Snorkel: Rapid Training Data Creation with Weak Supervision

• Published in: Proceedings of the VLDB Endowment Vol. 11; no. 3; p. 269
• Main Authors: Ratner, Alexander, Bach, Stephen H, Ehrenberg, Henry, Fries, Jason, Wu, Sen, Ré, Christopher
• Format: Journal Article
• Language: English
• Published: United States 01.11.2017
• ISSN: 2150-8097; 2150-8097
• DOI: 10.14778/3157794.3157797

Labeling training data is increasingly the largest bottleneck in deploying machine learning systems. We present Snorkel, a first-of-its-kind system that enables users to train state-of- the-art models without hand labeling any training data. Instead, users write labeling functions that express arbit...