Trusted Source Identification System for Search Reliability

Resolve Bottlenecks,
Find Innovative SolutionsGenerate Solutions

Solution Overview

Problem

Current search engines prioritize paid advertisements and clickstream data, leading to biased results that can promote misinformation and 'clickbait,' making it difficult to find reliable sources of knowledge online.

Innovation Solution

A system that scores URLs based on their usage in documents and user interactions, accumulating a 'Trusted Web' corpus of vetted sources, where URLs with high scores are considered trusted and accessible for research and search engines.

Engineering Contradictions & Design Principles

VSEngineering Contradiction Analysis

1Productivity

If search engines use clickstream data and popular content ranking, then user engagement and visibility are improved, but reliability and accuracy of information deteriorate

Engineering Contradiction:

Improveuser engagementVSAvoidinformation reliability

Core Design Contradiction:

ProductivityVSReliability

Solution Approach 1:

The patent inverts the traditional search ranking approach by not ranking based on popularity or clicks, but by ranking based on trust scores derived from expert endorsements and document quality metrics. This inversion allows the system to prioritize reliability over engagement while still providing useful search results.

Inventive Principle:

Principle #13The other way round (Inversion)

Solution Approach 2:

The patent introduces trust scores as an intermediary metric that mediates between document quality and search ranking. Instead of directly ranking by popularity, the system uses trust scores calculated from multiple factors including expert endorsements, citation patterns, and document attributes to determine search result ordering.

Inventive Principle:

Principle #24Intermediary (Mediator)

2Productivity

If search engines prioritize paid advertisements and popular content, then revenue and user engagement are improved, but information accuracy and trustworthiness deteriorate

Engineering Contradiction:

Improverevenue generationVSAvoidcontent trustworthiness

Core Design Contradiction:

ProductivityVSReliability

Solution Approach 1:

The patent segments search results into different categories including trusted sources, paid advertisements, and general results. By separating trusted content from commercial content, the system can maintain the revenue model while ensuring that trustworthy information is not compromised by paid promotions.

Inventive Principle:

Principle #1Segmentation

Solution Approach 2:

The trust score system acts as an intermediary layer that evaluates and validates content quality independent of payment status. Even paid advertisements can be evaluated for trustworthiness, and the system can distinguish between sponsored content and organic trusted content.

Inventive Principle:

Principle #24Intermediary (Mediator)

3Adaptability or versatility

If search results reflect user behavior and popularity, then relevance to user interests is improved, but bias and misinformation propagation worsen

Engineering Contradiction:

Improveuser relevanceVSAvoidbias reduction

Core Design Contradiction:

Adaptability or versatilityVSReliability

Solution Approach 1:

The patent applies local quality by allowing different ranking criteria for different types of queries and users. Trust scores are calculated based on local document attributes, expert endorsements in specific domains, and citation patterns relevant to the query topic, rather than applying a single global popularity metric.

Inventive Principle:

Principle #3Local quality

4Productivity

If clickstream data drives ranking, then visibility and engagement are improved, but clickbait promotion and information quality deteriorate

Engineering Contradiction:

Improvecontent visibilityVSAvoidcontent quality

Core Design Contradiction:

ProductivityVSReliability

Solution Approach 1:

The patent inverts the visibility mechanism by making trust scores the primary driver of ranking rather than clickstream data. High-quality content with expert endorsements receives visibility regardless of initial click counts, while clickbait content with low trust scores remains invisible even if it attracts clicks.

Inventive Principle:

Principle #13The other way round (Inversion)

Data Source

PatentUS11023553B2Identifying and managing trusted sources in online and networked content for professional knowledge exchange

Publication Date: 2021.06.01 MICROSOFT TECHNOLOGY LICENSING LLC

AI summary

Systems and methods are provided for identifying and managing trusted sources in online and networked content. URLs and other objects are vetted and identified as trusted sources based on the documents in which these objects are used or otherwise acted upon and not on the content of the URLs or other objects themselves. When a URL or other object is used in a document, the URL is given a score based on the use activity, attributes of the document itself, and, optionally, user history of the user. As users use that same URL with respect to their documents, the scores from those activities are accrued to that URL. A URL with an accrued score (value) above (or equal to) a threshold value can be added to a corpus of trusted sources that can be accessed when conducting research and/or by other applications and search engines.