Pokemon GO Data Used to Train Geospatial AI Model

by

Pokemon GO Data Used to Train Geospatial AI Model

Key Takeaways

  • Niantic is building a geospatial AI model using data from popular AR games like Pokemon GO.
  • The model, called the Large Geospatial Model, aims to enhance interactions in physical spaces globally.
  • The technology has vast applications in AR glasses, robotics, content creation, and autonomous systems.

As a long-time enthusiast of AR gaming and a witness to the evolution of spatial computing, I find Niantic’s new geospatial AI model, the Large Geospatial Model (LGM), absolutely fascinating! The potential applications, from AR glasses to autonomous systems, are mind-boggling.

The creators of Pokemon GO, Niantic, have disclosed that they are utilizing game data to construct a cutting-edge geospatial AI model. This model is designed to enhance the ability of artificial intelligence, powered by neural networks, to interact more effectively with physical environments worldwide. In simpler terms, Pokemon GO, an AR adaptation of the renowned franchise, captures player data from smartphones and neural networks to integrate virtual items into real-world settings.

John Hanke, the founder of Niantic, previously worked at Google’s Geo division, where he played a significant role in developing Google Maps, Earth, and Street View. Prior to joining Google, his company Keyhole developed an interactive mapping interface for real-time tracking of conflicts during the Iraq War. Nowadays, Niantic is renowned for its collection of popular augmented reality games such as Pokemon GO, Ingress Prime, Pikmin Bloom, and others.

In a recent blog post, Niantic is developing a comprehensive geospatial model known as the Large Geospatial Model (LGM). This innovative project will consolidate over 50 million trained neural networks into a unified system that maps out geographical locations and their spatial relationships. The LGM is expected to be integral in various fields, such as augmented reality glasses, robotics, content creation, and autonomous systems.

Niantic Looks Ahead With New Tech Investment

Niantic is confident that wearable XR tech and self-driving machines will dominate the tech landscape in the coming years, and they’re banking on it. Their current 3D maps are shaped by countless scans from users of their AR games and Scaniverse app. These scans have been pooled to develop their Visual Positioning System, which precisely places virtual items in real-world spaces. The Large Spatial Model is the next step, merging all the neural networks generated by the VPS into a single entity that can estimate information about unscanned locations based on local spatial and position data.

While it’s not shocking for users of “Pokemon GO” to learn that the company behind such a popular game is gathering their data as part of its broader business objectives, the consequences of creating a vast spatial database are extensive.

With geospatial models advancing rapidly, it’s likely that Niantic or another organization will delve into fully utilizing the technology’s capabilities. It’s unclear at this point what rules or moral guidelines will be set for deploying this groundbreaking innovation, and what benefits and drawbacks consumers may experience with this new phase of AI development.

Read More

2024-11-22 23:13