Entity Augmentation Service Using Relational Data Indexes

Resolve Bottlenecks,
Find Innovative SolutionsGenerate Solutions

Solution Overview

Problem

Existing methods for augmenting data in spreadsheets and databases are unsatisfactory due to low precision and poor coverage, requiring users to manually find and merge information from web sources, which is time-consuming and often results in incorrect data.

Innovation Solution

A method and system that process augmentation tasks based on direct and indirect relationships between entities and attributes mined from a corpus, using a holistic matching framework and personalized page rank to identify and aggregate data from multiple tables, ensuring high precision and coverage.

Engineering Contradictions & Design Principles

VSEngineering Contradiction Analysis

1Quantity of substance

If users manually find and merge information from web sources, then data coverage can be improved, but time consumption and error rate increase

Engineering Contradiction:

Improvedata coverageVSAvoidtime consumption

Core Design Contradiction:

Quantity of substanceVSLoss of time

Solution Approach 1:

The system automatically performs entity augmentation by querying pre-processed relational tables and aggregating results, eliminating the need for manual information gathering and merging by users

Inventive Principle:

Principle #25Self-service

Solution Approach 2:

Relational tables are pre-processed and stored in an optimized format with computed features before queries are executed, enabling fast automated augmentation without manual intervention

Inventive Principle:

Principle #10Preliminary action

2Productivity

If existing automated approaches are used, then time consumption is reduced, but data precision and recall deteriorate

Engineering Contradiction:

Improveautomation efficiencyVSAvoiddata precision

Core Design Contradiction:

ProductivityVSMeasurement precision

Solution Approach 1:

The system transforms relational tables by computing additional features and representations that capture semantic relationships, improving the precision of automated queries while maintaining efficiency

Inventive Principle:

Principle #35Parameter changes

Solution Approach 2:

Pre-processed relational tables serve as an intermediary structure between raw web data and query results, enabling automated systems to achieve high precision through optimized data representations

Inventive Principle:

Principle #24Intermediary (Mediator)

3Quantity of substance

If comprehensive data aggregation is performed, then data coverage is improved, but system complexity increases

Engineering Contradiction:

Improvedata coverageVSAvoidsystem complexity

Core Design Contradiction:

Quantity of substanceVSDevice complexity

Solution Approach 1:

Complex data processing and relationship extraction are performed in advance during the pre-processing stage, simplifying the query execution phase while maintaining comprehensive coverage

Inventive Principle:

Principle #10Preliminary action

Solution Approach 2:

The system divides data processing into separate stages: pre-processing of relational tables, query execution, and result aggregation, reducing overall system complexity through modular design

Inventive Principle:

Principle #1Segmentation

Data Source

PatentEP2823410B1Entity augmentation service from latent relational data

Publication Date: 2019.12.25 MICROSOFT TECHNOLOGY LICENSING LLC

Figure 1
Figure 2
Figure 3A~3C

AI summary

The subject disclosure is directed towards providing data for augmenting an entity-attribute-related task. Pre-processing is preformed on entity-attribute tables extracted from the web, e.g., to provide indexes that are accessible to find data that completes augmentation tasks. The indexes are based on both direct mappings and indirect mappings between tables. Example augmentation tasks include queries for augmented data based on an attribute name or examples, or finding synonyms for augmentation. An online query is efficiently processed by accessing the indexes to return augmented data related to the task.