Document Summarization via Term Scoring and Combinatorial Optimization
Find Innovative SolutionsGenerate Solutions
Solution Overview
Problem
Users face challenges in efficiently finding relevant content among numerous search results, particularly on mobile devices, due to the need to sift through lengthy documents and the difficulty in determining the interest value of electronic documents from URLs or titles alone.
Innovation Solution
A method for summarizing electronic documents by splitting them into terms with informativeness and coherence scores, using a combinatorial optimizer to select a subset of terms that maximizes the aggregate informativeness score while keeping the summary length manageable, and arranging them in a concise summary.
Engineering Contradictions & Design Principles
Engineering Contradiction Analysis
1Measurement precision
If users review documents one-by-one to find relevant content, then they can evaluate document quality, but the time required increases significantly
Solution Approach 1:
The system generates summaries of search result documents before the user views them. This preliminary action provides users with condensed information about document contents, enabling them to assess relevance quickly without reading entire documents, thus resolving the contradiction between evaluation accuracy and time consumption
Solution Approach 2:
The summary acts as an intermediary between the full document and the user. Instead of directly reviewing complete documents, users interact with condensed summaries that capture essential information, facilitating efficient relevance assessment while reducing time investment
2Loss of information
If documents are displayed in full to provide complete information, then information completeness is maintained, but the complexity of information processing increases
Solution Approach 1:
The system segments information by extracting and presenting only the most relevant portions of documents in summary form. Instead of displaying complete documents, it divides content into essential segments that capture key information, reducing processing complexity while maintaining information completeness
Solution Approach 2:
The system extracts essential information from full documents to create summaries. By taking out only the most relevant content elements and presenting them in condensed form, it maintains information completeness while significantly reducing the complexity of information processing for users
3Productivity
If summaries are made more concise to save time, then reading speed increases, but the amount of information retained decreases
Solution Approach 1:
The system applies local quality by concentrating information density in specific areas of the summary. Instead of uniformly distributing content, it places higher informational value in key segments that capture essential document elements, enabling concise presentation while retaining critical information
Data Source
Figure 1
Figure 2
Figure 3
AI summary
One embodiment of a method for summarizing an electronic document includes splitting the electronic document into a plurality of terms, wherein each of the plurality of terms is associated with a respective length, a respective informativeness score, and a respective coherence score, automatically selecting a subset of the plurality of terms, such that an aggregate informativeness score of the subset is maximized while an aggregate length of the subset is less than or equal to a maximum length, and arranging the subset as a summary of the electronic document.