US 11,874,763 B2
Unsupervised integration test builder
Austin Walters, Savoy, IL (US); Jeremy Goodsitt, Champaign, IL (US); Fardin Abdi Taghi Abad, Champaign, IL (US); Mark Watson, Urbana, IL (US); Anh Truong, Champaign, IL (US); and Reza Farivar, Champaign, IL (US)
Assigned to Capital One Services, LLC, McLean, VA (US)
Filed by Capital One Services, LLC, McLean, VA (US)
Filed on Nov. 4, 2022, as Appl. No. 18/052,844.
Application 18/052,844 is a continuation of application No. 17/026,461, filed on Sep. 21, 2020, granted, now 11,494,290.
Application 17/026,461 is a continuation of application No. 16/697,561, filed on Nov. 27, 2019, granted, now 10,783,064, issued on Sep. 22, 2020.
Prior Publication US 2023/0161689 A1, May 25, 2023
This patent is subject to a terminal disclaimer.
Int. Cl. G06F 9/44 (2018.01); G06F 9/445 (2018.01); G06F 9/455 (2018.01); G06F 11/36 (2006.01); G06N 3/086 (2023.01); G06N 3/044 (2023.01)
CPC G06F 11/3684 (2013.01) [G06N 3/044 (2023.01); G06N 3/086 (2013.01)] 20 Claims
OG exemplary drawing
 
1. A method of testing a target website, comprising:
identifying, by a computing system, scripting code associated with a plurality of websites hosted by a plurality of third party servers;
generating, by the computing system, a recurrent neural network configured to identify broken paths within a target website based on the scripting code, the generating comprising:
training the recurrent neural network to learn a first plurality of paths through the plurality of websites, wherein a genetic algorithm is used to inject randomness into the learning by adjusting variables of the recurrent neural network, wherein the variables define how the recurrent neural network simulates a user navigating the plurality of websites; and
deploying, by the computing system, the recurrent neural network to test the target website for errors.