• Home
  • Technology
  • Gaming
  • Entertainment
  • World & Business
  • Science
  • Sports
  • AI
HomeTechnologyGamingEntertainmentWorld & BusinessScienceSportsAI
  • HomeTechnologyGamingEntertainmentWorld & BusinessScienceSportsAI
    • Home
    • Technology
    • Gaming
    • Entertainment
    • World & Business
    • Science
    • Sports
    • AI
    AI

    Google and NASA JPL Release MAPL-EMIT Methane Model

    Tweet by Google DeepMind engineer shares release of MAPL-EMIT vision transformer developed with NASA JPL.

    👩‍💻P
    1 Source, 29d ago, first seen 29d ago

    TLDR

    Paige Bailey posted that Google and NASA JPL released MAPL-EMIT, a vision transformer that detects methane leaks from space and identifies their sources. The model was trained using physics-simulated plumes. The post links to a Google Research blog titled Mapping global methane emissions from space with deep learning, written by Vishal Batchu and Michelangelo Conserva. The tweet notes methane drives about 25 percent of global warming and is invisible to the naked eye. An attached video is mentioned but not described further.

    Combined views

    4.7K

    1 Source, first seen 29d ago

    Combined views

    4.7K

    1 Source, first seen 29d ago

    89 likes
    89 likes
    2 comments
    21 saves
    13 reposts

    Sentiment

    Positive——Negative

    Summary

    Not enough discussion yet.

    No sentiment analysis available yet.

    2 comments
    21 saves
    13 reposts

    Sentiment

    Positive——Negative

    Summary

    Not enough discussion yet.

    No sentiment analysis available yet.

    Today's Rank

    —

    Not ranked yet

    Today's Rank

    —

    Not ranked yet

    1 Source

    @DynamicWebPaige👀 You can’t fix what you can’t see. Methane is invisible to the naked eye, but drives ~25% of global warming. @Google + @NASAJPL just released MAPL-EMIT: a vision transformer that spots methane leaks from space and then pinpoints their exact source. To train it, the researchers injected 3.6M physics-simulated plumes into real NASA EMIT satellite scans. It captures 84% of expert plumes and maps 50% more leaks than existing methods. @GoogleAI's open-sourcing the model, synthetic data, @GoogleEarth Engine app, and code - take a look! 🛰️🌍 https://research.google/blog/mapping-global-methane-emissions-from-space-with-deep-learning/

    1 Source

    @DynamicWebPaige👀 You can’t fix what you can’t see. Methane is invisible to the naked eye, but drives ~25% of global warming. @Google + @NASAJPL just released MAPL-EMIT: a vision transformer that spots methane leaks from space and then pinpoints their exact source. To train it, the researchers injected 3.6M physics-simulated plumes into real NASA EMIT satellite scans. It captures 84% of expert plumes and maps 50% more leaks than existing methods. @GoogleAI's open-sourcing the model, synthetic data, @GoogleEarth Engine app, and code - take a look! 🛰️🌍 https://research.google/blog/mapping-global-methane-emissions-from-space-with-deep-learning/