US 7,925,496 B1Granted

Method for summarizing natural language text

AssigneeThe United States of America as represented by the Secretary of the Navy

InventorStuart Harvey Rubin

Life of this patent

Estimating the term…

Loading patent…

Abstract

A method includes the steps of comparing a first body of text with a user-created summary of the first body of text, creating rules based on the comparison of the first body of text with the user-created summary of the first body of text, selecting one or more summary rules for generating a computer-created summary of a second body of text, and applying the selected summary rules to the second body of text to generate a computer-created summary of the second body of text. The first body of text may be a user-corrected summary of a computer-created summary of the first body of text. The rules may be selected based on previous use, frequency of use, context of the body of text, or most-specific applicability. The rules may be iteratively applied to generate a summary. A method is also provided for generating a heading for a summary of text.

Claims6

Dependent claims

A computer-implemented method comprising the steps of:

eliminating, from a portion of a body of text, the most-frequently occurring words that do not relate to the specific context of the portion of the body of text;
searching a data storage having a plurality of words and at least one summary rule associated with each word stored therein, for the remaining words within the portion of the body of text;
selecting, from the data storage, one of the summary rules associated with each remaining word found in the data storage, wherein
if more than one summary rule is associated with one of the remaining words, a validated summary rule is selected over a non-validated summary rule,
if more than one validated summary rule is associated with one of the remaining words, the validated summary rule that covers the most-specific context of the remaining word is selected, and
if equally-specific validated summary rules are associated with one of the remaining words, the most-frequently-used validated summary rule that covers the most-specific context of the remaining word is selected;
creating a computer-generated summary by applying the selected summary rules to the body of text;
determining that a user has made user corrections to the computer-generated summary;
creating one or more new summary rules based upon the user corrections by
modifying one or more existing associations of a word and a summary rule stored in the data storage to reflect the user corrections if the user added one or more words to the computer-generated summary that are stored in the data storage, removed one or more words from the computer-generated summary, or performed spelling/grammar corrections to the computer-generated summary, and
creating a new association of a word and a summary rule if the user added one or more words to the computer-generated summary that are not stored in the data storage, the new association associating the non-stored word with a word sequence from a portion of the computer-generated summary;
and
storing the one or more new summary rules in the data storage.

10 more elements

A non-transitory computer-readable medium having a method stored thereon, the method represented by computer readable programming code, the method comprising the steps of:

eliminating, from a portion of a body of text, the most-frequently occurring words that do not relate to the specific context of the portion of the body of text;
searching a data storage having a plurality of words and at least one summary rule associated with each word stored therein, for the remaining words within the portion of the body of text;
selecting, from the data storage, one of the summary rules associated with each remaining word, wherein
if more than one summary rule is associated with one of the remaining words, a validated summary rule is selected over a non-validated summary rule,
if more than one validated summary rule is associated with one of the remaining words, the validated summary rule that covers the most-specific context of the remaining word is selected, and
if equally-specific validated summary rules are associated with one of the remaining words, the most-frequently-used validated summary rule that covers the most-specific context of the remaining word is selected;
creating a computer-generated summary by applying the selected summary rules to the body of text;
determining that a user has corrected the computer-generated summary;
creating one or more new summary rules based upon the user corrections by,
modifying one or more existing associations of a word and a summary rule stored in the data storage to reflect the user corrections if the user added one or more words to the computer-generated summary that are stored in the data storage, removed one or more words from the computer-generated summary, or performed spelling/grammar corrections to the computer-generated summary, and
creating a new association of a word and a summary rule if the user added one or more words to the computer-generated summary that are not stored in the data storage, the new association associating the non-stored word with a word sequence from a portion of the computer-generated summary;
and
storing the one or more new summary rules in the data storage.

10 more elements

Dependencies are inferred from the claim text.