Real-time Time Series Segmentation via Sparse Graph Recovery

Resolve Bottlenecks,
Find Innovative SolutionsGenerate Solutions

Solution Overview

Problem

Conventional systems are inefficient in real-time segmentation of time series data, often processing data offline and failing to accurately identify salient patterns due to complexities and noise, especially with multivariate time series data.

Innovation Solution

A real-time segmentation system that converts multivariate and univariate time series data into segmented data using time-based data windows, graph recovery models, and similarity models, with minimal memory and processing requirements, allowing for on-the-fly processing and accurate pattern identification.

Engineering Contradictions & Design Principles

VSEngineering Contradiction Analysis

1Measurement precision

If conventional systems store large amounts of time series data and process offline, then measurement precision is improved, but productivity deteriorates and time loss increases

Engineering Contradiction:

Improveaccuracy in identifying salient patternsVSAvoidprocessing efficiency

Core Design Contradiction:

Measurement precisionVSProductivity

Solution Approach 1:

The patent segments the time series data processing into streaming real-time operations and offline batch operations. The streaming segment processes data continuously with linear time complexity using graph recovery models, while the offline segment handles comprehensive analysis. This segmentation allows the system to achieve both real-time processing speed and accurate pattern identification without requiring storage of large amounts of historical data.

Inventive Principle:

Principle #1Segmentation

Solution Approach 2:

The patent changes the computational parameters from traditional quadratic-time offline processing to linear-time streaming processing. By using graph recovery models and sparse graph representations, the system transforms the processing complexity parameter, enabling real-time segmentation while maintaining measurement precision through adaptive graph updates rather than bulk reprocessing.

Inventive Principle:

Principle #35Parameter changes

2Manufacturing precision

If conventional systems process time series data offline, then manufacturing precision is improved, but loss of time worsens

Engineering Contradiction:

Improvesegmentation accuracyVSAvoidtime delay in segmentation

Core Design Contradiction:

Manufacturing precisionVSLoss of time

Solution Approach 1:

The patent implements preliminary action by pre-computing graph structures and recovery models that can be rapidly applied to incoming data streams. The system prepares graph recovery algorithms and similarity models in advance, allowing immediate processing of new time series data without waiting for offline batch processing cycles. This preliminary preparation enables real-time segmentation while maintaining the precision previously achievable only through offline methods.

Inventive Principle:

Principle #10Preliminary action

3Productivity

If real-time processing is implemented, then productivity is improved, but device complexity increases

Engineering Contradiction:

Improvereal-time processing capabilityVSAvoidcomputational complexity

Core Design Contradiction:

ProductivityVSDevice complexity

Solution Approach 1:

The patent extracts only the essential features and patterns from time series data by using graph recovery models that identify salient structures. Instead of processing entire datasets with full computational complexity, the system extracts key graph representations and similarity metrics, reducing the computational burden while maintaining real-time processing capability. This extraction approach simplifies the device requirements for real-time operation.

Inventive Principle:

Principle #2Taking out (Extraction)

4Measurement precision

If offline processing is used, then measurement precision is improved, but loss of information increases

Engineering Contradiction:

Improvepattern identification accuracyVSAvoiddata relevance over time

Core Design Contradiction:

Measurement precisionVSLoss of information

Solution Approach 1:

The patent implements continuous processing of time series data through streaming graph recovery operations. Instead of discontinuous offline batch processing that loses temporal context, the system continuously updates graph representations and detects patterns in real-time. This continuous action preserves the temporal relationships and relevance of data points, maintaining measurement precision while preventing information loss that occurs with delayed processing.

Inventive Principle:

Principle #20Continuity of useful action

Data Source

PatentUS20240354551A1Real-time segmentation of time series data using sparse graph recovery algorithms

Publication Date: 2024.10.24 MICROSOFT TECHNOLOGY LICENSING LLC

AI summary

This disclosure relates to a real-time segmentation system that utilizes graph objects and models to efficiently and accurately generate segmented real-time time series data. The real-time segmentation system achieves this by efficiently generating new current graph objects as data points are received using a graph recovery model. Additionally, the real-time segmentation system removes previously generated graph objects beyond the current graph object and the previous graph object to reduce the amount of stored data. These object graphs can include conditional independence (CI) graphs, which are probabilistic graphical models that include nodes connected by edges to exhibit partial correlations between the nodes. Furthermore, the time series segmentation system determines segmentation timestamps from the graph objects using a similarity model.