US 12,712,918 B2Granted

Systems and methods for evaluating an electronic communication

AssigneeWells Fargo Bank, N.A.

InventorsKelvin Hay, Kathryn Lynne Sanders, Angela Melita Maryland, Leo Rinaldi, Abe Akhimiomoh

Life of this patent

Estimating the term…

Loading patent…

Abstract

Systems, apparatuses, methods, and computer program products are disclosed for evaluating an electronic communication. An example method includes receiving an electronic communication and generating, a plurality of tokens based on the plurality of characters. The example method further includes determining whether the electronic communication includes disallowed content based on the plurality of tokens using a contextual analysis machine learning model and in an instance in which the electronic communication is determined to include disallowed content, determining, an infraction event using the contextual analysis machine learning model. The example method further includes determining an infraction alleviation action based on the infraction event type and causing performance of the infraction alleviation action.

Claims20

Dependent claims

A method for evaluating an electronic communication, the method comprising:

receiving, by communications hardware, the electronic communication, wherein a body of the electronic communication comprises a plurality of characters;
generating, by a contextual analysis engine and based on the plurality of characters, a plurality of tokens;
determining, by the contextual analysis engine and using a contextual analysis machine learning model, an infraction token probability score for a token of the plurality of tokens, wherein (a) the infraction token probability score for the token is determined based on semantic information about the token's role in a sentence derived from a surrounding token, (b) the infraction token probability score indicates a probability that the token corresponds to disallowed content, and (c) the disallowed content is associated with a rule;
determining, by the contextual analysis engine and using the contextual analysis machine learning model, whether the electronic communication includes the disallowed content based on the infraction token probability score for the token;
in response to determining that the electronic communication includes the disallowed content, determining, by the contextual analysis engine and using the contextual analysis machine learning model, an infraction event, wherein the infraction event corresponds to an infraction event type and the infraction event type is based on the disallowed content included in the electronic communication;
determining, by the contextual analysis engine and based on the infraction event type, an infraction alleviation action;
and
causing, by the contextual analysis engine, performance of the infraction alleviation action.

5 more elements

An apparatus for evaluating an electronic communication, the apparatus comprising:

communications hardware configured to receive the electronic communication, wherein a body of the electronic communication comprises a plurality of characters;
and
a contextual analysis engine configured to:
generate, based on the plurality of characters, a plurality of tokens,
determine, using a contextual analysis machine learning model, an infraction token probability score for a token of the plurality of tokens, wherein (a) the infraction token probability score for the token is determined based on semantic information about the token's role in a sentence derived from a surrounding token, (b) the infraction token probability score indicates a probability that the token corresponds to disallowed content, and (c) the disallowed content is associated with a rule;
determine, using the contextual analysis machine learning model, whether the electronic communication includes the disallowed content based on the infraction token probability score for the token,
in response to determining that the electronic communication includes the disallowed content, determine, using the contextual analysis machine learning model, an infraction event, wherein the infraction event corresponds to an infraction event type and the infraction event type is based on the disallowed content included in the electronic communication,
determine, based on the infraction event type, an infraction alleviation action, and
cause performance of the infraction alleviation action.

6 more elements

A computer program product for evaluating an electronic communication, the computer program product comprising a non-transitory computer-readable storage medium storing instructions that, when executed by an apparatus, cause the apparatus to:

receive the electronic communication, wherein a body of the electronic communication comprises a plurality of characters;
generate, based on the plurality of characters, a plurality of tokens;
determine, using a contextual analysis machine learning model, an infraction token probability score for a token of the plurality of tokens, wherein (a) the infraction token probability score for the token is determined based on semantic information about the token's role in a sentence derived from a surrounding token, (b) the infraction token probability score indicates a probability that the token corresponds to disallowed content, and (c) the disallowed content is associated with a rule;
determine, using the contextual analysis machine learning model, whether the electronic communication includes the disallowed content based on the infraction token probability score for the token;
in response to determining that the electronic communication includes the disallowed content, determine, using the contextual analysis machine learning model, an infraction event, wherein the infraction event corresponds to an infraction event type and the infraction event type is based on the disallowed content included in the electronic communication;
determine, based on the infraction event type, an infraction alleviation action;
and
cause performance of the infraction alleviation action.

5 more elements

Dependencies are inferred from the claim text.