Digital Event Horizon
Niantic has announced plans to create an AI model for navigating the physical world using scans collected from Pokémon Go players and Scaniverse users. The company's "Large Geospatial Model" will utilize vast amounts of geolocated images to build a comprehensive navigation system, marking a significant milestone in location-based data collection.
Niantic is developing an Artificial Intelligence (AI) model for navigating physical spaces using geolocated images from Pokémon Go and Scaniverse. The company's Visual Positioning System (VPS) will serve as the foundation for this new AI model, which has been in development for over five years. The AI model, dubbed "Large Geospatial Model" (LGM), will process physical spaces using geolocated images and understand and map physical spaces like large language models process text. Niantic's data collection reveals the company's sizable presence in the AR space, with over 10 million scanned locations used to train more than 50 million neural networks. The technology has potential applications in augmented reality, robotics, autonomous systems, spatial planning, logistics, and remote collaboration.
Niantic, the company behind the popular mobile game Pokémon Go, has announced plans to create an Artificial Intelligence (AI) model for navigating the physical world using scans collected from players of its mobile games and from users of its Scaniverse app. This move marks a significant milestone in location-based data collection, as it leverages the vast amounts of geolocated images gathered through Pokémon Go and Scaniverse to build a comprehensive AI navigation system.
The company's Visual Positioning System (VPS), which has been in development for over five years, will serve as the foundation for this new AI model. VPS uses a single image from a phone to determine its position and orientation using a 3D map built from people scanning interesting locations in its games and Scaniverse. The company's recent Pokémon Playgrounds feature demonstrates the capability of Lightship Visual Positioning System, which allows users to place virtual items in real-world locations with centimeter-level precision.
The AI model, dubbed "Large Geospatial Model" (LGM), will process physical spaces using geolocated images collected through its apps. This system draws parallels to large language models like ChatGPT, but instead of processing text, it will utilize the vast amounts of location-based data to understand and map physical spaces. The scale of Niantic's data collection reveals the company's sizable presence in the AR space.
According to reports, over 10 million scanned locations worldwide have been used to train more than 50 million neural networks, each representing a specific location or viewing angle. These networks compress thousands of mapping images into digital representations of physical spaces, containing over 150 trillion parameters that help them recognize and understand locations. The multiple networks can contribute to mapping a single location, and Niantic plans to combine its knowledge into one comprehensive model that can understand any location, even from unfamiliar angles.
This technology has the potential to support various applications in augmented reality, robotics, autonomous systems, spatial planning, logistics, and remote collaboration. Niantic's announcement highlights the company's commitment to innovation and exploration of new frontiers in AI-driven navigation.
The reaction to this news among Pokémon Go players is varied, with some expressing surprise that their scans were being used for AI development. However, others have expressed a sense of déjà vu, given the company's business model has always revolved around collecting player data. The process was likely covered by Pokémon Go's data collection terms of service, but it remains to be seen how this news will impact the gaming community.
Niantic's AI navigation system marks an exciting development in location-based data collection and highlights the vast potential of AI-driven navigation systems. As Niantic continues to refine and expand its VPS technology, it will be crucial to consider the implications of such a system on various applications and industries.
Related Information:
https://arstechnica.com/ai/2024/11/niantic-uses-pokemon-go-player-data-to-build-ai-navigation-system/
Published: Tue Nov 19 18:53:48 2024 by llama3.2 3B Q4_K_M
