Sentiment Classification Using BERT on IMDB Dataset
Worked on training and fine-tuning BERT models for binary sentiment classification using the IMDB Reviews dataset, including both classical and hybrid quantum-classical computing platforms. The role involved annotating and preparing text data for training, evaluating, and comparing model performance across CPU, GPU, and quantum backends. Responsible for designing and labeling sentiment classes, curating datasets, and verifying model predictions for accuracy. • Developed baseline and quantum-enhanced sentiment analysis pipelines. • Managed and processed large-scale movie review text data for training and validation tasks. • Performed data preprocessing, annotation, and binary sentiment labeling tasks. • Collaborated on comparative performance evaluation and error analysis across models.