US 12,462,103 B2
Representing confidence in natural language processing
Jeremy Goodsitt, Champaign, IL (US); Austin Walters, Savoy, IL (US); and Anh Truong, Champaign, IL (US)
Assigned to Capital One Services, LLC, McLean, VA (US)
Filed by Capital One Services, LLC, McLean, VA (US)
Filed on Jun. 20, 2023, as Appl. No. 18/211,874.
Application 18/211,874 is a continuation of application No. 17/961,915, filed on Oct. 7, 2022, granted, now 11,720,753.
Application 17/961,915 is a continuation of application No. 17/004,728, filed on Aug. 27, 2020, granted, now 11,501,074, issued on Nov. 15, 2022.
Prior Publication US 2023/0359827 A1, Nov. 9, 2023
This patent is subject to a terminal disclaimer.
Int. Cl. G06F 40/295 (2020.01)
CPC G06F 40/295 (2020.01) 20 Claims
OG exemplary drawing
 
1. A method comprising:
processing, using a natural language processing algorithm, a plurality of characters of a word of text content by associating different characters of the plurality of characters of the word of the text content with a corresponding one or more of a plurality of categories;
for each character of the plurality of characters of the text content:
determining, based on the processing, a plurality of confidence values that indicate, for the character, a degree of confidence corresponding to one or more of the plurality of categories, wherein at least two of the plurality of confidence values are different; and
determining, for the character and based on the plurality of confidence values, a corresponding display indication that indicates the degree of confidence corresponding to one or more of the plurality of categories when processing, using the natural language processing algorithm, the character, wherein at least two characters of the word correspond to different display indications; and
causing display, by a computing device, of a user interface comprising a visualization of the text content, wherein the user interface is configured to display the corresponding display indication of each character of the plurality of characters of the word of the text content such that the user interface indicates a change in confidence when processing, using the natural language processing algorithm, different characters of the plurality of characters of the word of the text content.