IP Library › Granted Patent US 12,579,362
Granted Patent B2
US 12,579,362 · App. 18/101,290 · Granted Mar 17, 2026

Automated text generation using artificial intelligence techniques

Inventors: Aaron K. Baughman (Research Triangle Park, NC); Nicholas Michael Wilkin (Seattle, WA); Eris Opal Rashon Calhoun (Atlanta, GA); Gray Franklin Cannon (Atlanta, GA)
Assignee: International Business Machines Corporation
G06F40/279G06F40/10G06N20/00
View Patent ↗
Loading inventors, assignments & file history…
Monitor This Case
Get email alerts when status or documents change.
Order Certified Copies
Most orders are placed with the USPTO same day — all within 24 business hours.
Order via The Patent Place →
Pre-filled with this patent's details
Quick Facts
Patent No.
US 12,579,362
App. No.
18/101,290
Granted
Mar 17, 2026
Kind
B2
Abstract

Methods, systems, and computer program products for automated text generation using artificial intelligence techniques are provided herein. A computer-implemented method includes simulating, using artificial intelligence techniques, at least one textual description of at least one event related to an input event; identifying historical text sources related to the input event based on processing the at least one simulated textual description; identifying content related to argument theses from at least a portion of the historical text sources; determining at least one measure of alignment between at least a portion of the at least one simulated textual description and at least a portion of the content related to argument theses; and automatically generating text describing the input event based on the at least one measure of alignment.

Claims (43)

1 . A system comprising:

a memory configured to store program instructions; and

a processor operatively coupled to the memory to execute the program instructions to:

simulate, using one or more artificial intelligence techniques, at least one textual description of at least one event related to an input event;

identify one or more historical text sources related to the input event based at least in part on processing the at least one simulated textual description;

identify content related to one or more argument theses from at least a portion of the one or more historical text sources;

determine at least one measure of alignment between at least a portion of the at least one simulated textual description and at least a portion of the content related to one or more argument theses; and

automatically generate text describing the input event based at least in part on the at least one measure of alignment between the at least a portion of the at least one simulated textual description and the at least a portion of the content related to one or more argument theses, wherein automatically generating text describing the input event comprises automatically generating at least one opinionated abstractive summary of the input event based at least in part on the at least one measure of alignment between the at least a portion of the at least one simulated textual description and the at least a portion of the content related to one or more argument theses.

2 . The system of claim 1 , wherein automatically generating text describing the input event comprises automatically generating at least one opinionated abstractive summary of the input event using text from the at least one simulated textual description that aligns with a selected one of the one or more argument theses.

3 . The system of claim 1 , wherein simulating at least one textual description of the at least one event comprises simulating the at least one textual description of the at least one event using one or more text-to-text models.

4 . The system of claim 3 , wherein the one or more text-to-text models comprise at least one a text-to-text transfer transformer.

5 . The system of claim 1 , wherein simulating at least one textual description of the at least one event comprises generating at least one abstractive text summary of the at least one event using the one or more artificial intelligence techniques.

6 . The system of claim 1 , wherein identifying one or more historical text sources comprises:

extracting one or more themes from the at least one simulated textual description; and

querying at least one database containing historical text sources using at least a portion of the one or more themes.

7 . The system of claim 1 , wherein identifying one or more historical text sources comprises identifying one or more historical opinionated text sources based at least in part on processing the at least one simulated textual description.

8 . The system of claim 1 , wherein identifying content related to one or more argument theses from at least a portion of the one or more historical text sources comprises identifying at least one of evidence from the one or more historical text sources that supports at least one of the one or more argument theses and evidence from the one or more historical text sources that refutes at least one of the one or more argument theses.

9 . The system of claim 1 , wherein the processor is further operatively coupled to the memory to execute the program instructions to:

automatically train at least a portion of the one or more artificial intelligence techniques based at least in part on the automatically generated text describing the input event.

10 . 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:

simulate, using one or more artificial intelligence techniques, at least one textual description of at least one event related to an input event;

identify one or more historical text sources related to the input event based at least in part on processing the at least one simulated textual description;

identify content related to one or more argument theses from at least a portion of the one or more historical text sources;

determine at least one measure of alignment between at least a portion of the at least one simulated textual description and at least a portion of the content related to one or more argument theses; and

automatically generate text describing the input event based at least in part on the at least one measure of alignment between the at least a portion of the at least one simulated textual description and the at least a portion of the content related to one or more argument theses, wherein automatically generating text describing the input event comprises automatically generating at least one opinionated abstractive summary of the input event based at least in part on the at least one measure of alignment between the at least a portion of the at least one simulated textual description and the at least a portion of the content related to one or more argument theses.

11 . The computer program product of claim 10 , wherein automatically generating text describing the input event comprises automatically generating at least one opinionated abstractive summary of the input event using text from the at least one simulated textual description that aligns with a selected one of the one or more argument theses.

12 . The computer program product of claim 10 , wherein simulating at least one textual description of the at least one event comprises simulating the at least one textual description of the at least one event using one or more text-to-text models.

13 . The computer program product of claim 10 , wherein simulating at least one textual description of the at least one event comprises generating at least one abstractive text summary of the at least one event using the one or more artificial intelligence techniques.

14 . A computer-implemented method comprising:

simulating, using one or more artificial intelligence techniques, at least one textual description of at least one event related to an input event;

identifying one or more historical text sources related to the input event based at least in part on processing the at least one simulated textual description;

identifying content related to one or more argument theses from at least a portion of the one or more historical text sources;

determining at least one measure of alignment between at least a portion of the at least one simulated textual description and at least a portion of the content related to one or more argument theses; and

automatically generating text describing the input event based at least in part on the at least one measure of alignment between the at least a portion of the at least one simulated textual description and the at least a portion of the content related to one or more argument theses, wherein automatically generating text describing the input event comprises automatically generating at least one opinionated abstractive summary of the input event based at least in part on the at least one measure of alignment between the at least a portion of the at least one simulated textual description and the at least a portion of the content related to one or more argument theses;

wherein the method is carried out by at least one computing device.

15 . The computer-implemented method of claim 14 , wherein automatically generating text describing the input event comprises automatically generating at least one opinionated abstractive summary of the input event using text from the at least one simulated textual description that aligns with a selected one of the one or more argument theses.

16 . The computer-implemented method of claim 14 , wherein simulating at least one textual description of the at least one event comprises simulating the at least one textual description of the at least one event using one or more text-to-text models.

17 . The computer-implemented method of claim 14 , wherein software implementing the method is provided as a service in a cloud environment.

18 . The computer-implemented method of claim 14 , wherein simulating at least one textual description of the at least one event comprises generating at least one abstractive text summary of the at least one event using the one or more artificial intelligence techniques.

19 . The computer-implemented method of claim 14 , wherein identifying one or more historical text sources comprises:

extracting one or more themes from the at least one simulated textual description; and

querying at least one database containing historical text sources using at least a portion of the one or more themes.

20 . The computer-implemented method of claim 14 , wherein identifying one or more historical text sources comprises identifying one or more historical opinionated text sources based at least in part on processing the at least one simulated textual description.

Assignments (1)
ASSIGNMENT OF ASSIGNOR'S INTEREST Recorded Jan 25, 2023
From: BAUGHMAN, AARON K.; WILKIN, NICHOLAS MICHAEL; CALHOUN, ERIS OPAL RASHON; CANNON, GRAY FRANKLIN
To: INTERNATIONAL BUSINESS MACHINES CORPORATION
Reel/Frame 062483/0486 →
Continuity (1)
Related Publication 20240249183A1 · Jul 25, 2024
References Cited (21)
US 8498930B2 · Chung et al. · 2013 [cited by applicant]
US 10395216B2 · Coffing · 2019 [cited by applicant]
US 10593002B2 · Palombi et al. · 2020 [cited by applicant]
US 11127099B2 · Eidelmal et al. · 2021 [cited by applicant]
US 20180349476A1 · Carmeli · 2018 [cited by examiner]
US 20210374338A1 · Shrivastava · 2021 [cited by examiner]
CN 104216954B · 2017 [cited by applicant]
CN 104380146B · 2018 [cited by applicant]
Genest et al., Framework for Abstractive Summarization using Text-to-Text Generation, Proceedings of the 49th Annual Meeting of the Association for Computational Linguistics, pp. 64-73, Jun. 24, 2011. [cited by applicant]
Buciumas, S. Reinforcement Learning Models for Abstractive Text Summarization, ACM SE '19: Proceedings of the 2019 ACM Southeast Conference, Apr. 2019 pp. 270-271. [cited by applicant]
Chen et al., Abstractive Snippet Generation, In Proceedings of the Web Conference 2020 (WWW '20), Apr. 20-24, 2020. [cited by applicant]
Yuan et al., Research on cross-language text similarity calculation, 2015 IEEE 5th International Conference on Electronics Information and Emergency Communication. [cited by applicant]
Runze et al., Spatial load forecasting based on unstructured information processing and multi-attribute deep learning, 2017 5th International Symposium on Electrical and Electronics Engineering. [cited by applicant]
Ahuja et al., Language2Pose: Natural Language Grounded Pose Forecasting, 2019 International Conference on 3D Vision (3DV). [cited by applicant]
Wang et al., Neural Network-Based Abstract Generation for Opinions and Arguments, Jun. 2016. [cited by applicant]
Kane et al., Prototypical Prospection: Future Events are More Prototypically Represented and Simulated than Past Events, European Journal of Social Psychology, Eur. J. Soc. Psychol. 42, 354-362 (2012). [cited by applicant]
Huang et al., Distinguishing Past, On-Going, and Future Events: The EventStatus Corpus, Nov. 2016. [cited by applicant]
IPCOM000270295D, IP.com, Method for Guided Data Exploration in a Multi-Widget Dashboard, Jun. 24, 2022. [cited by applicant]
Luk et al., IPCOM000264007D, IP.com, Social Listening Advanced Analytics Solution (SOLIAS) for Data Storage Industry, Oct. 30, 2020. [cited by applicant]
IPCOM000157414D, IP.com, A Method for Relating and Displaying IT Capacity Forecasts with Business Metadata, Aug. 28, 2007. [cited by applicant]
Mell et al., The NSIT Definition of Cloud Computing, NIST Special Publication 800-145, Sep. 2011. [cited by applicant]