US 11,687,716 B2Granted

Machine-learning techniques for augmenting electronic documents with data-verification indicators

AssigneeAdobe Inc.

InventorsNavita Goyal, Vipul Shankhpal, Priyanshu Gupta, Ishika Singh, Baldip Singh Bijlani, Anandha velu Natarajan

Life of this patent

Estimating the term…

Loading patent…

Abstract

This disclosure involves executing machine-learning techniques for transforming or otherwise processing electronic data. This disclosure, for example, relates to executing machine-learning techniques to generate data-verification indicators that augment electronic documents to represent the veracity of text. The machine-learning techniques include neural networks trained to retrieve and analyze evidence regarding content of electronic documents and to generate indicators of veracity to be displayed with that content via electronic reading software.

Claims20

Dependent claims

A computer-implemented method, comprising:

receiving an input indicating a selection of a target text segment;
determining a veracity metric of the target text segment by, at least:
extracting a first text segment from an electronic document;
generating, using a trained reasoning model, an output classifying the first text segment as lacking information to support or refute the target text segment;
extracting a second text segment from the electronic document or another electronic document in response to classifying the first text segment as lacking information to support or refute the target text segment;
and
classifying, with the trained reasoning model, a combination of the first text segment and the second text segment according to a degree to which the trained reasoning model measures whether the combination supports the target text segment;
and
generating an augmentation for a graphical interface to display, adjacent to the target text segment, a verification indicator representing the determined veracity metric of the target text segment.

6 more elements

A computer-program product tangibly embodied in a non-transitory machine-readable storage medium, including instructions configured to cause a processing apparatus to perform operations including:

identifying a target text segment included in a target electronic document;
determining a veracity metric of the target text segment at least in part by:
determining that a first text segment lacks information to support or refute the target text segment;
obtaining a second text segment in response to determining that the first text segment lacks information to support or refute the target text segment;
and
generating, using a trained reasoning model, an indicator of the veracity metric based on the first text segment and the second text segment;
and
outputting an augmentation for the target electronic document, the augmentation including a verification indicator indicating the determined veracity metric of the target text segment.

5 more elements

A system comprising:

one or more processors;
and
a non-transitory computer-readable medium communicatively coupled to the one or more processors and storing program code executable by the one or more processors, the program code implementing a text verification system configured to determine a veracity metric of a target text segment, the text verification system comprising:
a first-stage text retrieval model configured to infer a relevance between a first text segment included in an electronic document and the target text segment, the first text segment being selected for extraction from the electronic document based on the inferred relevance;
a reasoning model configured to classify the first text segment as lacking information to support or refute the target text segment;
and
a second-stage text retrieval model configured to extract a second text segment from the electronic document or another electronic document in response to the first text segment being classified as lacking information to support or refute the target text segment, and
wherein the reasoning model is further configured to generate an indicator of the veracity metric based on a combination of the first and second text segments, wherein the indicator of the veracity metric is usable for augmenting a target electronic document with a verification indicator that is selectively displayable adjacent to the target text segment.

5 more elements

Dependencies are inferred from the claim text.