Back to results
US 2018/0300306 A1Published application
Techniques for automatic generation of natural language text
AssigneeYseop SA
InventorsAlain Kaeser, Emmanuel Vignon, Ludan Stoeckle
Life of this patent
Estimating the term…
Abstract
Techniques for creating a template to be used in connection with automatically generating text. Techniques include creating a template to include human language text and at least a first tag that serves as a placeholder for a text portion referring to at least one referent; and allowing a user to specify multiple options to be used in place of the first tag when generating output text using the created template, the options comprising at least a first referential expression for the at least one referent and at least a first anaphoric expression for the at least one referent.
Claims21
1-15. (canceled)
A method for use in connection with automatic text generation, the method comprising:
- using at least one processor to perform:
- obtaining a template to include human language text and at least a first tag that serves as a first placeholder for a first text portion referring to a first referent, and a second tag that serves as a second placeholder for a second text portion referring to the first referent;
- obtaining, from a user, multiple language options to be used in place of the first tag and the second tag when generating output text using the obtained template, the multiple language options comprising at least a first referential expression for the first referent and a plurality of anaphoric expressions for the first referent including a first anaphoric expression for the first referent and a second anaphoric expression for the first referent;
- obtaining, from the user, a second referential expression for referring to a second referent, wherein the first referent and the second referent are of a same type;
- storing, in a first data structure including a first set of fields associated to the first referent, information about the first referent and a first link to a second data structure;
- storing, in the second data structure including a second set of fields associated to the first referent and the second referent, the multiple language options to use when referring to the first referent and the second referent, including the first referential expression, the second referential expression and the plurality of anaphoric expressions;
- storing, in a third data structure including a third set of fields associated to the second referent, information about the second referent and a second link to the second data structure;
- providing the obtained template, at least a portion of the information about the first referent stored in the first data structure, at least a portion of the information about the second referent stored in the third data structure and the multiple language options stored in the second data structure to a natural language generation system for use in generating the output text by using the obtained template, the natural language generation system programmed to identify which of the multiple language options to use for referring to the first referent and which of the multiple language options to use for referring to the second referent in the output text by accessing one or more fields of the second set of fields in the second data structure;
- determining to use the first referential expression in place of the first tag;
- selecting one of the first and second anaphoric expressions to use in place of the second tag by using the accessed one or more fields of the second set of fields;
- generating, using the natural language generation system and based on the template, the output text to include the first referential expression in place of the first tag and the selected one anaphoric expression in place of the second tag;
- and
- presenting, via a device, the output text to the user.
10 more elements
A system for use in connection with automatic text generation, the system comprising:
- at least one processor configured to perform:
- obtaining a template to include human language text and at least a first tag that serves as a first placeholder for a first text portion referring to a first referent, and a second tag that serves as a second placeholder for a second text portion referring to the first referent;
- obtaining, from a user, multiple language options to be used in place of the first tag and the second tag when generating output text using the obtained template, the multiple language options comprising at least a first referential expression for the first referent and a plurality of anaphoric expressions for the first referent including a first anaphoric expression for the first referent and a second anaphoric expression for the first referent;
- obtaining, from the user, a second referential expression for referring to a second referent, wherein the first referent and the second referent are of a same type;
- storing, in a first data structure including a first set of fields associated to the first referent, information about the first referent and a first link to a second data structure;
- storing, in the second data structure including a second set of fields associated to the first referent and the second referent, the multiple language options to use when referring to the first referent and the second referent, including the first referential expression, the second referential expression and the plurality of anaphoric expressions;
- storing, in a third data structure including a third set of fields associated to the second referent, information about the second referent and a second link to the second data structure;
- providing the obtained template, at least a portion of the information about the first referent stored in the first data structure, at least a portion of the information about the second referent stored in the third data structure and the multiple language options stored in the second data structure to a natural language generation system for use in generating the output text by using the obtained template, the natural language generation system programmed to identify which of the multiple language options to use for referring to the first referent and which of the multiple language options to use for referring to the second referent in the output text by accessing one or more fields of the second set of fields in the second data structure;
- determining to use the first referential expression in place of the first tag;
- selecting one of the first and second anaphoric expressions to use in place of the second tag by using the accessed one or more fields of the second set of fields;
- generating, using the natural language generation system and based on the template, the output text to include the first referential expression in place of the first tag and the selected one anaphoric expression in place of the second tag;
- and
- presenting, via a device, the output text to the user.
10 more elements
At least one non-transitory computer-readable storage medium storing processor-executable instructions that, when executed by at least one processor, cause the at least one processor to perform a method for use in connection with automatic speech generation, the method comprising:
- obtaining a template to include human language text and at least a first tag that serves as a first placeholder for a first text portion referring to a first referent, and a second tag that serves as a second placeholder for a second text portion referring to a the first referent;
- obtaining, from a user, multiple language options to be used in place of the first tag and the second tag when generating output text using the obtained template, the multiple language options comprising at least a first referential expression for the first referent and a plurality of anaphoric expressions for the first referent including a first anaphoric expression for the first referent and a second anaphoric expression for the first referent;
- obtaining, from the user, a second referential expression for referring to a second referent, wherein the first referent and the second referent are of a same type;
- storing, in a first data structure including a first set of fields associated to the first referent, information about the first referent and a first link to a second data structure;
- storing, in the second data structure including a second set of fields associated to the first referent and the second referent, the multiple language options to use when referring to the first referent and the second referent, including the first referential expression, the second referential expression and the plurality of anaphoric expressions;
- storing, in a third data structure including a third set of fields associated to the second referent, information about the second referent and a second link to the second data structure;
- providing the obtained template, at least a portion of the information about the first referent stored in the first data structure, at least a portion of the information about the second referent stored in the third data structure and the multiple language options stored in the second data structure to a natural language generation system for use in generating the output text by using the obtained template, the natural language generation system programmed to identify which of the multiple language options to use for referring to the first referent and which of the multiple language options to use for referring to the second referent in the output text by accessing a field of the second set of fields in the second data structure;
- determining to use the first referential expression in place of the first tag;
- selecting one of the first and second anaphoric expressions to use in place of the second tag by using the accessed one or more fields of the second set of fields;
- generating, using the natural language generation system and based on the template, the output text to include the first referential expression in place of the first tag and the selected one anaphoric expression in place of the second tag;
- and
- presenting, via a device, the output text to the user.
9 more elements
The at least one non-transitory computer-readable storage medium of claim 32, wherein the second tag precedes the first tag in the template.
The at least one non-transitory computer-readable storage medium of claim 32, where the method further comprises:
- obtaining, from the user, a third referential expression for referring to a third referent, wherein the first referent and the third referent are of the same type;
- and
- storing, in a fourth data structure including a fourth set of fields associated to the third referent, information about the third referent and a third link to the second data structure.
The at least one non-transitory computer-readable storage medium of claim 32, wherein selecting one of the first and second anaphoric expressions to use in place of the second tag comprises selecting one of the first and second anaphoric expressions based on the first referential expression and text preceding the first tag in the template.
Dependencies are inferred from the claim text.