Indirect Concept Inference Using Linked Documents and User Patterns
Find Innovative SolutionsGenerate Solutions
Solution Overview
Problem
Existing concept inference techniques for documents, such as web pages, are limited to inferring concepts directly from textual content, domain, and URL, failing to identify other relevant concepts that are not directly inferable from these characteristics.
Innovation Solution
The method involves indirectly inferring concepts by retrieving and analyzing associated URLs, linked documents, user retrieval patterns, and successful advertisements to label concepts as useful to the document's audience, expanding the scope beyond traditional characteristics.
Engineering Contradictions & Design Principles
Engineering Contradiction Analysis
1Measurement precision
If concepts are inferred solely from document characteristics (textual content, domain, URL), then the inference process is simple and direct, but relevant concepts that are not directly inferable from these characteristics are missed
Solution Approach 1:
The patent transitions from inferring concepts solely from document characteristics (text, domain, URL) to incorporating external dimensions such as linked documents, user retrieval patterns, and advertisement data. This multi-dimensional approach enables discovery of concepts not directly inferable from the document itself, thereby improving concept inference accuracy while accepting increased process complexity
Solution Approach 2:
The patent introduces intermediary elements (linked documents, user behavior data, advertisement content) that mediate between the target document and the concepts to be inferred. These intermediaries provide indirect evidence about relevant concepts, allowing the system to discover concepts that would not be apparent from direct analysis of the document characteristics alone
2Adaptability or versatility
If only direct document characteristics are used for concept inference, then the system is easy to implement, but the relevance of associated content for broader audience is limited
Solution Approach 1:
The patent makes the concept inference system multi-functional by incorporating multiple data sources (document characteristics, linked documents, user behavior, advertisements) that serve different functions. This universality allows the system to adapt to diverse document types and audience preferences, thereby expanding audience coverage while managing processing complexity through modular architecture
3Measurement precision
If external information sources (URLs, linked documents, user patterns, advertisements) are incorporated for concept inference, then more relevant concepts are identified, but the processing complexity increases
Solution Approach 1:
The patent segments the concept inference process into distinct modules: extracting concepts from document characteristics, extracting concepts from linked documents, analyzing user retrieval patterns, and processing advertisement data. Each segment handles a specific data source independently, improving concept relevance while managing system complexity through modular, manageable components that can be implemented and maintained separately
Data Source
AI summary
A system indirectly infers concepts associated with a document. The concepts may be indirectly inferred based on information that does not include characteristics of the document, such as the characteristics that include a textual content of the document not associated with links included in the document, a domain of the document, and the document's Uniform Resource Locator (URL). The system may label the inferred concepts as useful to an audience of the document.