# US 12,602,374 B2 · Data acquisition method and apparatus, computer device and storage medium

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

- Number: US 12,602,374 B2
- Document ID: US12602374B2
- Status: Granted
- Kind: B2
- Application number: 19/052,975
- Filed: 2025-02-13
- Published: 2025-08-28
- Granted: 2026-04-14
- Earliest claimed: 2024-02-22
- Assignees: Beijing Volcano Engine Technology Co., Ltd.
- Applicants: Beijing Volcano Engine Technology Co., Ltd.
- Inventors: Wenjing Chen; Shuqing He; Guoxiang Zhao; Hongjian Peng; Yuan Song; Xiaowei Chen; Fangyuan Zhou; Yu Chen; Mingjun Tang; Huixiang Wang
- Web page: https://patentgrep.com/patent/US12602374B2

## CPC

- G06F 16/243: {Natural language query formulation}
- G06F 16/24522: {Translation of natural language queries to structured queries}
- G06F 16/24573: {using data annotations, e.g. user-defined metadata}
- G06F 16/248: Presentation of query results
- G06F 40/20: Natural language analysis (semantic analysis of natural language G06F40/30)
- G06F 40/279: Recognition of textual entities
- G06F 40/30: Semantic analysis
- G06N 3/00: Computing arrangements based on biological models
- G06N 3/044: Recurrent networks, e.g. Hopfield networks
- G06N 3/045: Combinations of networks
- G06N 3/08: Learning methods
- G06N 20/00: Machine learning
- G06F: ELECTRIC DIGITAL DATA PROCESSING (computer systems based on specific computational models G06N)
- G06N: COMPUTING ARRANGEMENTS BASED ON SPECIFIC COMPUTATIONAL MODELS

## Abstract

The present disclosure relates to the field of data processing technology, and discloses a data acquisition method and apparatus, a computer device and a storage medium. The method includes: acquiring question information; determining task information corresponding to the question information using a pre-trained language model, where the task information comprises a keyword in the question information and a target data source type; obtaining a search result by searching, using the pre-trained language model and according to a prompt corresponding to the task information, an associated data source in at least one candidate data source; and determining at least one data source to be provided from the at least one candidate data source according to the search result, and providing metadata of the at least one data source to be provided to the user.

## Claims

### Claim 1

A data acquisition method, comprising:
acquiring question information, wherein the question information indicates a need to search an associated data source, and the associated data source is a data source related to data that a user needs;
determining task information corresponding to the question information using a pre-trained language model, wherein the task information comprises a keyword in the question information and a target data source type;
obtaining a search result by searching, using the pre-trained language model and according to a prompt corresponding to the task information, an associated data source in at least one candidate data source, wherein the prompt is used to prompt the pre-trained language model to search the associated data source according to the task information corresponding to the question information;
determining at least one data source to be provided from the at least one candidate data source according to the search result;
transmitting metadata of the at least one data source to be provided to a user terminal to cause the user terminal to display the metadata;
receiving data usage indication information, wherein the data usage indication information indicates performing an operation by using a field of at least one data source; and
determining, using the pre-trained language model and according to the data usage indication information, at least one target data source, a target field of the at least one target data source, and a target operation for the target field of the at least one target data source;
wherein determining, using the pre-trained language model and according to the data usage indication information, the at least one target data source, the target field of the at least one target data source, and the target operation for the target field of the at least one target data source comprises:
in response to determining, using the pre-trained language model, that a data source whose name matches a sentence describing a data source in the data usage indication information exists in a plurality of data sources, determining the data source whose name matches the sentence describing the data source as a target data source corresponding to the sentence describing the data source.

### Claim 2 (depends on 1)

The method according to claim 1, wherein determining the at least one data source to be provided from the at least one candidate data source according to the search result comprises:
in response to the search result comprising an associated data source of the target data source type, determining the associated data source of the target data source type as the data source to be provided; and
in response to the search result not comprising the associated data source of the target data source type, selecting the data source to be provided from the at least one candidate data source according to a degree of association between metadata of the at least one candidate data source and the question information.

### Claim 3 (depends on 1)

The method according to claim 1, wherein the method further comprises:
in response to the search result comprising the associated data source of the target data source type, generating, using the pre-trained language model and according to the prompt corresponding to the task information, summary information of the associated data source of the target data source type, wherein the summary information comprises introduction information indicating a feature of the associated data source of the target data source type, and usage information indicating a usage of the associated data source of the target data source type; and
providing the summary information to the user.

### Claim 4 (depends on 1)

The method according to claim 1, wherein the method further comprises:
calculating a degree of association between metadata of each of a plurality of data sources and the question information; and
determining a data source, from the plurality of data sources, with a degree of association between metadata and the question information being greater than an association degree threshold, as the candidate data source.

### Claim 5 (depends on 1)

The method according to claim 1, wherein acquiring the question information comprises:
obtaining the question information by rewriting, using the pre-trained language model, at least one question sentence input by the user.

### Claim 6 (depends on 1)

The method according to claim 1, wherein the method further comprises:
obtaining an operation result by performing the target operation on the target field of the at least one target data source, and providing the operation result to the user.

### Claim 7 (depends on 6)

The method according to claim 6, wherein determining, using the pre-trained language model and according to the data usage indication information, the at least one target data source, the target field of the at least one target data source, and the target operation for the target field of the at least one target data source comprises:
in response to determining, using the pre-trained language model, that the data source whose name matches the sentence describing the data source does not exist in the plurality of data sources, generating, using the pre-trained language model, a sentence indicating a need to search a data source described by the sentence describing the data source; determining task information corresponding to the generated sentence, and searching, using the pre-trained language model and according to a prompt corresponding to the task information corresponding to the generated sentence, a target data source corresponding to the sentence describing the data source from the plurality of data sources.

### Claim 8

A computer device, comprising:
a memory and a processor, the memory and the processor communicatively connected with each other, the memory storing computer instructions, and the computer instructions, when executed by the processor, causing the computer device to:
acquire question information, wherein the question information indicates a need to search an associated data source, and the associated data source is a data source related to data that a user needs;
determine task information corresponding to the question information using a pre-trained language model, wherein the task information comprises a keyword in the question information and a target data source type;
obtain a search result by searching, using the pre-trained language model and according to a prompt corresponding to the task information, an associated data source in at least one candidate data source, wherein the prompt is used to prompt the pre-trained language model to search the associated data source according to the task information corresponding to the question information;
determine at least one data source to be provided from the at least one candidate data source according to the search result;
transmit metadata of the at least one data source to be provided to a user terminal to cause the user terminal to display the metadata;
receive data usage indication information, wherein the data usage indication information indicates performing an operation by using a field of at least one data source; and
determine, using the pre-trained language model and according to the data usage indication information, at least one target data source, a target field of the at least one target data source, and a target operation for the target field of the at least one target data source;
wherein the computer instructions causing the computer device to determine, using the pre-trained language model and according to the data usage indication information, the at least one target data source, the target field of the at least one target data source, and the target operation for the target field of the at least one target data source further cause the computer device to:
in response to determining, using the pre-trained language model, that a data source whose name matches a sentence describing a data source in the data usage indication information exists in a plurality of data sources, determine the data source whose name matches the sentence describing the data source as a target data source corresponding to the sentence describing the data source.

### Claim 9 (depends on 8)

The computer device according to claim 8, wherein the computer instructions causing the computer device to determine the at least one data source to be provided from the at least one candidate data source according to the search result further cause the computer device to:
in response to the search result comprising an associated data source of the target data source type, determine the associated data source of the target data source type as the data source to be provided; and
in response to the search result not comprising the associated data source of the target data source type, select the data source to be provided from the at least one candidate data source according to a degree of association between metadata of the at least one candidate data source and the question information.

### Claim 10 (depends on 8)

The computer device according to claim 8, wherein the computer instructions further cause the computer device to:
in response to the search result comprising the associated data source of the target data source type, generate, using the pre-trained language model and according to the prompt corresponding to the task information, summary information of the associated data source of the target data source type, wherein the summary information comprises introduction information indicating a feature of the associated data source of the target data source type, and usage information indicating a usage of the associated data source of the target data source type; and
provide the summary information to the user.

### Claim 11 (depends on 8)

The computer device according to claim 8, wherein the computer instructions further cause the computer device to:
calculate a degree of association between metadata of each of a plurality of data sources and the question information; and
determine a data source, from the plurality of data sources, with a degree of association between metadata and the question information being greater than an association degree threshold, as the candidate data source.

### Claim 12 (depends on 8)

The computer device according to claim 8, wherein the computer instructions causing the computer device to acquire the question information further cause the computer device to:
obtain the question information by rewriting, using the pre-trained language model, at least one question sentence input by the user.

### Claim 13 (depends on 8)

The computer device according to claim 8, wherein the computer instructions further cause the computer device to:
obtain an operation result by performing the target operation on the target field of the at least one target data source, and providing the operation result to the user.

### Claim 14 (depends on 13)

The computer device according to claim 13, wherein the computer instructions causing the computer device to determine, using the pre-trained language model and according to the data usage indication information, the at least one target data source, the target field of the at least one target data source, and the target operation for the target field of the at least one target data source further cause the computer device to:
in response to determining, using the pre-trained language model, that the data source whose name matches the sentence describing the data source does not exist in the plurality of data sources, generate, using the pre-trained language model, a sentence indicating a need to search a data source described by the sentence describing the data source; determine task information corresponding to the generated sentence, and search, using the pre-trained language model and according to a prompt corresponding to the task information corresponding to the generated sentence, a target data source corresponding to the sentence describing the data source from the plurality of data sources.

### Claim 15

A non-transitory computer-readable storage medium, wherein the computer-readable storage medium stores computer instructions, and the computer instructions are configured to cause a computer to:
acquire question information, wherein the question information indicates a need to search an associated data source, and the associated data source is a data source related to data that a user needs;
determine task information corresponding to the question information using a pre-trained language model, wherein the task information comprises a keyword in the question information and a target data source type;
obtain a search result by searching, using the pre-trained language model and according to a prompt corresponding to the task information, an associated data source in at least one candidate data source, wherein the prompt is used to prompt the pre-trained language model to search the associated data source according to the task information corresponding to the question information;
determine at least one data source to be provided from the at least one candidate data source according to the search result;
transmit metadata of the at least one data source to be provided to a user terminal to cause the user terminal to display the metadata;
receive data usage indication information, wherein the data usage indication information indicates performing an operation by using a field of at least one data source; and
determine, using the pre-trained language model and according to the data usage indication information, at least one target data source, a target field of the at least one target data source, and a target operation for the target field of the at least one target data source;
wherein the computer instructions causing the computer to determine, using the pre-trained language model and according to the data usage indication information, the at least one target data source, the target field of the at least one target data source, and the target operation for the target field of the at least one target data source further cause the computer to:
in response to determining, using the pre-trained language model, that a data source whose name matches a sentence describing a data source in the data usage indication information exists in a plurality of data sources, determine the data source whose name matches the sentence describing the data source as a target data source corresponding to the sentence describing the data source.

### Claim 16 (depends on 15)

The medium according to claim 15, wherein the computer instructions causing the computer to determine the at least one data source to be provided from the at least one candidate data source according to the search result further cause the computer to:
in response to the search result comprising an associated data source of the target data source type, determine the associated data source of the target data source type as the data source to be provided; and
in response to the search result not comprising the associated data source of the target data source type, select the data source to be provided from the at least one candidate data source according to a degree of association between metadata of the at least one candidate data source and the question information.

### Claim 17 (depends on 15)

The medium according to claim 15, wherein the computer instructions further cause the computer to:
in response to the search result comprising the associated data source of the target data source type, generate, using the pre-trained language model and according to the prompt corresponding to the task information, summary information of the associated data source of the target data source type, wherein the summary information comprises introduction information indicating a feature of the associated data source of the target data source type, and usage information indicating a usage of the associated data source of the target data source type; and
provide the summary information to the user.

### Claim 18 (depends on 15)

The medium according to claim 15, wherein the computer instructions further cause the computer to:
calculate a degree of association between metadata of each of a plurality of data sources and the question information; and
determine a data source, from the plurality of data sources, with a degree of association between metadata and the question information being greater than an association degree threshold, as the candidate data source.

### Claim 19 (depends on 15)

The medium according to claim 15, wherein the computer instructions causing the computer to acquire the question information further cause the computer to:
obtain the question information by rewriting, using the pre-trained language model, at least one question sentence input by the user.

### Claim 20 (depends on 15)

The medium according to claim 15, wherein the computer instructions further cause the computer to:
obtain an operation result by performing the target operation on the target field of the at least one target data source, and providing the operation result to the user.