# US 2021/0240752 A1 · Schema-Informed Extraction for Unstructured Data

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

- Number: US 2021/0240752 A1
- Document ID: US20210240752A1
- Status: Published application
- Kind: A1
- Application number: 16/779,833
- Filed: 2020-02-03
- Published: 2021-08-05
- Earliest claimed: 2020-02-03
- Assignees: S&P Global Inc.
- Applicants: S&P Global
- Inventors: Chester Curme; Delphine Vendryes; Baojia Tong; Matthew Theisen; David Relyea
- Web page: https://patentgrep.com/patent/US20210240752A1

## CPC

- G06F 16/3344: {using natural language analysis}
- G06F 16/3329: Natural language query formulation
- G06F 16/3334: {Selection or weighting of terms from queries, including natural language queries}
- G06F 40/205: Parsing
- G06F 40/30: Semantic analysis
- G06F 40/177: of tables; using ruled lines
- G06F: ELECTRIC DIGITAL DATA PROCESSING (computer systems based on specific computational models G06N)

## Abstract

A method of extracting data from documents is provided. The method comprises receiving input of a number of documents and input of a schema of data items available for extraction from the documents. The documents are parsed into a machine-readable representation, and data items in the machine-readable representation are identified according to the schema. Interpretations of data items are propagated within the documents to disambiguate identified data items, and identified data items are matched with other data items in the documents according to the schema. Only identified data items that include a minimal set of interpretations specified by the schema are extracted.

## Claims

### Claim 1

A computer-implemented method of extracting data from documents, the method comprising:
using one or more processors to perform the steps of:
receiving input of a number of documents;
receiving input of a schema of data items available for extraction from the documents;
parsing the documents into a machine-readable representation;
identifying data items in the machine-readable representation according to the schema;
propagating interpretations of data items within the documents to disambiguate identified data items;
matching identified data items with other data items in the documents according to the schema; and
extracting only identified data items that include a minimal set of interpretations specified by the schema.

### Claim 2 (depends on 1)

The method of claim 1, wherein the schema comprises known attributes of data items, known aliases of data items, and specified relationships between data items.

### Claim 3 (depends on 2)

The method of claim 2, where in identified data items are matched with other data according to the specified relationships in the schema.

### Claim 4 (depends on 1)

The method of claim 1, wherein parsing the documents comprises:
text parsing;
line parsing; and
image parsing.

### Claim 5 (depends on 4)

The method of claim 4, further comprising detecting and reconstructing tables from text, lines, and images.

### Claim 6 (depends on 4)

The method of claim 4, further comprising detecting and reconstructing charts from text, lines, and images.

### Claim 7 (depends on 1)

The method of claim 1, wherein, for tabular data, interpretations of data items are propagated across rows and down columns.

### Claim 8 (depends on 1)

The method of claim 1, wherein, for floating text, interpretation of a matched data item to a corresponding value is propagated according to geometric and semantic information.

### Claim 9

A system for extracting data from documents, the system comprising:
a storage device configured to store program instructions; and
one or more processors operably connected to the storage device and configured to execute the program instructions to cause the system to:
receive input of a number of documents;
receive input of a schema of data items available for extraction from the documents;
parse the documents into a machine-readable representation;
identify data items in the machine-readable representation according to the schema;
propagate interpretations of data items within the documents to disambiguate identified data items;
match identified data items with other data items in the documents according to the schema; and
extract only identified data items that include a minimal set of interpretations specified by the schema.

### Claim 10 (depends on 9)

The system of claim 9, wherein the schema comprises known attributes of data items, known aliases of data items, and specified relationships between data items.

### Claim 11 (depends on 10)

The system of claim 10, where in identified data items are matched with other data according to the specified relationships in the schema.

### Claim 12 (depends on 9)

The system of claim 9, wherein parsing the documents comprises:
text parsing;
line parsing; and
image parsing.

### Claim 13 (depends on 12)

The system of claim 12, further comprising detecting and reconstructing tables from text, lines, and images.

### Claim 14 (depends on 12)

The system of claim 12, further comprising detecting and reconstructing charts from text, lines, and images.

### Claim 15 (depends on 9)

The system of claim 9, wherein, for tabular data, interpretations of data items are propagated across rows and down columns.

### Claim 16 (depends on 9)

The system of claim 9, wherein, for floating text, interpretation of a matched data item to a corresponding value is propagated according to geometric and semantic information.

### Claim 17

A computer program product for extracting data from documents, the computer program product comprising:
a non-volatile computer readable storage medium having program instructions stored thereon to perform the steps of:
receiving input of a number of documents;
receiving input of a schema of data items available for extraction from the documents;
parsing the documents into a machine-readable representation;
identifying data items in the machine-readable representation according to the schema;
propagating interpretations of data items within the documents to disambiguate identified data items;
matching identified data items with other data items in the documents according to the schema; and
extracting only identified data items that include a minimal set of interpretations specified by the schema.

### Claim 18 (depends on 17)

The computer program product of claim 17, wherein the schema comprises known attributes of data items, known aliases of data items, and specified relationships between data items.

### Claim 19 (depends on 18)

The computer program product of claim 18, wherein identified data items are matched with other data according to the specified relationships in the schema.

### Claim 20 (depends on 17)

The computer program product of claim 17, wherein parsing the documents comprises:
text parsing;
line parsing; and
image parsing.

### Claim 21 (depends on 20)

The computer program product of claim 20, further comprising detecting and reconstructing tables from text, lines, and images.

### Claim 22 (depends on 20)

The computer program product of claim 20, further comprising detecting and reconstructing charts from text, lines, and images.

### Claim 23 (depends on 17)

The computer program product of claim 17, wherein, for tabular data, interpretations of data items are propagated across rows and down columns.

### Claim 24 (depends on 17)

The computer program product of claim 17, wherein, for floating text, interpretation of a matched data item to a corresponding value is propagated according to geometric and semantic information.
