System and method for generating preferred ameliorative actions using generative adversarial networks

Patent metadata

Field	Value
Publication number	US-11429996-B2
Application number	US-202016748264-A
Country	US
Kind code	B2
Filing date	Jan 21, 2020
Priority date	Jan 21, 2020
Publication date	Aug 30, 2022
Grant date	Aug 30, 2022

Abstract

A method, system and computer-usable medium are disclosed for improving likelihood of user to purchase a product or service. Interactions of a person, not necessarily the user, are monitored related to the product or service, along with actions by a business related to the person and the product or service. A trained Generative Adversarial Network model is applied to the monitored interactions to form recommend actions that the business should take to achieve ameliorative actions by the user. A reward feedback iterative adjustment of the GAN model is used to facilitate purchase of the product or service by the user.

First claim

What is claimed is:

1. A computer-implementable method for improving a likelihood of a person making a purchase comprising:
monitoring interactions of a person related to a product and actions by a business related to the person and the product;
applying a trained Generative Adversarial Network (GAN), wherein the GAN is initially trained with random noise or corpus of data, and training of the GAN is dynamic with actions taken and will be taken by users of the product;
monitoring by the GAN, interactions to generate a sequence of actions by the business in response to the monitored interactions and actions, wherein the GAN applies the sequence of actions within a reinforcement learning solution, and wherein state and action are determined by a generator component and reward is decided by a discriminator component, wherein the generator component generates data instances, and the discriminator component decides whether instances are truthful or untruthful; and
utilizing a reward feedback iterative adjustment of the model to facilitate a purchase of the product by the user.

2. The method of claim 1 , wherein the GANs model represents the interactions in an embedded form.

3. The method of claim 1 , wherein the GANs model includes the generator component initially trained via recurrent neural network (RNN) with long short-term memory (LSTM) and discriminator component initially trained via convolutional neural network (CNN).

4. The method of claim 3 , wherein the CNN is used to distinguish between real and fake sequence of actions.

5. The method of claim 1 , wherein a reinforcement learning policy is based on the interactions over a period of time.

6. The method of claim 1 , wherein the actions by the business are interactions with the user selected from a group consisting of showing relevant content, triggering a chat, emailing specific demos for use cases, offering a discount, offering technical expertise, and inviting to a webinar.

7. The method of claim 1 , wherein the interactions of a person related to the product is selected from the group consisting of searching on the product, reading marketing content, signing up for a trial account, executing an application program interface, increasing application program interface usage, reducing application program interface usage, and upgrading to a premium prescription.

8. A system comprising:
a processor;
a data bus coupled to the processor; and
a non-transitory, computer-readable storage medium embodying computer program code, the non-transitory, computer-readable storage medium being coupled to the data bus, the computer program code interacting with a plurality of computer operations for improving a likelihood of a user making a purchase executable by the processor and configured for:
monitoring interactions of a person related to a product and actions by a business related to the person and the product;
applying a trained Generative Adversarial Network (GAN), wherein the GAN is initially trained with random noise or corpus of data, and training of the GAN is dynamic with actions taken and will be taken by users of the product;
monitoring by the GAN, interactions to generate a sequence of actions by the business in response to the monitored interactions and actions, wherein the GAN applies the sequence of actions within a reinforcement learning solution, and wherein state and action are determined by a generator component and reward is decided by a discriminator component, wherein the generator component generates data instances, and the discriminator component decides whether instances are truthful or untruthful; and
utilizing a reward feedback iterative adjustment of the model to facilitate a purchase of the product by the user.

9. The system of claim 8 , wherein the GANs model represents the interactions in an embedded form.

10. The system of claim 8 , wherein the GANs model includes the generator component initially trained via recurrent neural network (RNN) with long short-term memory (LSTM) and discriminator component initially trained via convolutional neural network (CNN).

11. The method of claim 10 , wherein the CNN is used to distinguish between real and fake sequence of actions.

12. The system of claim 8 , wherein a reinforcement learning policy is based on the interactions over a period of time.

13. The system of claim 8 , wherein the actions by the business are interactions with the user selected from a group consisting of showing relevant content, triggering a chat, emailing specific demos for use cases, offering a discount, offering technical expertise, and inviting to a webinar.

14. The system of claim 8 , wherein the interactions of a person related to the product is selected from the group consisting of searching on the product, reading marketing content, signing up for a trial account, executing an application program interface, increasing application program interface usage, reducing application program interface usage, and upgrading to a premium prescription.

15. A non-transitory, computer-readable storage medium embodying computer program code, the computer program code comprising computer executable instructions configured for:
monitoring interactions of a person related to a product and actions by a business related to the person and the product;
applying a trained Generative Adversarial Network (GAN), wherein the GAN is initially trained with random noise or corpus of data, and training of the GAN is dynamic with actions taken and will be taken by users of the product;
monitoring by the GAN, interactions to generate a sequence of actions by the business in response to the monitored interactions and actions, wherein the GAN applies the sequence of actions within a reinforcement learning solution, and wherein state and action are determined by a generator component and reward is decided by a discriminator component, wherein the generator component generates data instances, and the discriminator component decides whether instances are truthful or untruthful; and
utilizing a reward feedback iterative adjustment of the model to facilitate a purchase of the product by the user.

16. The non-transitory, computer-readable storage medium of claim 15 , wherein the GANs model represents the interactions in an embedded form.

17. The non-transitory, computer-readable storage medium of claim 15 , wherein the GANs model includes the generator component initially trained via recurrent neural network (RNN) with long short-term memory (LSTM) and discriminator component initially trained via convolutional neural network (CNN).

18. The non-transitory, computer-readable storage medium of claim 15 , wherein a reinforcement learning policy is based on the interactions over a period of time.

19. The non-transitory, computer-readable storage medium of claim 15 , wherein the actions by the business are interactions with the user selected from a group consisting of showing relevant content, triggering a chat, emailing specific demos for use cases, offering a discount, offering technical expertise, and inviting to a webinar.

20. The non-transitory, computer-readable storage medium of claim 15 , wherein the interactions of a person related to the product is selected from the group consisting of searching on the product, reading marketing content, signing up for a trial account, executing an application program interface, increasing application program interface usage, reducing application program interface usage, and upgrading to a premium prescription.

Assignees

• IBM

Inventors

• Mazumder Sourav
• Kozhaya Joseph N
• Kwatra Shikhar

Classifications

• G06N3/045

  Combinations of networks · CPC title

• G06N3/044

  Recurrent networks, e.g. Hopfield networks · CPC title

• G06N3/092

  Reinforcement learning · CPC title

• G06N3/0464

  Convolutional networks [CNN, ConvNet] · CPC title

• G06N3/0442

  characterised by memory or gating, e.g. long short-term memory [LSTM] or gated recurrent units [GRU] · CPC title