Incremental category embedding for categorization
AssigneeDisney Enterprises, Inc.
InventorsSung Ju Hwang, Leonid Sigal
Life of this patent
Estimating the term…
Abstract
There are provided systems and methods of incremental category embedding for categorization. One method including selecting one or more input categories from a plurality of input categories to be added to learned categories, determining at least one representative category from the learned categories for each input category from the one or more input categories, the at least one representative category representing the input category, and approximating the input category using the at least one representative category.
Claims18
A system comprising:
- a memory storing learned categories, each of the learned categories including learned data;
- and
- a processor configured to:
- select one or more input categories from a plurality of input categories to be added to the learned categories;
- determine at least one representative category from the learned categories for each input category from the one or more input categories, the at least one representative category representing the input category, wherein the at least one representative category includes a category that is similar to the input category based on the learned data;
- approximate the input category using the at least one representative category to create an approximate input category, and
- add the approximate input category to the learned categories.
A method for incrementally embedding categories for categorization, the method comprising:
- selecting one or more input categories from a plurality of input categories to be added to learned categories, each of the learned categories including learned data;
- determining at least one representative category from the learned categories for each input category from the one or more input categories, the at least one representative category representing the input category, wherein the at least one representative category includes a category that is similar to the input category based on the learned data;
- approximating the input category using the at least one representative category to create an approximate input category, and
- adding the approximate input category to the learned categories.
Dependencies are inferred from the claim text.