Cognitive Assistant for Co-Generating Creative Content

• US 20200134089A1
• Filed: 10/24/2018
• Published: 04/30/2020
• Est. Priority Date: 10/24/2018
• Status: Active Grant

• Current Assignee: International Business Machines Corporation
• Original Assignee: International Business Machines Corporation
• Inventors: Sankaran, Anush, Lohia, Pranay, Agrawal, Priyanka, Shrivastava, Disha, Laha, Anirban, Jain, Parag
• Granted Patent: US 11,301,640 B2
• Time in Patent Office: Days
• CPC Class Codes: G06F 40/20 Natural language analysis semantic analysis of natural language G06F40/30
• CPC Class Codes: G06F 40/56 Natural language generation
• CPC Class Codes: G06N 3/044 Recurrent networks, e.g. Hopfield networks
• CPC Class Codes: G06N 3/045 Combinations of networks
• CPC Class Codes: G06N 3/047 Probabilistic or stochastic networks
• CPC Class Codes: G06N 3/088 Non-supervised learning, e.g. competitive learning
• CPC Class Codes: G06Q 30/0276 Advertisement creation

Abstract

Methods, systems, and computer program products related to a cognitive assistant for co-generating creative content are provided herein. A computer-implemented method includes obtaining semantic-level inputs from at least one user, wherein the semantic-level inputs pertain to multiple aspects of a desired content narrative; generating textual content based at least in part on the semantic-level inputs, wherein said generating the textual content comprises applying one or more deep learning algorithms to the semantic-level inputs; generating image content based at least in part on the generated textual content; creating the desired content narrative by integrating (i) the generated textual content and (ii) the generated image content; and outputting the desired content narrative to the at least one user.

Claims

1. A computer-implemented method, the method comprising steps of:

obtaining semantic-level inputs from at least one user, wherein the semantic-level inputs pertain to multiple aspects of a desired content narrative;

generating textual content based at least in part on the semantic-level inputs, wherein said generating the textual content comprises applying one or more deep learning algorithms to the semantic-level inputs;

generating image content based at least in part on the generated textual content;

creating the desired content narrative by integrating (i) the generated textual content and (ii) the generated image content; and

outputting the desired content narrative to the at least one user;

wherein the steps are carried out by at least one computing device.

2. The computer-implemented method of claim 1, wherein the semantic-level inputs comprise mood information.

3. The computer-implemented method of claim 1, wherein the semantic-level inputs comprise setting information.

4. The computer-implemented method of claim 1, wherein the semantic-level inputs comprise product information.

5. The computer-implemented method of claim 1, wherein the semantic-level inputs comprise target audience information.

6. The computer-implemented method of claim 1, wherein the semantic-level inputs comprise entity-related information.

7. The computer-implemented method of claim 1, wherein the semantic-level inputs comprise thematic information.

8. The computer-implemented method of claim 1, wherein the desired content narrative comprises an advertisement for at least one of (i) one or more products and (ii) one or more services.

9. The computer-implemented method of claim 1, wherein the one or more deep learning algorithms comprises one or more bag-of-words models.

10. The computer-implemented method of claim 1, wherein said generating the image content comprises implementing a pre-trained conditional generative model.

11. A computer program product comprising a computer readable storage medium having program instructions embodied therewith, the program instructions executable by a computing device to cause the computing device to:

obtain semantic-level inputs from at least one user, wherein the semantic-level inputs pertain to multiple aspects of a desired content narrative;

generate textual content based at least in part on the semantic-level inputs, wherein said generating the textual content comprises applying one or more deep learning algorithms to the semantic-level inputs;

generate image content based at least in part on the generated textual content;

create the desired content narrative by integrating (i) the generated textual content and (ii) the generated image content; and

output the desired content narrative to the at least one user.

12. The computer program product of claim 11, wherein the semantic-level inputs comprise at least one of (i) mood information, (ii) setting information, (iii) product information, (iv) target audience information, (v) entity-related information, and (vi) thematic information.

13. The computer program product of claim 11, wherein the desired content narrative comprises an advertisement for at least one of (i) one or more products and (ii) one or more services.

14. The computer program product of claim 11, wherein the one or more deep learning algorithms comprises one or more bag-of-words models.

15. The computer program product of claim 11, wherein said generating the image content comprises implementing a pre-trained conditional generative model.

16. A system comprising:

a memory; and

at least one processor operably coupled to the memory and configured for;

obtaining semantic-level inputs from at least one user, wherein the semantic-level inputs pertain to multiple aspects of a desired content narrative;

generating textual content based at least in part on the semantic-level inputs, wherein said generating the textual content comprises applying one or more deep learning algorithms to the semantic-level inputs;

generating image content based at least in part on the generated textual content;

creating the desired content narrative by integrating (i) the generated textual content and (ii) the generated image content; and

outputting the desired content narrative to the at least one user.

17. A computer-implemented method, the method comprising steps of:

obtaining semantic-level inputs from at least one user, wherein the semantic-level inputs pertain to multiple aspects of a desired content narrative;

generating textual content based at least in part on the semantic-level inputs, wherein said generating the textual content comprises applying one or more rule-based heuristic techniques to the semantic-level inputs;

generating image content based at least in part on the generated textual content;

creating the desired content narrative by integrating (i) the generated textual content and (ii) the generated image content;

generating one or more items of feedback pertaining to modifying one or more of the semantic-level inputs based at least in part on an evaluation score calculated for one or more portions of the desired content narrative;

creating an updated version of the desired content narrative by updating each portion of the desired content narrative that is related to the one or more modified semantic-level inputs; and

outputting the updated content narrative to the at least one user;

wherein the steps are carried out by at least one computing device.

18. The computer-implemented method of claim 17, wherein the semantic-level inputs comprise at least one of (i) mood information, (ii) setting information, (iii) product information, (iv) target audience information, (v) entity-related information, and (vi) thematic information.

19. The computer-implemented method of claim 17, wherein the desired content narrative comprises an advertisement for at least one of (i) one or more products and (ii) one or more services.

20. The computer-implemented method of claim 17, wherein said generating the image content comprises implementing a pre-trained conditional generative model.