Data Programming

Presented two papers at EMNLP 2022 Findings - a benchmark for post-OCR text correction in Sanskrit and a demonstration of SPEAR data programming library.

A widely-adopted open-source Python library for programmatic data labeling with 100+ GitHub stars. SPEAR reduces data labeling efforts by implementing cutting-edge data programming approaches (Snorkel, ImplyLoss, Learning to Reweight). Integrates semi-supervised learning for efficient training and inference. Featured at EMNLP 2021.