Share your thoughts, 1 month free Claude Pro on usSee more
WorkDL logo mark

Snorkel DryBell: A Case Study in Deploying Weak Supervision at Industrial Scale

About

Labeling training data is one of the most costly bottlenecks in developing machine learning-based applications. We present a first-of-its-kind study showing how existing knowledge resources from across an organization can be used as weak supervision in order to bring development time and cost down by an order of magnitude, and introduce Snorkel DryBell, a new weak supervision management system for this setting. Snorkel DryBell builds on the Snorkel framework, extending it in three critical aspects: flexible, template-based ingestion of diverse organizational knowledge, cross-feature production serving, and scalable, sampling-free execution. On three classification tasks at Google, we find that Snorkel DryBell creates classifiers of comparable quality to ones trained with tens of thousands of hand-labeled examples, converts non-servable organizational resources to servable models for an average 52% performance improvement, and executes over millions of data points in tens of minutes.

Stephen H. Bach, Daniel Rodriguez, Yintao Liu, Chong Luo, Haidong Shao, Cassandra Xia, Souvik Sen, Alexander Ratner, Braden Hancock, Houman Alborzi, Rahul Kuchhal, Christopher R\'e, Rob Malkin• 2018

Related benchmarks

TaskDatasetResultRank
Comment ClassificationCivil Comments
Accuracy73.9
30
Binary/Pairwise ClassificationSummarize
Accuracy70.5
9
Binary/Pairwise ClassificationChatbot Arena
Accuracy54.3
9
Binary/Pairwise ClassificationSHP
Accuracy61.9
9
Binary/Pairwise ClassificationPKU-BETTER
Accuracy57.5
9
Binary/Pairwise ClassificationPKU-SAFER
Accuracy57
9
scoringASSET
MAE29.073
5
scoringFeedbackQA
MAE0.793
5
scoringReview-5K
MAE2.593
5
scoringSummarize
MAE1.364
5
Showing 10 of 12 rows

Other info

Follow for update