Methods and apparatus for processing grammatical tags in a template to generate text
First Claim
Patent Images
1. A method for automatically generating text, the method comprising acts of:
- accessing human-language text automatically generated using at least one template that includes at least some fixed text and at least one tag that serves as a placeholder to be filled in with automatically generated text;
automatically identifying instances of at least one portion of the human-language text appearing multiple times in the human language text, wherein the automatically identifying is performed via at least one processor; and
automatically generating output text in a human-readable language at least in part by substituting one or more synonyms of the at least one portion for one or more of the identified instances of the at least one portion in the human-language text;
wherein the one or more synonyms comprises a first synonym;
wherein identifying instances of the at least one portion comprises identifying two instances of the at least one portion that appear in close proximity to each other in the human-language text;
wherein substituting the one or more synonyms comprises substituting the first synonym for one of the two identified instances of the at least one portion in the human-language text; and
wherein identifying the two instances of the at least one portion comprises identifying two instances that appear within a threshold number of characters or words of one another in the human-language text.
1 Assignment
0 Petitions
Accused Products
Abstract
Method and apparatus for automatically generating text in a human language using a template. The template may include at least one grammatical tag that implicates at least one actor in a sentence in the template. Human-language text may be determined to fill in the tag based on a characteristic of the actor, such as, for example, the actor'"'"'s gender, whether the actor is singular or plural, or some other characteristic of the actor.
96 Citations
18 Claims
-
1. A method for automatically generating text, the method comprising acts of:
-
accessing human-language text automatically generated using at least one template that includes at least some fixed text and at least one tag that serves as a placeholder to be filled in with automatically generated text; automatically identifying instances of at least one portion of the human-language text appearing multiple times in the human language text, wherein the automatically identifying is performed via at least one processor; and automatically generating output text in a human-readable language at least in part by substituting one or more synonyms of the at least one portion for one or more of the identified instances of the at least one portion in the human-language text; wherein the one or more synonyms comprises a first synonym; wherein identifying instances of the at least one portion comprises identifying two instances of the at least one portion that appear in close proximity to each other in the human-language text; wherein substituting the one or more synonyms comprises substituting the first synonym for one of the two identified instances of the at least one portion in the human-language text; and wherein identifying the two instances of the at least one portion comprises identifying two instances that appear within a threshold number of characters or words of one another in the human-language text. - View Dependent Claims (2, 3, 4, 5, 6)
-
-
7. At least one computer readable medium encoded with instructions that, when executed on a computer system, perform a method for automatically generating text, the method comprising acts of:
-
accessing human-language text automatically generated using at least one template that includes at least some fixed text and at least one tag that serves as a placeholder to be filled in with automatically generated text; identifying instances of at least one portion of the human-language text appearing multiple times in the human language text; and automatically generating output text in a human-readable language at least in part by substituting one or more synonyms of the at least one portion for one or more of the identified instances of the at least one portion in the human-language text; wherein the one or more synonyms comprises a first synonym; wherein identifying instances of the at least one portion comprises identifying two instances of the at least one portion that appear in close proximity to each other in the human-language text; wherein substituting the one or more synonyms comprises substituting the first synonym for one of the two identified instances of the at least one portion in the human-language text; and wherein identifying the two instances of the at least one portion comprises identifying two instances that appear within a threshold number of characters or words of one another in the human-language text. - View Dependent Claims (8, 9, 10, 11, 12)
-
-
13. A computer system comprising:
-
at least one memory; and at least one controller, coupled to the at least one memory, that; accesses, from the at least one memory, human-language text automatically generated using at least one template that includes at least some fixed text and at least one tag that serves as a placeholder to be filled in with automatically generated text; identifies instances of at least one portion of the human-language text appearing multiple times in the human language text; and automatically generates output text in a human-readable language at least in part by substituting one or more synonyms of the at least one portion for one or more of the identified instances of the at least one portion in the human-language text; wherein the one or more synonyms comprises a first synonym; wherein the at least one controller identifies instances of the at least one portion at least in part by identifying two instances of the at least one portion that appear in close proximity to each other in the human-language text; wherein substituting the one or more synonyms comprises substituting the first synonym for one of the two identified instances of the at least one portion in the human-language text; and wherein identifying the two instances of the at least one portion comprises identifying two instances that appear within a threshold number of characters or words of one another in the human-language text. - View Dependent Claims (14, 15, 16, 17, 18)
-
Specification