Topic-Specific Knowledge Graph Generation via AI Entity Filtering

Resolve Bottlenecks,
Find Innovative SolutionsGenerate Solutions

Solution Overview

Problem

Existing knowledge graph generation techniques produce extensive, unfocused graphs that make it difficult to identify relevant information and are inefficient to update when documents are added or removed, requiring the creation of new graphs from scratch.

Innovation Solution

Systems and methods using artificial intelligence models and user-provided guidance to generate topic-specific knowledge graphs, filtering out irrelevant entities and allowing for efficient updates by merging new graphs with existing ones.

Engineering Contradictions & Design Principles

VSEngineering Contradiction Analysis

1Quantity of substance

If existing knowledge graph generation techniques are used to represent all important information in documents, then comprehensive information representation is achieved, but the knowledge graph becomes extensive and difficult to navigate for topic-specific queries

Engineering Contradiction:

Improveamount of information representedVSAvoidease of identifying relevant information

Core Design Contradiction:

Quantity of substanceVSEase of operation

Solution Approach 1:

The patent extracts and filters out only the entities and relationships relevant to a specific topic from the complete document set. By using topic modeling algorithms, the system identifies and extracts only those knowledge graph elements that pertain to the user's specified topic, removing irrelevant information while preserving comprehensive coverage of the target topic.

Inventive Principle:

Principle #2Taking out (Extraction)

Solution Approach 2:

The patent segments the knowledge graph generation process into topic-specific subsets. Instead of creating a single monolithic knowledge graph containing all information, the system divides the document collection into topic-related groups and generates separate knowledge graphs for each topic, making the information more manageable and easier to navigate.

Inventive Principle:

Principle #1Segmentation

2Reliability

If a new knowledge graph is created from scratch when documents are added or removed, then the knowledge graph reflects current document set, but the process is time-consuming especially for large document sets

Engineering Contradiction:

Improveaccuracy of knowledge graph representationVSAvoidtime required for graph generation

Core Design Contradiction:

ReliabilityVSLoss of time

Solution Approach 1:

The patent merges the process of knowledge graph generation with the document indexing process. When new documents are added to the collection, the system combines the extraction of entities and relationships from these new documents with the existing knowledge graph structure, integrating them through merging operations rather than creating separate graphs and manually reconciling differences.

Inventive Principle:

Principle #5Merging (Combining)

Solution Approach 2:

The patent performs preliminary processing of documents by extracting entities, relationships, and topic assignments before the knowledge graph is fully constructed. This preliminary action allows the system to prepare and organize information in advance, so when documents are added or removed, only updates are needed rather than regenerating the entire graph from scratch.

Inventive Principle:

Principle #10Preliminary action

3Ease of operation

If topic-specific filtering is applied to create focused knowledge graphs, then ease of identifying relevant documents is improved, but the complexity of the generation process increases

Engineering Contradiction:

Improveease of identifying relevant documentsVSAvoidcomplexity of generation process

Core Design Contradiction:

Ease of operationVSDevice complexity

Solution Approach 1:

The patent implements a universal knowledge graph generation framework that can handle multiple topics simultaneously. The same core algorithms and processing steps are used for each topic, allowing the system to generate topic-specific knowledge graphs using a unified approach rather than requiring separate specialized processes for each topic.

Inventive Principle:

Principle #6Universality (Multi-functionality)

Solution Approach 2:

The patent introduces topic models as intermediary components that bridge the gap between raw documents and the final knowledge graph. These topic models act as mediators that first process and organize document information into topic-specific representations, which are then used to construct the final knowledge graph, simplifying the overall generation process.

Inventive Principle:

Principle #24Intermediary (Mediator)

Data Source

PatentUS20240394563A1Systems and methods for constructing topic-specific knowledge graphs

Publication Date: 2024.11.28 PWC PRODUCT SALES LLC

AI summary

Provided is a method for generating a knowledge graph for a topic. First, a set of documents may be received. A user may then provide an indication of exemplary entities associated with a topic of interest. Then, using one or more artificial intelligence models, a plurality of textual entities may be extracted from the documents. A quality level of each extracted textual entity may be determined. Each quality level may indicate a degree of similarity between a textual entity and each exemplary entity of the one or more exemplary entities. Next, a plurality of high-quality textual entities may be identified and categorized according to one or more sub-topics associated with the topic. Connection information indicating relationships between the sub-topics may be determined. Finally, a knowledge graph for the topic that represents the sub-topics in the documents and the relationships between said sub-topics may be generated.