On-Device Generative AI for Customized Digital Maps
Find Innovative SolutionsGenerate Solutions
Solution Overview
Problem
Current methods for providing customized digital maps require significant bandwidth and compute resources, as users rely on server computing systems to render and download map tiles, which can be inefficient in low bandwidth or low network coverage areas.
Innovation Solution
Implementing a generative machine-learned model on a user's computing device allows for the direct rendering of customized digital maps, reducing the need for constant network communication and enabling personalized map content while conserving bandwidth.
Engineering Contradictions & Design Principles
Engineering Contradiction Analysis
1Adaptability or versatility
If server computing systems render and download map tiles for customized digital maps, then map customization and rendering capability are improved, but bandwidth consumption and network dependency increase
Solution Approach 1:
The patent extracts the map rendering capability from the server computing system and relocates it to the user computing device through on-device machine learning models. This allows the device to generate customized map tiles independently without continuous network communication, significantly reducing bandwidth consumption while maintaining customization capabilities.
Solution Approach 2:
The user computing device performs self-service by executing machine learning models locally to generate customized map tiles. The device autonomously processes customization requests and renders maps without requiring constant server intervention or network resources, enabling independent operation in low-network conditions.
2Manufacturing precision
If server computing systems render customized map tiles, then map rendering quality is improved, but network coverage dependency and latency increase
Solution Approach 1:
The rendering capability is extracted from the network-dependent server system and embedded within the user computing device through machine learning models. This enables high-quality map rendering to occur locally without network coverage, eliminating the reliability issue while maintaining rendering precision.
Solution Approach 2:
The system performs preliminary actions by pre-loading and caching machine learning models and base map data on the user device before customization is needed. This allows the device to immediately generate customized maps locally without waiting for network requests, reducing latency and improving reliability.
3Loss of energy
If machine learning models are deployed on user computing devices, then bandwidth savings and offline capability are improved, but device computational resources and storage requirements increase
Solution Approach 1:
The patent segments the map rendering task into two parts: a lightweight machine learning model that runs on the user device for customization, and a separate server system that provides computational-intensive base map generation and rendering. This segmentation allows the device to handle only the customization portion, reducing the computational burden while still achieving bandwidth savings.
Data Source
AI summary
A computing device for generating a customized digital map includes one or more memories to store instructions and one or more processors to execute the instructions to perform operations, the operations including: receiving an input from a user relating to customizing features associated with a location viewable on a digital map; in response to receiving the input, implementing a generative machine-learned model to generate the customized digital map which depicts the location with one or more customized features which are generated via the generative machine-learned model based on the input; and providing the customized digital map for presentation via a display device.