# US 2026/0065025 A1 · Machine learning based simulated control evaluation

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

- Number: US 2026/0065025 A1
- Document ID: US20260065025A1
- Status: Published application
- Kind: A1
- Application number: 19/318,321
- Filed: 2025-09-03
- Published: 2026-03-05
- Earliest claimed: 2024-09-04
- Assignees: Interstice Labs, Inc. dba Trustero
- Applicants: Interstice Labs, Inc. dba Trustero
- Inventors: Phillip Liu; Aleksandr Olegovich Gonopolskiy; Martha Raber; Nicholas Adam Martin; Max Hirsch
- Web page: https://patentgrep.com/patent/US20260065025A1

## CPC

- G06N 3/0455: Auto-encoder networks; Encoder-decoder networks
- G06F 21/577: {Assessing vulnerabilities and evaluating computer system security}
- G06F 21/6245: {Protecting personal data, e.g. for financial or medical purposes}
- G06N 3/092: Reinforcement learning
- G06N 3/096: Transfer learning
- G06F 2221/034: Test or assess a computer or a system
- G06N: COMPUTING ARRANGEMENTS BASED ON SPECIFIC COMPUTATIONAL MODELS
- G06F: ELECTRIC DIGITAL DATA PROCESSING (computer systems based on specific computational models G06N)

## Abstract

A system evaluates controls established for an organization. The system receives a description of a set of controls and a set of constraints. The system identifies a set of data sources storing information associated with the set of controls. The system stores vector representations of information obtained from the set of data sources in a vector database. For each control from the set of controls, the system extracts supporting information related to the control from the vector database based on vector distances. The system generates a structured query input describing the control and the supporting information, and inputs the structured query input into the trained transformer-based neural network. The system processes the output sequence generated by the trained transformer-based neural network to obtain a result indicating whether the control satisfies a constraint from the set of constraints. The system evaluates the set of controls based on the results.

## Claims

### Claim 1

A computer-implemented method, comprising:
receiving description of a set of controls and a set of constraints;
identifying a set of data sources storing information associated with the set of controls;
storing in a vector database, vector representations of information obtained from the set of data sources;
for each control from the set of controls:
extracting supporting information related to the control from the vector database based on vector distances,
generating a structured query input describing the control and the supporting information, the structured query input formatted according to a pre-defined schema for interaction with a trained transformer-based neural network,
inputting the structured query input into the trained transformer-based neural network, and
processing an output sequence generated by the trained transformer-based neural network to obtain a result indicating whether the control satisfies a constraint from the set of constraints;
generate evaluation of the set of controls based on results obtained for the set of controls; and
performing a target action based on the evaluation of the set of controls.

### Claim 2 (depends on 1)

The computer-implemented method of claim 1, further comprising:
executing a series of classifiers to identify tests applicable to the set of controls, and
executing the tests identified.

### Claim 3 (depends on 1)

The computer-implemented method of claim 1, wherein a data source provides one or more screenshots of user interfaces of applications, wherein the vector database stores a vector representation of the one or more screenshots.

### Claim 4 (depends on 1)

The computer-implemented method of claim 1, wherein a data source stores an audio recording, wherein the vector database stores a vector representation of a transcript of the audio recording.

### Claim 5 (depends on 1)

The computer-implemented method of claim 1, further comprising:
sampling a set of entities, and
cross-referencing entities from the set of entities against support information.

### Claim 6 (depends on 1)

The computer-implemented method of claim 1, wherein the trained transformer-based neural network has been trained on a corpus of domain-specific data using supervised fine-tuning and reinforcement learning.

### Claim 7 (depends on 1)

The computer-implemented method of claim 1, wherein processing an output sequence generated by the trained transformer-based neural network comprises applying post-generation parsing rules to extract data fields and generate an interpretable response.

### Claim 8

A non-transitory computer readable storage medium storing instructions that when executed by one or more computer processors, cause the one or more computer processors to perform steps comprising:
receiving description of a set of controls and a set of constraints;
identifying a set of data sources storing information associated with the set of controls;
storing in a vector database, vector representations of information obtained from the set of data sources;
for each control from the set of controls:
extracting supporting information related to the control from the vector database based on vector distances,
generating a structured query input describing the control and the supporting information, the structured query input formatted according to a pre-defined schema for interaction with a trained transformer-based neural network,
inputting the structured query input into the trained transformer-based neural network, and
processing an output sequence generated by the trained transformer-based neural network to obtain a result indicating whether the control satisfies a constraint from the set of constraints;
generate evaluation of the set of controls based on results obtained for the set of controls; and
performing a target action based on the evaluation of the set of controls.

### Claim 9 (depends on 8)

The non-transitory computer readable storage medium of claim 8, wherein the instructions when executed by one or more computer processors, further cause the one or more computer processors to perform steps comprising:
executing a series of classifiers to identify tests applicable to the set of controls, and
executing the tests identified.

### Claim 10 (depends on 8)

The non-transitory computer readable storage medium of claim 8, wherein a data source provides one or more screenshots of user interfaces of applications, wherein the vector database stores a vector representation of the one or more screenshots.

### Claim 11 (depends on 8)

The non-transitory computer readable storage medium of claim 8, wherein a data source stores an audio recording, wherein the vector database stores a vector representation of a transcript of the audio recording.

### Claim 12 (depends on 8)

The non-transitory computer readable storage medium of claim 8, wherein the instructions when executed by one or more computer processors, further cause the one or more computer processors to perform steps comprising:
sampling a set of entities, and
cross-referencing entities from the set of entities against support information.

### Claim 13 (depends on 8)

The non-transitory computer readable storage medium of claim 8, wherein the trained transformer-based neural network has been trained on a corpus of domain-specific data using supervised fine-tuning and reinforcement learning.

### Claim 14 (depends on 8)

The non-transitory computer readable storage medium of claim 8, wherein processing an output sequence generated by the trained transformer-based neural network comprises applying post-generation parsing rules to extract data fields and generate an interpretable response.

### Claim 15

A computer system comprising:
one or more computer processors; and
a non-transitory computer readable storage medium storing instructions that when executed by the one or more computer processors, cause the one or more computer processors to perform steps comprising:
receiving description of a set of controls and a set of constraints;
identifying a set of data sources storing information associated with the set of controls;
storing in a vector database, vector representations of information obtained from the set of data sources;
for each control from the set of controls:
extracting supporting information related to the control from the vector database based on vector distances,
generating a structured query input describing the control and the supporting information, the structured query input formatted according to a pre-defined schema for interaction with a trained transformer-based neural network,
inputting the structured query input into the trained transformer-based neural network, and
processing an output sequence generated by the trained transformer-based neural network to obtain a result indicating whether the control satisfies a constraint from the set of constraints;
generate evaluation of the set of controls based on results obtained for the set of controls; and
performing a target action based on the evaluation of the set of controls.

### Claim 16 (depends on 15)

The computer system of claim 15, wherein the instructions when executed by one or more computer processors, further cause the one or more computer processors to perform steps comprising:
executing a series of classifiers to identify tests applicable to the set of controls, and
executing the tests identified.

### Claim 17 (depends on 15)

The computer system of claim 15, wherein a data source provides one or more screenshots of user interfaces of applications, wherein the vector database stores a vector representation of the one or more screenshots.

### Claim 18 (depends on 15)

The computer system of claim 15, wherein a data source stores an audio recording, wherein the vector database stores a vector representation of a transcript of the audio recording.

### Claim 19 (depends on 15)

The computer system of claim 15, wherein the instructions when executed by one or more computer processors, further cause the one or more computer processors to perform steps comprising:
sampling a set of entities, and
cross-referencing entities from the set of entities against support information.

### Claim 20 (depends on 15)

The computer system of claim 15, wherein the trained transformer-based neural network has been trained on a corpus of domain-specific data using supervised fine-tuning and reinforcement learning.
