Niantic is pioneering the concept of a Large Geospatial Model using machine learning to understand scenes globally. Spatial intelligence is the next frontier of AI models, allowing for new interactions with physical spaces. Geospatial models go beyond 3D vision by capturing 3D entities rooted in geographic locations. The Large Geospatial Model enables strong positioning capabilities globally, akin to how humans perceive and imagine the world. MicKey, a neural network, showcases the potential of an LGM for advanced camera positioning. With a unique data influx, Niantic aims to lead the way in developing the Large Geospatial Model for spatial computing’s future.
https://nianticlabs.com/news/largegeospatialmodel