Co-Relation Graphs for Preference-Based Content Summaries

Resolve Bottlenecks,
Find Innovative SolutionsGenerate Solutions

Solution Overview

Problem

Existing content summaries are often generic and fail to reflect user preferences, leading to misrepresentation of content characteristics and reduced user engagement.

Innovation Solution

A system and method that generates personalized summaries based on user preferences and a co-relation graph, identifying relevant segments of content using characteristics like characters, actions, and locations, and adjusting summary length based on priority rankings.

Engineering Contradictions & Design Principles

VSEngineering Contradiction Analysis

1Ease of operation

If generic template-based summaries are used, then the summary generation process is simple and fast, but the summary does not reflect user preferences and misrepresents content characteristics

Engineering Contradiction:

Improvesummary generation processVSAvoiduser preference information

Core Design Contradiction:

Ease of operationVSLoss of information

Solution Approach 1:

The system pre-generates a co-relation graph that maps relationships between content segments, characters, and attributes before user interaction. This preliminary structuring enables rapid personalized summary generation without complex real-time processing, as the graph already contains pre-computed relationships that can be quickly queried based on user preferences

Inventive Principle:

Principle #10Preliminary action

Solution Approach 2:

The co-relation graph serves as an intermediary data structure that bridges generic content metadata and user preferences. Instead of directly matching user preferences against raw content data, the system uses the pre-built co-relation graph to efficiently retrieve and aggregate relevant segment information, thereby preserving user preference information while maintaining operational simplicity

Inventive Principle:

Principle #24Intermediary (Mediator)

2Reliability

If personalized summaries based on user preferences are generated, then user engagement and content selection accuracy improve, but the system complexity increases due to co-relation graph construction and segment identification

Engineering Contradiction:

Improvecontent recommendation accuracyVSAvoidsystem structure

Core Design Contradiction:

ReliabilityVSDevice complexity

Solution Approach 1:

The system segments content into discrete units with defined attributes (characters, locations, actions) and builds the co-relation graph by mapping relationships between these segmented elements. This segmentation approach transforms the complex problem of personalized summarization into manageable components: segment identification, relationship mapping, and preference-based filtering, thereby reducing overall system complexity while maintaining recommendation accuracy

Inventive Principle:

Principle #1Segmentation

3Loss of information

If all identified segments are included in the summary, then the summary is comprehensive, but the summary length exceeds user attention thresholds and reduces readability

Engineering Contradiction:

Improvecontent detail informationVSAvoidsummary length

Core Design Contradiction:

Loss of informationVSLength of moving object

Solution Approach 1:

The system applies local quality by differentiating segment importance within the summary based on user preferences and segment characteristics. Instead of treating all segments uniformly, the system identifies and prioritizes high-value segments (e.g., those containing preferred characters or plot-critical actions) while omitting or condensing less important segments, thereby maintaining information quality while controlling summary length to fit user attention thresholds

Inventive Principle:

Principle #3Local quality

Data Source

PatentUS20250330674A1Methods and systems for generating a summary of content based on a co-relation graph

Publication Date: 2025.10.23 ADEIA GUIDES INC

AI summary

Provided are systems and methods for generating a customized summary of content for a user using a co-relation graph and a user preference. The co-relation graph maps, for each segment of the content, summary information of the segment and a characteristic of the segment. The system identifies, using the co-relation graph, a subset of the segments of the content based on the user preference and the characteristics of the segments. The customized summary of the content is generated based on the summary information for the identified subset of segments. The system may also use a current presentation time of the content to limit which segments are used to generate the summary. The system can also use a priority ranking to exclude summary information from less important segments to keep the summary within a desired size.