US 11,816,696 B2
Machine-learning based multi-step engagement strategy modification
Pankhri Singhai, Rohini (IN); Sundeep Parsa, San Jose, CA (US); Piyush Gupta, Noida (IN); Nupur Kumari, Noida (IN); Nikaash Puri, New Delhi (IN); Mayank Singh, Noida (IN); Eshita Shah, San Francisco, CA (US); Balaji Krishnamurthy, Noida (IN); and Akash Rupela, Rohini (IN)
Assigned to Adobe Inc., San Jose, CA (US)
Filed by Adobe Inc., San Jose, CA (US)
Filed on Jun. 23, 2021, as Appl. No. 17/355,907.
Application 17/355,907 is a continuation of application No. 16/057,743, filed on Aug. 7, 2018, granted, now 11,107,115.
Prior Publication US 2021/0319473 A1, Oct. 14, 2021
This patent is subject to a terminal disclaimer.
Int. Cl. G06Q 30/00 (2023.01); G06Q 30/0242 (2023.01); G06Q 30/0251 (2023.01); G06N 20/00 (2019.01); G06N 5/00 (2023.01); G05B 19/418 (2006.01)
CPC G06Q 30/0244 (2013.01) [G06N 5/00 (2013.01); G06N 20/00 (2019.01); G06Q 30/0242 (2013.01); G06Q 30/0254 (2013.01); G06Q 30/0255 (2013.01); G06Q 30/0264 (2013.01)] 19 Claims
OG exemplary drawing
 
1. A method comprising:
receiving, via a user interface and by at least one computing device, user input to define aspects of a user-created multi-step engagement strategy for controlling delivery of content associated with a campaign, the defined aspects including an entry condition, an exit condition, and at least two content deliveries to client device users targeted by the campaign based on the entry condition;
generating, by the at least one computing device, one or more randomly varied multi-step engagement strategies that differ from the user-created multi-step engagement strategy;
training, using data describing interactions with the content delivered according to both the user-created multi-step engagement strategy and the one or more randomly varied multi-step engagement strategies, one or more machine learning models to predict a class corresponding to a different multi-step engagement strategy;
generating, by the at least one computing device and using the one or more machine learning models, a prediction of the different multi-step engagement strategy based on the class corresponding to the different multi-step engagement strategy and based on the delivery of the content associated with the campaign according, in part, to both the user-created multi-step engagement strategy and the one or more randomly varied multi-step engagement strategies; and
presenting, by the at least one computing device and via the user interface, a visualization of the prediction of the different multi-step engagement strategy, wherein the visualization includes indications of at least two content deliveries controlled by the prediction of the different multi-step engagement strategy.