DBTB INT Barak Turovsky
summary
Google Translate is presented as one of the oldest and largest machine learning systems with a 10th anniversary milestone and a recent disruptive shift toward neural networks. The product and design team focuses on more than model training: heavy investment goes to data preprocessing and postprocessing, pipeline engineering, and tooling to massage corpora. The claim: training data alone is insufficient; practical gains come from pipelines, metrics and deployment work, and talent and education remain scarce.
AI-written summary · about
related talks
07mMachine learning at Google
27mKeynote - Google Cloud, Machine Learning
39mDataEngBytes 2023 - BRI-T1-02 - Mikiko Bazeley
09mInterview with Martin Görner
38mBuilding robust machine learning systems
26mQuoc Le, Software Engineer, Google
38mFrom Experimentation to Products: The Production Machine Learning Journey
38mData science at The New York Times