# US 2025/0156439 A1 · Optimization Engine in a Structured and Unstructured Data System

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

- Number: US 2025/0156439 A1
- Document ID: US20250156439A1
- Status: Published application
- Kind: A1
- Application number: 18/389,436
- Filed: 2023-11-14
- Published: 2025-05-15
- Earliest claimed: 2023-11-14
- Assignees: The Boston Consulting Group, Inc.
- Applicants: The Boston Consulting Group, Inc.
- Inventors: Pavel Erofeev; Ning Yang; Marcin Druzkowski; Ronny Fehling
- Web page: https://patentgrep.com/patent/US20250156439A1

## CPC

- G06F 16/258: {Data format conversion from or to a database}
- G06F 16/285: {Clustering or classification}
- G06F: ELECTRIC DIGITAL DATA PROCESSING (computer systems based on specific computational models G06N)

## Abstract

Disclosed are techniques that generate a structural representation of a plurality of documents, the structural representation including a plurality of nodes and a plurality of edges, with the plurality of nodes being representations of the plurality of documents and the plurality of edges representing a feature in common between nodes of the plurality of nodes, with each node holding a vector of confidence values for weak models on a current optimization step and a weighted prediction for each of the weak models, generate a local ensemble model from the structural representation of the plurality of documents combined with the weighted prediction of the weak models, with the generated local ensemble model having a higher predictive power than any weak model individually, and generate a label for each node based on the local ensemble model.

## Claims

### Claim 1

A computer implemented method comprises:
generating by a computer system, a structural representation of a plurality of documents, the structural representation including a plurality of nodes and a plurality of edges, with the plurality of nodes being representations of the plurality of documents and the plurality of edges representing a feature in common between nodes of the plurality of nodes, with each node holding a vector of confidence values for weak models on a current optimization step and a weighted prediction for each of the weak models;
generating by the computer system, a local ensemble model from the structural representation of the plurality of documents combined with the weighted prediction of the weak models, with the generated local ensemble model having a higher predictive power than any weak model individually; and
generating by the computer system, a label for each node based on the local ensemble model.

### Claim 2 (depends on 1)

The method of claim 1 wherein each edge has a type and each type of an edge represents a feature of the plurality of documents.

### Claim 3 (depends on 2)

The method of claim 2 wherein the features of the plurality of documents is a meta-feature or an associated target metric, or processed unstructured data.

### Claim 4 (depends on 1)

The method of claim 1 wherein generating a label for each node provides either a manually confirmed and/or corrected label that is assumed to be a true label or a resulting weighted prediction of the weak models.

### Claim 5 (depends on 1)

The method of claim 1 wherein confidence values vary from 0 to 1 and the confidence values represent uncertainty in predictions made by the weak models.

### Claim 6 (depends on 1)

The method of claim 1 wherein the confidence values are probability estimates or normalized relative distances to known exemplars of a class of the weak models.

### Claim 7 (depends on 6)

The method of claim 6 wherein each node holds a set of weights for each of the weak models.

### Claim 8 (depends on 7)

The method of claim 7 wherein a sum of the weights is equal to a value of 1.

### Claim 9

A computing system comprises:
one or more processor devices;
memory operatively coupled to the one or more processor devices; and
computer storage that stores executable computer instructions to cause the one or more processor devices to:
generate a structural representation of a plurality of documents, the structural representation including a plurality of nodes and a plurality of edges, with the plurality of nodes being representations of the plurality of documents and the plurality of edges representing a feature in common between nodes of the plurality of nodes, with each node holding a vector of confidence values for weak models on a current optimization step and a weighted prediction for each of the weak models;
generate a local ensemble model from the structural representation of the plurality of documents combined with the weighted prediction of the weak models, with the generated local ensemble model having a higher predictive power than any weak model individually; and
generate a label for each node based on the local ensemble model.

### Claim 10 (depends on 9)

The computing system of claim 9 wherein each edge has a type and each type of an edge represents a feature of the plurality of documents.

### Claim 11 (depends on 10)

The computing system of claim 10 wherein the features of the plurality of documents is a meta-feature or an associated target metric, or processed unstructured data.

### Claim 12 (depends on 9)

The computing system of claim 9 wherein generating a label for each node provides either a manually confirmed and/or corrected label that is assumed to be a true label or a resulting weighted prediction of the weak models.

### Claim 13 (depends on 9)

The computing system of claim 9 wherein confidence values vary from 0 to 1 and the confidence values represent uncertainty in predictions made by the weak models.

### Claim 14 (depends on 9)

The computing system of claim 9 wherein the confidence values are probability estimates or normalized relative distances to known exemplars of a class of the weak models.

### Claim 15 (depends on 14)

The computing system of claim 14 wherein each node holds a set of weights for each of the weak models.

### Claim 16 (depends on 15)

The computing system of claim 15 wherein a sum of the weights is equal to a value of 1.

### Claim 17

A computer program product tangibly stored on a non-transitory computer readable medium, the computer program product comprising instructions to cause one or more processor devices to:
generate a structural representation of a plurality of documents, the structural representation including a plurality of nodes and a plurality of edges, with the plurality of nodes being representations of the plurality of documents and the plurality of edges representing a feature in common between nodes of the plurality of nodes, with each node holding a vector of confidence values for weak models on a current optimization step and a weighted prediction for each of the weak models;
generate a local ensemble model from the structural representation of the plurality of documents combined with the weighted prediction of the weak models, with the generated local ensemble model having a higher predictive power than any weak model individually; and
generate a label for each node based on the local ensemble model.

### Claim 18 (depends on 17)

The computer program product of claim 17 wherein each edge has a type and each type of an edge represents a feature of the plurality of documents.

### Claim 19 (depends on 18)

The computer program product of claim 18 wherein the features of the plurality of documents is a meta-feature or an associated target metric, or processed unstructured data.

### Claim 20 (depends on 17)

The computer program product of claim 17 wherein generating a label for each node provides either a manually confirmed and/or corrected label that is assumed to be a true label or a resulting weighted prediction of the weak models.

### Claim at position 21

21-60. (canceled)
