The National Weather Service's Challenge

The National Weather Service protects the life and property of all Americans during hazardous weather events. To reach U.S. residents that speak a language other than English, NWS requires the highest possible translation accuracy and ability to instantly translate weather products into multiple languages, including languages NWS had not previously supported.

A Dynamic, Adaptive Solution

LILT’s neural translation software learns in real-time from NWS forecaster input. Our CAT tool provides translation suggestions, which linguists can approve or edit. The adaptive system learns from corrections and incorporates feedback in future translations. This innovative technology allows linguists to translate more quickly without sacrificing quality and enables organizations to scale adapted models for more accurate machine translation.

"[Our project with LILT] applies automation to gain efficiency with translation turnaround times while reducing the burden of translation on our personnel, especially during critical hazardous weather events."

Monica Bozeman, Automated Language Translation Lead

Faster Weather Report Publication

In dangerous weather situations, every second counts. LILT’s adaptive translation engine is continuously learning agency-specific terminology and using NWS linguist activity to improve translation velocity. With LILT, NWS translations that previously would have taken hours will take minutes.

Lower Risk of Mistranslations

A translation error in emergency weather alerts could have severe human consequences. To ensure the highest possible accuracy, NWS linguists work directly in the LILT platform, reviewing and approving predictive translation recommendations. The resulting adapted memories also enable higher accuracy for fully automated workflows.

Tools to Streamline Translation

NWS linguists rely on the LILT translator platform to make checking and editing weather products easier. Our intuitively designed, AI-powered platform removes manual and redundant steps, so linguists can focus where their skills are most needed.