Case Study: Georgetown University’s CSET achieves 85% classification accuracy and 50% faster labeling with Snorkel AI

Georgetown University’s CSET - Customer Case Study

Georgetown University’s CSET partnered with Snorkel AI to accelerate AI development and improve cross-team collaboration for NLP models that classify scientific articles (for example, technical papers on virology). The team’s challenge was that manual labeling was impractical at scale, and they needed high-quality, trustworthy models to support policy recommendations.

Snorkel AI implemented Snorkel Flow’s programmatic labeling, autosuggest and cluster labeling‑function features, plus integrated analysis and an active-learning workflow so domain experts could spot-check data slices and troubleshoot. The solution produced 107k programmatic labels, cut labeling time by 50%, and yielded a classification model reaching 85% accuracy within days.

Georgetown University’s CSET