US 12,118,079 B2Granted

Resource conservation system for scalable identification of a subset of inputs from among a group of inputs that contributed to an output

AssigneeBank of America Corporation

InventorsJustin Horowitz, Melissa Podrazka, Sameer Sharma

Life of this patent

Estimating the term…

Loading patent…

Abstract

A resource conservation system, including a determination processor may be provided. The determination processor may identify a characterization output that characterizes a plurality of data structures. The characterization output may be based on plurality of inputs. The inputs may be processed through a plurality, or cascade, of artificial intelligence models both in sequence and in parallel. A numerical value may be identified for each data structure. The value may identify a degree of certainty that the determination processor accurately characterized each data structure. When the degree is above a threshold, the determination processor may identify a subset of inputs that most contributed to the characterization output. The determination processor may execute an equation to identify a subset of inputs that most contributed to the output. The equation may involve inputs and/or outputs of each of the cascade of models. Identified inputs may be ranked based on contribution to the outcome.

Claims5

Dependent claims

A method for conserving resources during a suspicious activity reporting (“SAR”) investigation, the method comprising:

receiving a SAR output characterizing a first entity, said SAR output being a number between zero and one;
determining a plurality of most important contributing factors to the SAR output, the determining comprising:
identifying a plurality of suspicious data elements at a monitoring data unit;
feeding the plurality of suspicious data elements into a plurality of models;
processing the plurality of suspicious data elements at the plurality of models;
identifying a plurality of suspicious activity outputs from the plurality of models;
feeding the plurality of suspicious activity outputs into an event processor;
processing the plurality of suspicious activity outputs at the event processor;
grouping the plurality of suspicious activity outputs into a plurality of suspicious events at the event processor;
inputting the plurality of suspicious events into a determination processor;
determining, at the determination processor, a probability of the first entity being associated with an affirmative SAR output;
determining, at the determination processor, that the determined probability is higher than a threshold probability of being associated with an affirmative SAR output;
in order to identify a predetermined number of data elements contributing to the affirmative SAR output:
multiplying the integrated gradient of the determination processor with respect to the suspicious events outputs by (the integrated gradient of the event processor with respect to the plurality of suspicious data elements divided by the suspicious events outputs), which results in a vector of:
a subset of the plurality of suspicious data elements;
and
a probability for each suspicious data element, included in the subset of suspicious data elements, that the suspicious data element contributed to the affirmative SAR output;
ranking the subset of the plurality of suspicious data elements included in the vector by the probability that the suspicious data element contributed to the affirmative SAR output;
and
identifying a predetermined number of suspicious data elements included in the subset of suspicious data elements, said predetermined number of suspicious data element that rank highest based on the probability, said predetermined number of suspicious data elements comprising a group of data elements most impactful on the SAR output vis-à-vis the plurality of suspicious data elements.

17 more elements

A method for conserving resources while identifying a subset of inputs, from among a set of inputs, that most contributed to an output, the method comprising:

receiving a characterization output characterizing a first data structure;
determining a plurality of most important contributing factors to the characterizing the first data structure, the determining comprising:
identifying a plurality of data elements associated with the first data structure at a monitoring data unit;
feeding the plurality of data elements into a plurality of models;
processing the plurality of data elements at the plurality of models;
identifying a plurality of outputs from the plurality of models;
feeding the plurality of outputs into an event processor;
processing the plurality of outputs at the event processor;
grouping the plurality of outputs into a plurality of events at the event processor;
inputting the plurality of events into a determination processor;
determining, at the determination processor, a probability of the first data structure being associated with the characterization output;
determining, at the determination processor, that the determined probability is over a threshold probability of being associated with the characterization output;
in order to identify a predetermined number of data elements, included in the plurality of data elements, contributing to the characterization output:
multiplying the integrated gradient of the determination processor with respect to the plurality of outputs by (the integrated gradient of the event processor with respect to the plurality of data elements divided by the plurality of outputs), which results in a vector of:
a subset of the plurality of data elements;
and
a probability that each data element, included in the subset of data elements, contributed to the characterization output;
ranking the subset of the plurality of data elements of the vector by the probability that each data element contributed to the characterization output;
and
identifying a top predetermined number of data elements from the ranked subset of the plurality of data elements, the top predetermined number of data elements being the plurality of most important contributing factors to the characterizing the first data structure;
and
wherein an equation for determining the integrated gradient of the event processor with respect to the plurality of data elements is:
IG
W
(
x
)
=
∫
t
0
t
f
∂
W
∂
x
⁢
dx
dt
⁢
dt
.

40 more elements

A method for conserving resources while identifying a subset of inputs, from among a set of inputs, that most contributed to an output, the method comprising:

receiving a characterization output characterizing a first data structure;
determining a plurality of most important contributing factors to the characterizing the first data structure, the determining comprising:
identifying a plurality of data elements associated with the first data structure at a monitoring data unit;
feeding the plurality of data elements into a plurality of models;
processing the plurality of data elements at the plurality of models;
identifying a plurality of outputs from the plurality of models;
feeding the plurality of outputs into an event processor;
processing the plurality of outputs at the event processor;
grouping the plurality of outputs into a plurality of events at the event processor;
inputting the plurality of events into a determination processor;
determining, at the determination processor, a probability of the first data structure being associated with the characterization output;
determining, at the determination processor, that the determined probability is over a threshold probability of being associated with the characterization output;
in order to identify a predetermined number of data elements, included in the plurality of data elements, contributing to the characterization output:
multiplying the integrated gradient of the determination processor with respect to the plurality of outputs by (the integrated gradient of the event processor with respect to the plurality of data elements divided by the plurality of outputs), which results in a vector of:
a subset of the plurality of data elements;
and
a probability that each data element, included in the subset of data elements, contributed to the characterization output;
ranking the subset of the plurality of data elements of the vector by the probability that each data element contributed to the characterization output;
and
identifying a top predetermined number of data elements from the ranked subset of the plurality of data elements, the top predetermined number of data elements being the plurality of most important contributing factors to the characterizing the first data structure;
and
wherein an equation for determining the integrated gradient of the determination processor with respect to the plurality of events is:
IGdetermination processor(plurality of events)=∫totfd(plurality of events).

20 more elements

Dependencies are inferred from the claim text.

Facts

Application: 18/418,502Filed 22 Jan 2024 · continuation
Priority: 3 Dec 2021
Published: US 2024/0160726 A116 May 2024
Granted: 15 Oct 2024
Term ends: Not estimatedThe source records don’t support an estimate.

Classification8 codes

G06F 21/552
Detecting local intrusion or implementing counter-measures: involving long-term monitoring or reporting
G06F 18/2193
Validation; Performance evaluation; Active pattern learning techniques: based on specific statistical tests
G06F 21/577
Certifying or maintaining trusted computer platforms, e.g. secure boots or power-downs, version controls, system software checks, secure updates…:…
G06N 3/045
Architecture, e.g. interconnection topology: combinations of networks

Family

Application17/541,428Parent of this continuation · filed 3 Dec 2021
US 2024/0160726 A1Pre-grant publication · published 16 May 2024
US 11,928,209 B2Parent of this continuation
US 2023/0177150 A1Published application
US 12,118,079 B2This patent

From the US filing history. Foreign family members aren’t listed.

Timeline

Citations

Open elsewhere

Building an agent or pulling many patents? Use the API or MCP server instead of scraping pages. Get an API key