also explored the application of simple word embeddings in semantic search. During my internship at IBM Research, I worked on information extraction from structured documents such as scanned images, pdfs etc.