Summarization of a Document

• US 20150293905A1
• Filed: 10/26/2012
• Published: 10/15/2015
• Est. Priority Date: 10/26/2012
• Status: Active Grant

• Current Assignee: Micro Focus LLC (Open Text Corporation)
• Original Assignee: Hewlett-Packard Development Company, L.P. (HP Inc.)
• Inventors: Wang, Lei, Wang, Xun, Zhang, Tong, Wang, Min, Bai, Bin
• Granted Patent: US 9,727,556 B2
• Time in Patent Office: Days
• US Class Current: 1/1
• CPC Class Codes: G06F 16/345 Summarisation for human users
• CPC Class Codes: G06F 16/353 into predefined classes
• CPC Class Codes: G06F 40/30 Semantic analysis
• CPC Class Codes: G06F 40/40 Processing or translation of natural language natural language analysis G06F40/20; semantic analysis G06F40/30
• CPC Class Codes: G06N 5/022 Knowledge engineering; Knowledge acquisition

Abstract

A method for summarizing a document is provided. A concept is detected for each sentence in said document. Relevance measures between the sentences are computed according to the detected concepts. And then a concept-aware graph is constructed, wherein a node in said graph represents a sentence in the document and an edge between two nodes represents a relevance measure between these two sentences.

Claims

1. A method for summarizing a document, comprising:

detecting, for each sentence in said document, a concept in said sentence;

according to the detected concepts, computing relevance measures between the sentences;

construct a concept-aware graph, wherein a node in said graph represents a sentence in the document and an edge between two nodes represents a relevance measure between these two sentences.

2. The method of claim 1, wherein detecting a concept in a sentence is performed based on a predefined concept library, which comprises a number of reference concepts.

3. The method of claim 2, wherein computing relevance measures between the sentences comprises:

computing said relevance measure according to reference concepts in said concept library corresponding to the detected concepts.

4. The method of claim 2, further comprising:

assigning a respective weight to a detected concept.

5. The method of claim 4, wherein said weight represents a degree of similarity between the detected concept and its corresponding reference concept in the concept library.

6. The method of claim 4, wherein computing a relevance measure between sentences comprises:

computing said relevance measure as a function of weights of concepts detected in the sentences and relevance measures between reference concepts in said concept library corresponding to the detected concepts.

7. A system for summarizing a document, the system comprising:

a processor that is adapted to execute stored instructions; and

a memory device that stores instructions, the memory device comprising processor-executable code, that when executed by the processor, is adapted to;

detect, for each sentence in said document, a concept in said sentence;

according to the detected concepts, compute relevance measures between the sentences;

construct a concept-aware graph, wherein a node in said graph represents a sentence in the document and an edge between two nodes represents a relevance measure between these two sentences.

8. The system of claim 7, wherein the memory stores processor-executable code adapted to detect a concept in a sentence based on a predefined concept library, which comprises a number of reference concepts.

9. The system of claim 8, wherein the memory stores processor-executable code adapted to compute relevance measures between the sentences by:

computing said relevance measure according to reference concepts in said concept library corresponding to the detected concepts.

10. The system of claim 8, wherein the memory stores processor-executable code adapted to assign a respective weight to a detected concept.

11. The system of claim 10, wherein said weight represents a degree of similarity between the detected concept and its corresponding reference concept in the concept library.

12. The system of claim 10, wherein the memory stores processor-executable code adapted to compute a relevance measure between sentences by:

computing said relevance measure as a function of weights of concepts detected in the sentences and relevance measures between reference concepts in said concept library corresponding to the detected concepts.

13. A non-transitory, computer-readable medium, comprising code to direct a processor to:

detect, for each sentence in a document which comprises a plurality of sentences, a concept in said sentence;

according to the detected concepts, compute relevance measures between the sentences;

based on said relevance measures, generate a summary for said document; and

construct a concept-aware graph, wherein a node in said graph represents a sentence in the document and an edge between two nodes represents a relevance measure between these two sentences.

14. The non-transitory, computer-readable medium of claim 13, wherein the non-transitory, computer-readable medium comprises code to detect a concept in a sentence based on a predefined concept library, which comprises a number of reference concepts.

15. The non-transitory, computer-readable medium of claim 14, wherein the non-transitory, computer-readable medium comprises code to assign a respective weight to a detected concept, wherein said weight represents a degree of similarity between the detected concept and its corresponding reference concept in the concept library.