# US 2023/0267276 A1 · System and method for determining context changes in text

> Agents: this data is available as JSON from the patentgrep API and MCP server — see https://patentgrep.com/llms.txt

- Number: US 2023/0267276 A1
- Document ID: US20230267276A1
- Status: Published application
- Kind: A1
- Application number: 17/677,259
- Filed: 2022-02-22
- Published: 2023-08-24
- Earliest claimed: 2022-02-22
- Assignees: Bank of America Corporation
- Applicants: Bank of America Corporation
- Inventors: Abhishek Trivedi; Shruti Nandini Thakur; Nimish Ravindra Deshpande; Prashant Khare
- Web page: https://patentgrep.com/patent/US20230267276A1

## CPC

- G06F 40/30: Semantic analysis
- G06F 40/166: Editing, e.g. inserting or deleting
- G06N 3/04: Architecture, e.g. interconnection topology
- G06N 3/065: Analogue means
- G06N 3/084: Backpropagation, e.g. using gradient descent
- G06N 3/09: Supervised learning
- G06F 40/205: Parsing
- G06F 40/289: Phrasal analysis, e.g. finite state techniques or chunking
- G06F: ELECTRIC DIGITAL DATA PROCESSING (computer systems based on specific computational models G06N)
- G06N: COMPUTING ARRANGEMENTS BASED ON SPECIFIC COMPUTATIONAL MODELS

## Abstract

A system splits the text into at least a first and a second portions. The system extracts a first context information from the first portion, and a second context information from a second portion in response to feeding the plurality of portions to a first plurality of neuron logic gates. The system compares the first context information with the second context information. If it is determined that the first context information is different from the second context information, the system dynamically activates at least one of a second plurality of neuron logic gates. The system determines an additional information from at least one of the first portion and second portions. The system updates at least one of the first context information and the second context information to include the additional information. The system generates a first output that comprises the updated first context information and the updated second context information.

## Claims

### Claim 1

A system for context building from text comprising:
a memory configured to store text comprising a plurality of sentences; and
a processor, operably coupled to the memory, and configured to:
split the text into a plurality of portions, wherein the plurality of portions comprises a first portion and a second portion;
extract a first context information from the first portion, wherein the first context information comprises a first meaning of the first portion;
extract a second context information from the second portion, wherein the second context information comprises a second meaning of the second portion, wherein:
extracting the first context information and the second context information is in response to feeding the plurality of portions to a first plurality of neuron logic gates;
each of the first plurality of neuron logic gates comprises a circuitry configured to determine context information from each of the plurality of portions;
compare the first context information with the second context information;
determine whether the first context information is different from the second context information;
in response to determining that the first context information is different from the second context information:
dynamically activate at least one of a second plurality of neuron logic gates, wherein the second plurality of neuron logic gates is different from the first plurality of neuron logic gates;
determine, via the at least one of the second plurality of neuron logic gates, an additional information from at least one of the first portion and the second portion, wherein the additional information comprises more detail about the at least one of the first context information and the second context information;
update the at least one of the first context information and the second context information to include the additional information; and
generate a first output that comprises the updated first context information and the updated second context information, wherein the first output indicates that the updated first context information is different from the updated second context information.

### Claim 2 (depends on 1)

The system of claim 1, wherein the processor is further configured to, in response to determining that the first context information is the same as the second context information:
bypass the second plurality of neuron logic gates; and
generate a second output that comprises the first context information and the second context information, wherein the second output indicates that the first context information is the same as the second context information.

### Claim 3 (depends on 1)

The system of claim 1, wherein extracting the first context information from the first portion is in response to feeding the first portion to a first neuron logic gate from among the first plurality of neuron logic gates, wherein the first neuron logic gate comprises a first circuitry configured to determine the first context information from the first portion.

### Claim 4 (depends on 3)

The system of claim 3, wherein extracting the second context information from the second portion is in response to feeding the second portion to a second neuron logic gate from among the first plurality of neuron logic gates, wherein the second neuron logic gate comprises a second circuitry configured to determine the second context information from the second portion.

### Claim 5 (depends on 4)

The system of claim 4, wherein the first neuron logic gate and the second neuron logic gate are operably coupled to each other such that:
the first neuron logic gate shares the first context information of the first portion with the second neuron logic gate; and
the second neuron logic gate shares the second context information of the second portion with the first neuron logic gate.

### Claim 6 (depends on 1)

The system of claim 1, wherein each portion from among the plurality of portions is a sentence from among the plurality of sentences.

### Claim 7 (depends on 1)

The system of claim 1, wherein the processor is further configured to:
feed the first context information to a first context cell, wherein the first context cell comprises a first storage unit, and is configured to store the first context information and the second context information; and
feed the second context information to a second context cell, wherein the second context cell comprises a second storage unit, and is configured to store the second context information and the first context information.

### Claim 8

A method for context building from text comprising:
splitting text into a plurality of portions, wherein the plurality of portions comprises a first portion and a second portion, wherein the text comprises a plurality of sentences;
extracting a first context information from the first portion, wherein the first context information comprises a first meaning of the first portion;
extracting a second context information from the second portion, wherein the second context information comprises a second meaning of the second portion, wherein:
extracting the first context information and the second context information is in response to feeding the plurality of portions to a first plurality of neuron logic gates;
each of the first plurality of neuron logic gates comprises a circuitry configured to determining context information from each of the plurality of portions;
comparing the first context information with the second context information;
determining whether the first context information is different from the second context information;
in response to determining that the first context information is different from the second context information:
dynamically activating at least one of a second plurality of neuron logic gates, wherein the second plurality of neuron logic gates is different from the first plurality of neuron logic gates;
determining, via the at least one of the second plurality of neuron logic gates, an additional information from at least one of the first portion and the second portion, wherein the additional information comprises more detail about the at least one of the first context information and the second context information;
updating the at least one of the first context information and the second context information to include the additional information; and
generating a first output that comprises the updated first context information and the updated second context information, wherein the first output indicates that the updated first context information is different from the updated second context information.

### Claim 8 (depends on 8)

The method of claim 8, wherein each portion from among the plurality of portions is a sentence from among the plurality of sentences.

### Claim 9 (depends on 8)

The method of claim 8, further comprising, in response to determining that the first context information is the same as the second context information:
bypassing the second plurality of neuron logic gates; and
generating a second output that comprises the first context information and the second context information, wherein the second output indicates that the first context information is the same as the second context information.

### Claim 10 (depends on 8)

The method of claim 8, wherein extracting the first context information from the first portion is in response to feeding the first portion to a first neuron logic gate from among the first plurality of neuron logic gates, wherein the first neuron logic gate comprises a first circuitry configured to determine the first context information from the first portion.

### Claim 11 (depends on 10)

The method of claim 10, wherein extracting the second context information from the second portion is in response to feeding the second portion to a second neuron logic gate from among the first plurality of neuron logic gates, wherein the second neuron logic gate comprises a second circuitry configured to determine the second context information from the second portion.

### Claim 12 (depends on 11)

The method of claim 11, wherein the first neuron logic gate and the second neuron logic gate are operably coupled to each other such that:
the first neuron logic gate shares the first context information of the first portion with the second neuron logic gate; and
the second neuron logic gate shares the second context information of the second portion with the first neuron logic gate.

### Claim 14 (depends on 8)

The method of claim 8, further comprising:
feeding the first context information to a first context cell, wherein the first context cell comprises a first storage unit, and is configured to store the first context information and the second context information; and
feeding the second context information to a second context cell, wherein the second context cell comprises a second storage unit, and is configured to store the second context information and the first context information.

### Claim 15

A computer program comprising executable instructions stored in a non-transitory computer-readable medium that when executed by a processor causes the processor to:
split text into a plurality of portions, wherein the plurality of portions comprises a first portion and a second portion, wherein the text comprises a plurality of sentences;
extract a first context information from the first portion, wherein the first context information comprises a first meaning of the first portion;
extract a second context information from the second portion, wherein the second context information comprises a second meaning of the second portion, wherein:
extracting the first context information and the second context information is in response to feeding the plurality of portions to a first plurality of neuron logic gates;
each of the first plurality of neuron logic gates comprises a circuitry configured to determine context information from each of the plurality of portions;
compare the first context information with the second context information;
determine whether the first context information is different from the second context information;
in response to determining that the first context information is different from the second context information:
dynamically activate at least one of a second plurality of neuron logic gates, wherein the second plurality of neuron logic gates is different from the first plurality of neuron logic gates;
determine, via the at least one of the second plurality of neuron logic gates, an additional information from at least one of the first portion and the second portion, wherein the additional information comprises more detail about the at least one of the first context information and the second context information;
update the at least one of the first context information and the second context information to include the additional information; and
generate a first output that comprises the updated first context information and the updated second context information, wherein the first output indicates that the updated first context information is different from the updated second context information.

### Claim 16 (depends on 15)

The computer program of claim 15, wherein the instructions when executed by the processor, further cause the processor to, in response to determining that the first context information is the same as the second context information:
bypass the second plurality of neuron logic gates; and
generate a second output that comprises the first context information and the second context information, wherein the second output indicates that the first context information is the same as the second context information.

### Claim 17 (depends on 15)

The computer program of claim 15, wherein extracting the first context information from the first portion is in response to feeding the first portion to a first neuron logic gate from among the first plurality of neuron logic gates, wherein the first neuron logic gate comprises a first circuitry configured to determine the first context information from the first portion.

### Claim 18 (depends on 17)

The computer program of claim 17, wherein extracting the second context information from the second portion is in response to feeding the second portion to a second neuron logic gate from among the first plurality of neuron logic gates, wherein the second neuron logic gate comprises a second circuitry configured to determine the second context information from the second portion.

### Claim 19 (depends on 18)

The computer program of claim 18, wherein the first neuron logic gate and the second neuron logic gate are operably coupled to each other such that:
the first neuron logic gate shares the first context information of the first portion with the second neuron logic gate; and
the second neuron logic gate shares the second context information of the second portion with the first neuron logic gate.

### Claim 20 (depends on 15)

The computer program of claim 15, wherein each portion from among the plurality of portions is a sentence from among the plurality of sentences.
