Multi-modal network-based assertion verification

Patent metadata

Field	Value
Publication number	US-12118306-B2
Application number	US-202117513169-A
Country	US
Kind code	B2
Filing date	Oct 28, 2021
Priority date	Oct 28, 2021
Publication date	Oct 15, 2024
Grant date	Oct 15, 2024

Abstract

A processing system may obtain first content including human language via a computing network. The processing system may next identify an assertion in the first content and identify one or more content sources containing second content relating to the assertion. The processing system may then determine whether the second content relating to the assertion corroborates or contradicts the assertion, and may present the first content with an indication of whether the second content corroborates or contradicts the assertion.

First claim

What is claimed is:

1. A method comprising:
obtaining, by a processing system including at least one processor, a first content comprising human language via at least one computing network;
identifying, by the processing system, at least one assertion in the first content, wherein the identifying comprises applying an open information extraction process to create a representation of the at least one assertion as a first formal relation comprising a first binary relation;
identifying, by the processing system, one or more content sources containing at least one second content relating to the at least one assertion;
determining, by the processing system, whether the at least one second content relating to the at least one assertion corroborates or contradicts the at least one assertion, wherein the determining whether the at least one second content corroborates or contradicts the at least one assertion comprises:
identifying at least a second assertion in the at least one second content;
creating a representation of the at least the second assertion as a second formal relation; and
determining whether the second formal relation comprises an affirmation of the first formal relation or a negation of the first formal relation; and

presenting, by the processing system, the first content with an indication of whether the at least one second content relating to the at least one assertion corroborates or contradicts the at least one assertion.

2. The method of claim 1 , wherein the identifying the one or more content sources containing at least one second content relating to the at least one assertion comprises:
obtaining the at least one second content;
calculating a term frequency-inverse document frequency metric between the at least one second content and the first content; and
determining that the at least one second content relates to the at least one assertion in response to the term frequency-inverse document frequency metric exceeding a threshold term frequency-inverse document frequency metric.

3. The method of claim 1 , further comprising:
obtaining an input requesting a verification of the at least one assertion.

4. The method of claim 1 , wherein the one or more content sources comprise:
an online news source;
a review website;
an online journal;
an online article repository; or
an online encyclopedia.

5. The method of claim 1 , wherein the first content comprises text.

6. The method of claim 5 , wherein the text comprises:
an article;
a short message service message;
a social-media posting;
a posting to an online forum;
a search result; or
a comment added to a comment section for an online media content.

7. The method of claim 1 , wherein the first content comprises speech.

8. The method of claim 7 , wherein the identifying the at least one assertion in the first content comprises:
converting the speech to text via a speech recognition process.

9. The method of claim 1 , wherein the identifying the one or more content sources containing the at least one second content relating to the at least one assertion comprises:
identifying the one or more content sources from among content sources having trust ratings above a threshold trust rating.

10. The method of claim 9 , wherein for each content source of the one or more content sources, a trust rating is based upon at least one of:
a user rating by a user to whom the first content is presented; or
a number of times or a percentage of agreement regarding assertions of fact with other content sources having trust ratings above the threshold trust rating.

11. The method of claim 1 , wherein the one or more content sources comprise at least one sensor device, and wherein the at least one second content comprises sensor data collected from the at least one sensor device.

12. The method of claim 11 , wherein the at least one sensor device comprises:
a microphone; or
a camera.

13. The method of claim 11 , wherein the identifying the one or more content sources includes identifying features from the sensor data, wherein the determining comprises determining whether the features from the sensor data corroborate or contradict the at least one assertion.

14. The method of claim 11 , wherein the at least one sensor device is deployed at a location that is geographically relevant to the at least one assertion.

15. The method of claim 14 , wherein the at least one assertion relates to the location.

16. A non-transitory computer-readable medium storing instructions which, when executed by a processing system including at least one processor, cause the processing system to perform operations, the operations comprising:
obtaining a first content comprising human language via at least one computing network;
identifying at least one assertion in the first content, wherein the identifying comprises applying an open information extraction process to create a representation of the at least one assertion as a first formal relation comprising a first binary relation;
identifying one or more content sources containing at least one second content relating to the at least one assertion;
determining whether the at least one second content relating to the at least one assertion corroborates or contradicts the at least one assertion, wherein the determining whether the at least one second content corroborates or contradicts the at least one assertion comprises:
identifying at least a second assertion in the at least one second content;
creating a representation of the at least the second assertion as a second formal relation; and
determining whether the second formal relation comprises an affirmation of the first formal relation or a negation of the first formal relation; and

presenting the first content with an indication of whether the at least one second content relating to the at least one assertion corroborates or contradicts the at least one assertion.

17. An apparatus comprising:
a processing system including at least one processor; and
a computer-readable medium storing instructions which, when executed by the processing system, cause the processing system to perform operations, the operations comprising:
obtaining a first content comprising human language via at least one computing network;
identifying at least one assertion in the first content, wherein the identifying comprises applying an open information extraction process to create a representation of the at least one assertion as a first formal relation comprising a first binary relation;
identifying one or more content sources containing at least one second content relating to the at least one assertion;
determining whether the at least one second content relating to the at least one assertion corroborates or contradicts the at least one assertion, wherein the determining whether the at least one second content corroborates or contradicts the at least one assertion comprises:
identifying at least a second assertion in the at least one second content;
creating a representation of the at least the second assertion as a second formal relation; and
determining whether the second formal relation comprises an affirmation of the first formal relation or a negation of the first formal relation; and

presenting the first content with an indication of whether the at least one second content relating to the at least one assertion corroborates or contradicts the at least one assertion.

18. The apparatus of claim 17 , wherein the identifying the one or more content sources containing at least one second c

Assignees

• At & T Ip I Lp

Inventors

• Chien Ginger
• Cui Zhi
• Zavesky Eric
• Moton Jr Robert T
• Luu Adrianne Binh
• Koch Robert

Classifications

• G06F40/284

  Lexical analysis, e.g. tokenisation or collocates · CPC title

• G06F40/216

  using statistical methods · CPC title

• G06F40/237

  Lexical tools · CPC title

• G06F40/279 Primary

  Recognition of textual entities · CPC title

• G06F40/30 Primary

  Semantic analysis · CPC title