Web Content Highlighting Service for Context-Rich Search
Find Innovative SolutionsGenerate Solutions
Solution Overview
Problem
Current methods for information retrieval on the web, such as bookmarks, lack context and persistence, are inefficient for sharing and often require manual effort from recipients, and existing search engines rely heavily on metadata manipulation and author-driven relevance, which does not account for user perspectives.
Innovation Solution
A highlighting service that allows users to select and highlight content within web documents, storing these highlights for easy retrieval and sharing, and uses user-generated highlights to determine document relevance, thereby improving search results by incorporating end-user feedback.
Engineering Contradictions & Design Principles
Engineering Contradiction Analysis
1Loss of information
If bookmarks are used for information retrieval, then users can save and retrieve URLs, but bookmarks lack context and require manual effort from recipients to understand what the bookmark is about
Solution Approach 1:
The patent segments the web page content by extracting and highlighting specific portions (text, images, videos) that are relevant to the bookmark. Instead of bookmarking the entire page, the system segments the content into manageable, context-rich portions that can be independently reviewed and shared.
Solution Approach 2:
The patent introduces an intermediary service between the user and the bookmarked content. This service automatically analyzes web pages, extracts relevant portions, and presents them to users with contextual information. The intermediary handles the manual effort of content selection and presentation, eliminating the need for users to manually review entire pages.
2Measurement precision
If traditional search engines are used, then users can search for documents, but search results rely on metadata manipulation and author-driven relevance rather than user perspectives
Solution Approach 1:
The patent inverts the traditional search model by shifting the focus from author-driven metadata to user-driven content selection. Instead of relying on authors to provide relevant metadata, the system empowers users to select and highlight the portions of content they find relevant, thereby inverting the source of relevance determination from author to user.
Solution Approach 2:
The patent implements feedback mechanisms where user selections and highlightings are used to refine and improve search results. The system learns from user interactions with highlighted content and uses this feedback to enhance the accuracy of future search results, creating a continuous improvement loop that adapts to user perspectives.
3Productivity
If bookmarks are shared with other users, then information can be transmitted, but the process is inefficient and often requires manual effort from recipients to locate and understand the content
Solution Approach 1:
The patent performs preliminary actions by automatically analyzing web pages and pre-selecting relevant content portions before sharing. Instead of requiring recipients to manually review entire pages, the system has already done the work of identifying and organizing relevant content, saving recipients significant time and effort.
Solution Approach 2:
The patent creates enriched copies of web page portions that include contextual information, highlights, and annotations. These copies are designed to be self-contained and easily reviewable, allowing recipients to quickly understand the bookmarked content without needing to access the original full page or spend time searching for relevant information.
Data Source
AI summary
A method, system, and program for searching for a portion content contained in a document is disclosed. Accordingly, a search engine, or some other application, analyzes a document with a user selected content portion to determine a match. The particular algorithm for determining the match will vary depending upon the particular implementation. Based on this analysis, one or more portions of the document are automatically highlighted in situ.