US 11,863,502 B2
Systems and methods for a two-tier machine learning model for generating conversational responses
Kunlaya Soiaporn, Vienna, VA (US); Victor Alvarez Miranda, McLean, VA (US); Pamela Katali, Gaithersburg, MD (US); Arturo Hernandez Zeledon, Arlington, VA (US); Rui Zhang, New York, NY (US); and Kwan-Yuet Ho, Silver Spring, MD (US)
Assigned to Capital One Services, LLC, McLean, VA (US)
Filed by Capital One Services, LLC, McLean, VA (US)
Filed on Feb. 7, 2023, as Appl. No. 18/165,917.
Application 18/165,917 is a continuation of application No. 17/738,224, filed on May 6, 2022, granted, now 11,616,741.
Application 17/738,224 is a continuation of application No. 16/908,116, filed on Jun. 22, 2020, granted, now 11,356,389, issued on Jun. 7, 2022.
Prior Publication US 2023/0188483 A1, Jun. 15, 2023
This patent is subject to a terminal disclaimer.
Int. Cl. H04L 51/02 (2022.01); G10L 15/16 (2006.01); G06N 20/20 (2019.01); G06F 18/23 (2023.01); G06V 10/762 (2022.01); G06V 10/80 (2022.01); G06V 10/82 (2022.01)
CPC H04L 51/02 (2013.01) [G06F 18/23 (2023.01); G06N 20/20 (2019.01); G10L 15/16 (2013.01); G06V 10/763 (2022.01); G06V 10/809 (2022.01); G06V 10/82 (2022.01)] 20 Claims
OG exemplary drawing
 
1. A system for generating responses using two-tier machine learning models, the system comprising:
cloud-based storage circuitry configured to:
store a first machine learning model, wherein the first machine learning model is trained to select an intent cluster from a plurality of intent clusters based on user actions, and wherein each intent cluster of the plurality of intent clusters corresponds to a respective intent of a user following a first user action; and
store a second machine learning model, wherein the second machine learning model is trained to select a specific intent from a plurality of specific intents of a selected intent cluster based, and wherein each specific intent of the plurality of specific intents corresponds to a respective specific intent of the user following the first user action;
cloud-based control circuitry configured to:
receive the first user action at a user interface;
select the second machine learning model, from a plurality of machine learning models, based on the intent cluster selected from the plurality of intent clusters, wherein each intent cluster of the plurality of intent clusters corresponds to a respective machine learning model from the plurality of machine learning models;
select, using the second machine learning model, a response from a plurality of responses; and
cloud-based input/output circuitry configured to:
generate the response.