IP Library Granted Patent US 12,423,337
Granted Patent B2
US 12,423,337 · App. 18/100,731 · Granted Sep 23, 2025

Context-based question generation from communication data

Inventors: Srini Bhagavan (Dublin, CA); Prasanna Alur Mathada (Bangalore, IN); Shwetha Gopalakrishna (Banagalore, IN); Hrishikesh Sujaya Kumar (Bangalore, IN)
Assignee: International Business Machines Corporation
G06F16/3329G06F40/40
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 12,423,337
App. No.
18/100,731
Granted
Sep 23, 2025
Kind
B2
Abstract

Systems and methods are provided to generate context-based questions from communications data. In embodiments, a method includes: determining, by a processor set, context of text under study obtained from digital communications data; generating, by the processor set, a text-based summary of the text under study using a natural language processing (NLP) tool to eliminate text from the text under study based on keywords of a dynamic knowledge base; constructing, by the processor set, a question phrase in sentence form based statements in the text under study, wherein the one or more statements constitute an answer to the question phrase; outputting, by the processor set, a question and answer set including the question phrase and the statements; and automatically updating, by the processor set, the dynamic knowledge base based on the question and answer set, thereby generating an updated dynamic knowledge base.

Claims (44)

1. A method, comprising:

determining, by a processor set, context of text under study obtained from digital communications data generated by a computer application;

determining, by the processor set, a number of contextual switches of the text under study that indicates a size of the text under study;

generating, by the processor set, a text-based summary of the text under study using a natural language processing (NLP) tool based on keywords of a dynamic knowledge base and the determined number of contextual switches, the dynamic knowledge base being a database of keywords associated with a source of the digital communications data;

constructing, by the processor set, a question phrase in sentence form based on one or more statements in the text under study, wherein the one or more statements constitute an answer to the question phrase;

outputting, by the processor set, a question and answer set to a user interface associated with the computer application, the question and answer set including the question phrase and the one or more statements; and

automatically transmitting, by the processor set, the question and answer set to the dynamic knowledge base, thereby generating an updated dynamic knowledge base.

2. The method of claim 1 , further comprising automatically obtaining, by the processor set, the digital communications data from a communications application based on stored user-configured rules, wherein the digital communications data is generated during a communication between two or more participants.

3. The method of claim 1 , further comprising selecting, by the processor set, the dynamic knowledge base of keywords from a plurality of dynamic knowledge bases of keywords based on the context of the text under study.

4. The method of claim 1 , wherein the context comprises one or more selected from the group consisting of: a source of the digital communications data, the determined number of contextual switches of the text under study, and content of the text under study.

5. The method of claim 4 , further comprising selecting, by the processor set, the NLP tool from a plurality of NLP tools based on the context of the text under study.

6. The method of claim 5 , wherein the NLP tool is selected based on the source of the digital communications data.

7. The method of claim 5 , wherein the NLP tool is selected based on the content of the text under study.

8. The method of claim 5 , wherein the NLP tool is selected from a group consisting of: a sentence scoring tool, a text ranking tool, and an unsupervised learning tool.

9. A computer program product comprising one or more computer readable storage media having program instructions collectively stored on the one or more computer readable storage media, the program instructions executable to:

determine context of text under study obtained from digital communications data generated by a computer application;

determine a number of contextual switches of the text under study that indicates a size of the text under study;

generate a text-based summary of the text under study using a natural language processing (NLP) tool, based on keywords of a dynamic knowledge base and the determined number of contextual switches, the dynamic knowledge base being a database of keywords associated with a source of the digital communications data;

construct a question phrase in sentence form based on one or more statements in the text under study, wherein the one or more statements constitute an answer to the question phrase;

output, to a user interface associated with the computer application based on stored rules, a question and answer set including the question phrase and the one or more statements; and

automatically transmit the question and answer set to the dynamic knowledge base, thereby generating an updated dynamic knowledge base.

10. The computer program product of claim 9 , wherein the program instructions are further executable to automatically obtain the digital communications data from a communications application based on stored user-configured rules, wherein the digital communications data is generated during a communication between two or more participants.

11. The computer program product of claim 9 , wherein the program instructions are further executable to select the dynamic knowledge base of keywords from a plurality of dynamic knowledge bases of keywords based on the context of the text under study.

12. The computer program product of claim 9 , wherein the context comprises one or more selected from the group consisting of: a source of the digital communications data, the determined number of contextual switches of the text under study, and content of the text under study.

13. The computer program product of claim 12 , wherein the program instructions are further executable to select the NLP tool from a plurality of NLP tools based on the context of the text under study.

14. The computer program product of claim 13 , wherein the digital communications data includes direct messages and the NLP tool is a sentence scoring tool that scores sentences based on word frequency in the direct messages.

15. The computer program product of claim 13 , wherein the digital communications data includes group chat messages and the NLP tool is text ranking tool that ranks text using a universal sentence encoder.

16. The computer program product of claim 13 , wherein the digital communications data includes audio or video recording of a meeting, and the NLP tool is an unsupervised learning tool that uses skip-thought vectors to:

perform text cleaning;

encode sentences to output vectors;

perform K means clustering to partition the digital communications data; and

extract sentences that are closest to cluster centers.

17. A system comprising:

a processor set, one or more computer readable storage media, and program instructions collectively stored on the one or more computer readable storage media, the program instructions executable to:

determine context of text under study obtained from digital communications data generated by a computer application;

determine a number of contextual switches of the text under study that indicates a size of the text under study;

generate a text-based summary of the text under study using a natural language processing (NLP) tool, based on keywords of a dynamic knowledge base and the determined number of contextual switches, the dynamic knowledge base being a database of keywords associated with a source of the digital communications data;

construct question phrases in sentence form based on one or more statements in the text under study, wherein the one or more statements constitute an answer to the question phrases;

rank the question phrases for importance using a passive aggressive learning algorithm;

output, to a user interface associated with the computer application, a question and answer set including a subset of the question phrases selected based on the ranking, and at least one of the one or more statements; and

automatically transmit the question and answer set to the dynamic knowledge base, thereby generating an updated dynamic knowledge base.

18. The system of claim 17 , wherein the program instructions are further executable to automatically obtain the digital communications data from a communications application based on stored user-configured rules, wherein the digital communications data is generated during a communication between two or more participants.

19. The system of claim 17 , wherein the program instructions are further executable to select the dynamic knowledge base of keywords from a plurality of dynamic knowledge bases of keywords based on the context of the text under study, and wherein the context comprises one or more selected from the group consisting of: a source of the digital communications data, the determined number of contextual switches of the text under study, and content of the text under study.

20. The system of claim 19 , wherein the program instructions are further executable to select the NLP tool from a plurality of NLP tools based on the context of the text under study.

Assignments (1)
ASSIGNMENT OF ASSIGNOR'S INTEREST Recorded Jan 24, 2023
From: BHAGAVAN, SRINI; MATHADA, PRASANNA ALUR; GOPALAKRISHNA, SHWETHA; KUMAR, HRISHIKESH SUJAYA
To: INTERNATIONAL BUSINESS MACHINES CORPORATION
Reel/Frame 062467/0630 →
Continuity (1)
Related Publication 20240248918A1 · Jul 25, 2024
References Cited (21)
US 20160062967A1 · Cantarero · 2016 [cited by examiner]
US 20170180284A1 · Smullen · 2017 [cited by examiner]
US 20230103076A1 · Zhou · 2023 [cited by examiner]
DE 102017124958 · 2019 [cited by applicant]
DE 102020003966 · 2022 [cited by applicant]
EP 3936234 · 2022 [cited by applicant]
WO 2018060450 · 2018 [cited by applicant]
WO WO2021195149A1 · 2021 [cited by examiner]
Dubey, “Understand Text Summarization and Create Your Own Summarizer in Python”, Towards Data Science, Dec. 22, 2018; 11 Pages. [cited by applicant]
Liu, “Question-based Text Summarization”, A Thesis Submitted to the Faculty of Drexel University, Dec. 2017; 101 Pages. [cited by applicant]
Vakulenko et al., “QRFA: A Data-Driven Model of Information-Seeking Dialogues”, Research Gate, Apr. 2019; 16 Pages. [cited by applicant]
Ma et al., “Multi-document Summarization via Deep Learning Techniques: A Survey”, Association for Computing Machinery, Publication, vol. 1, No. 1, Dec. 2021; 35 Pages. [cited by applicant]
Ling et al., “Leveraging Context for Neural Question Generation in Open-domain Dialogue Systems”, International World Wide Web Conference Committee, Apr. 2020; pp. 2486-2492. [cited by applicant]
Wang et al., “Learning to Ask Questions in Open-domain Conversational Systems with Typed Decoders”, Proceedings of the 56th Annual Meeting of the Association for Computational Linguistics, Jul. 2018; pp. 2193-2203. [cited by applicant]
OptiSol, “Automating Minutes of Meetings using Text Analytics”, Jun. 13, 2022; 4 Pages. [cited by applicant]
Ali et al., “Automatic Question Generation from Sentences”, TALN 2010, Jul. 19-23, 2010; 6 Pages. [cited by applicant]
Joshi, “An Introduction to Text Summarization using the TextRank Algorithm (with Python implementation)”, Automatic Text Summarization Using TextRank Algorithm, https://www.analyticsvidhya.com/blog/2018/11/introduction-… [cited by applicant]
Quora, “How can I generate questions automatically based on the text using NLP techniques?”, https://www.quora.com/How-can-I-generate-questions-automatically-based-on-the-text-using-NLP-techniques, Accessed Jan. 23, 202… [cited by applicant]
Heilman, “Automatic Factual Question Generation from Text”, Thesis Language Technologies Institute School of Computer Science Carnegie Mellon University, 2011; 203 Pages. [cited by applicant]
Heilman et al., “Good Question! Statistical Ranking for Question Generation”, Human Language Technologies: The 2010 Annual Conference of the North American Chapter of the ACL, Jun. 2010; pp. 609-617. [cited by applicant]
Thaker, “Comparing Text Summarization Techniques”, https://medium.com/@thakermadhav/comparing-text-summarization-techniques-d1e2e465584e , Mar. 24, 2019; 18 Pages. [cited by applicant]