Mixed Data Learning for Approximate Spatiotemporal Queries
Find Innovative SolutionsGenerate Solutions
Solution Overview
Problem
Existing technologies have not effectively addressed the limitations of applying approximate query techniques to large-scale mixed data, particularly those with time and space concepts such as vehicle travel paths, and there is a need for improved methods to efficiently analyze data in areas like traffic/navigation, autonomous vehicles, and bio/medical data analysis.
Innovation Solution
A machine-learning-based apparatus and method that discretizes and encodes data using a spatiotemporal model, utilizing a probabilistic inference model to generate a mixed learning model that learns relational data and spatiotemporal data, utilizing a probabilistic model for solving the technical problem.
Engineering Contradictions & Design Principles
Engineering Contradiction Analysis
1Productivity
If approximate query techniques based on machine learning are applied to relational data, then query efficiency is improved, but applicability to spatiotemporal data (e.g., vehicle travel paths) is limited
Solution Approach 1:
The patent segments the learning process into two distinct models: a relational data learning model and a spatiotemporal data learning model. Each model is specialized for its data type, allowing the system to efficiently process relational data while also being applicable to spatiotemporal data like vehicle trajectories. The segmentation enables targeted optimization for each data type without compromising the other.
Solution Approach 2:
The patent creates a universal learning framework that can handle both relational data and spatiotemporal data through multiple specialized models. The system is designed to be multi-functional, accommodating different data types (relational, spatiotemporal, and their mixtures) within a single architecture, thereby improving adaptability while maintaining query efficiency across diverse data types.
2Adaptability or versatility
If mixed data including relational data and spatiotemporal data is processed, then analysis capability is improved, but computational complexity increases
Solution Approach 1:
The patent divides the computational process into separate relational and spatiotemporal learning models, each handling specific aspects of the mixed data. This segmentation reduces the computational complexity of the overall system by avoiding the need for a single monolithic model to handle all data types simultaneously, while still providing comprehensive analysis capability.
Solution Approach 2:
The patent introduces an intermediary processing layer that manages the interaction between relational and spatiotemporal models. This intermediary coordinates the learning processes and integrates results, simplifying the computational complexity by providing a structured interface between different data types and their respective models.
3Measurement precision
If data is discretized based on level of detail for each designated area, then query accuracy is improved, but processing time increases
Solution Approach 1:
The patent applies local quality by discretizing spatiotemporal data with different levels of detail based on designated areas. High-precision discretization is applied only where needed (in designated areas of interest), while other regions use coarser discretization. This approach maintains query accuracy for important locations while reducing overall processing time through selective detail levels.
Data Source
AI summary
Disclosed herein is an apparatus and method for learning mixed data for approximate queries. The apparatus receives mixed data including relational data about information for identifying an object and spatiotemporal data about the trajectory of the object moving in a target space, discretizes the relational data and the spatiotemporal data based on a level of detail that is preset for each designated area of the target space corresponding to the trajectory of the object, and generates a mixed learning model that learns the relational data and the spatiotemporal data for each level of detail using multiple relational models and spatiotemporal models.