Early versions of Snorkel Flow's core technology have been developed in partnership with — and deployed at — some of the world’s most sophisticated ML organizations, including several deployments publicly described in peer-reviewed case studies:
Google used Snorkel to replace 10-100K hand-annotated labels in key ML pipelines
Content, product, and event classification problems change too fast to hand-label, even with significant annotation budget
We deployed early versions of Snorkel Flow's core technology with three high-impact teams at Google, repurposing many organizational resources as labeling functions
Hours of labeling function development replaced 10-100K+ hand labels, significantly impacting the bottom line and acceleration of ML solution adoption
Hand labels replaced
Improvement by repurposing resources
Labels in < 30 min.
Intel used Snorkel to replace a high-cost, high-latency crowdsourcing pipeline and accelerate sales and marketing agents
Rapidly changing sales goals make social media monitoring difficult to maintain
We deployed a prototype version of Snorkel Flow ("Snorkel Osprey") to replace months-long crowdworker processes with cheap and fast template-based programmatic labeling
Better performance and major cost savings in Sales & Marketing and Advanced Analytics
Months of crowdworker labels replaced
Precision percentage points
Coverage percentage points
Researchers at Stanford Medicine used Snorkel to label medical imaging & monitoring datasets, replacing person-years of hand labeling with several hours of using Snorkel
Labeling training data for triaging models takes person-months to person-years of radiologist time
We deployed a cross-modal Snorkel pipeline, matching or exceeding the performance of painstakingly gathered manual labels in hours
Currently being tested for deployment in Stanford & Department of Vetaran Affairs (VA) hospital systems
Person-months of labeling replaced
ROC AUC Performance
Images labeled in minutes
Top U.S. Bank
A top U.S. bank uses Snorkel Flow to quickly build AI applications that classify and extract information from their documents.
The bank estimated that, for a time-sensitive use case, hand-labeling data would take over a month.
With Snorkel Flow, the team produced a solution that was over 99% accurate in under 24 hours.
The resulting AI application could be quickly and easily adapted to new problems and business lines.
Snorkel Flow Accuracy
From problem start
# Documents processed