# US 2019/0095537 A1 · Method and apparatus for expanding query

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

- Number: US 2019/0095537 A1
- Document ID: US20190095537A1
- Status: Published application
- Kind: A1
- Application number: 16/052,414
- Filed: 2018-08-01
- Published: 2019-03-28
- Earliest claimed: 2017-09-22
- Assignees: Baidu Online Network Technology (Beijing) Co., Ltd.
- Applicants: Baidu Online Network Technology (Beijing) Co., Ltd.
- Inventors: Huiqiang ZHONG; Cunxiang YIN; Jianping SHEN; Guoqiang XU
- Web page: https://patentgrep.com/patent/US20190095537A1

## CPC

- G06F 16/9535: Search customisation based on user profiles and personalisation
- G06F 16/56: having vectorial format
- G06F 16/90328: {using search space presentation or visualization, e.g. category or range presentation and selection}
- G06F: ELECTRIC DIGITAL DATA PROCESSING (computer systems based on specific computational models G06N)

## Abstract

The present disclosure discloses a method and apparatus for expanding a query. The method comprises: calculating an image click characteristic vector of a target query based on an image click log associated with the target query; finding a similar query of the target query based on the image click characteristic vector, to obtain a candidate expansion query set of the target query; matching the target query and each candidate expansion query in the candidate expansion query set with a knowledge graph, to extract an entity word and a qualifier of the target query and an entity word and a qualifier of the each candidate expansion query; expanding the entity words and the qualifiers of the target query and the each candidate expansion query in combination with the knowledge graph; and matching using expansion results of the entity words and the qualifiers, to determine an expanded query of the target query.

## Claims

### Claim 1

A method for expanding a query, comprising:
calculating an image click characteristic vector of a target query based on an acquired image click log associated with the target query;
finding, in a preset query set, a similar query of the target query based on the image click characteristic vector of the target query, to obtain a candidate expansion query set of the target query;
matching the target query and each candidate expansion query in the candidate expansion query set with an entity word and a qualifier in a knowledge graph, to extract an entity word and a qualifier of the target query and an entity word and a qualifier of each candidate expansion query;
expanding the entity word and the qualifier of the target query, and the entity word and the qualifier of each candidate expansion query in combination with the knowledge graph; and
matching using expansion results of the entity word and the qualifier of the target query and the entity word and the qualifier of the each candidate expansion query, to determine an expanded query of the target query from the candidate expansion query set.

### Claim 2 (depends on 1)

The method according to claim 1, wherein the calculating an image click characteristic vector of a target query based on an acquired image click log associated with the target query comprises:
counting numbers of clicks on images associated with the target query in each preset time window based on the acquired image click log associated with the target query; and
sanitizing a number of clicks less than a preset click threshold in a counting result, and constructing the image click characteristic vector based on sanitized numbers of clicks on the images associated with the target query in each preset time window after the sanitizing.

### Claim 3 (depends on 1)

The method according to claim 1, wherein the finding, in a preset query set, a similar query of the target query based on the image click characteristic vector of the target query to obtain a candidate expansion query set of the target query comprises:
counting, for each query in the preset query set, a number of clicks on an image associated with the query to construct an image click characteristic vector of the each query in the preset query set;
calculating a similarity between the image click characteristic vector of the each query in the preset query set and the image click characteristic vector of the target query; and
using a query corresponding to an image click characteristic vector having a similarity to the image click characteristic vector of the target query greater than a preset similarity threshold as a candidate expansion query to form the candidate expansion query set.

### Claim 4 (depends on 1)

The method according to claim 1, wherein the matching the target query and each candidate expansion query in the candidate expansion query set with an entity word and a qualifier in a knowledge graph to extract an entity word and a qualifier of the target query and an entity word and a qualifier of the each candidate expansion query comprises:
performing word segmentation on the target query and the each candidate expansion query in the candidate expansion query set;
matching words obtained by the word segmentation with the entity word in the knowledge graph to obtain the entity word of the target query and the entity word of the each candidate expansion query; and
matching the words obtained by the word segmentation with a qualifier of the entity word in the knowledge graph to obtain the qualifier of the target query and the qualifier of the each candidate expansion query.

### Claim 5 (depends on 1)

The method according to claim 1, wherein the expanding the entity word and the qualifier of the target query, and the entity word and the qualifier of the each candidate expansion query in combination with the knowledge graph comprises:
expanding the entity word of the target query and the entity word of the each candidate expansion query based on a knowledge structure relation between different entity words in the knowledge graph, to obtain an expanded entity word of the target query and an expanded entity word of the each candidate expansion query; and
matching the qualifier of the target query and the qualifier of the each candidate expansion query with an associated qualifier set constructed based on the knowledge structure relation in the knowledge graph, to obtain an expanded qualifier of the target query and an expanded qualifier of the each candidate expansion query.

### Claim 6 (depends on 5)

The method according to claim 5, wherein the knowledge graph comprises knowledge graphs of a plurality of domain categories; and
the expanding the entity word and the qualifier of the target query, and the entity word and the qualifier of the each candidate expansion query in combination with the knowledge graph further comprises:
determining a target domain category of the knowledge graph matching the entity word of the target query and the entity word of the each candidate expansion query,
wherein the expanding the entity word of the target query and the entity word of the each candidate expansion query based on a knowledge structure relation between different entity words in the knowledge graph comprises:
expanding the entity word of the target query and the entity word of the each candidate expansion query based on a knowledge structure relation between different entity words in the knowledge graph of the target domain category; and
the matching the qualifier of the target query and the qualifier of the each candidate expansion query with an associated qualifier set constructed based on the knowledge structure relation in the knowledge graph comprises:
matching the qualifier of the target query and the qualifier of the each candidate expansion query with the associated qualifier set constructed based on the knowledge structure relation in the knowledge graph of the target domain category.

### Claim 7 (depends on 5)

The method according to claim 5, wherein the matching using expansion results of the entity word and the qualifier of the target query and the entity word and the qualifier of the each candidate expansion query to determine an expanded query of the target query from the candidate expansion query set comprises:
using a candidate expansion query having an entity word identical to the expanded entity word of the target query and/or a qualifier identical to the expanded qualifier of the target query in the candidate expansion query set as the expanded query of the target query.

### Claim 8

An apparatus for expanding a query, comprising:
at least one processor; and
a memory storing instructions, the instructions when executed by the at least one processor, cause the at least one processor to perform operations, the operations comprising:
calculating an image click characteristic vector of a target query based on an acquired image click log associated with the target query;
finding, in a preset query set, a similar query of the target query based on the image click characteristic vector of the target query, to obtain a candidate expansion query set of the target query;
matching the target query and each candidate expansion query in the candidate expansion query set with an entity word and a qualifier in a knowledge graph, to extract an entity word and a qualifier of the target query and an entity word and a qualifier of each candidate expansion query;
expanding the entity word and the qualifier of the target query, and the entity word and the qualifier of each candidate expansion query in combination with the knowledge graph; and
matching using expansion results of the entity word and the qualifier of the target query, and the entity word and the qualifier of each candidate expansion query, to determine an expanded query of the target query from the candidate expansion query set.

### Claim 9 (depends on 8)

The apparatus according to claim 8, wherein the calculating an image click characteristic vector of a target query based on an acquired image click log associated with the target query comprises:
counting numbers of clicks on images associated with the target query in each preset time window based on the acquired image click log associated with the target query; and
sanitizing a number of clicks less than a preset click threshold in a counting result, and constructing the image click characteristic vector based on sanitized numbers of clicks on the images associated with the target query in each preset time window after the sanitizing.

### Claim 10 (depends on 8)

The apparatus according to claim 8, wherein the finding, in a preset query set, a similar query of the target query based on the image click characteristic vector of the target query to obtain a candidate expansion query set of the target query comprises:
counting, for each query in the preset query set, a number of clicks on an image associated with the query to construct an image click characteristic vector of the each query in the preset query set;
calculating a similarity between the image click characteristic vector of the each query in the preset query set and the image click characteristic vector of the target query; and
using a query corresponding to an image click characteristic vector having a similarity to the image click characteristic vector of the target query greater than a preset similarity threshold as a candidate expansion query to form the candidate expansion query set.

### Claim 11 (depends on 8)

The apparatus according to claim 8, wherein the matching the target query and each candidate expansion query in the candidate expansion query set with an entity word and a qualifier in a knowledge graph to extract an entity word and a qualifier of the target query and an entity word and a qualifier of the each candidate expansion query comprises:
performing word segmentation on the target query and the each candidate expansion query in the candidate expansion query set;
matching words obtained by the word segmentation with the entity word in the knowledge graph to obtain the entity word of the target query and the entity word of the each candidate expansion query; and
matching the words obtained by the word segmentation with a qualifier of the entity word in the knowledge graph to obtain the qualifier of the target query and the qualifier of the each candidate expansion query.

### Claim 12 (depends on 8)

The apparatus according to claim 8, wherein the expanding the entity word and the qualifier of the target query, and the entity word and the qualifier of the each candidate expansion query in combination with the knowledge graph comprises:
expanding the entity word of the target query and the entity word of the each candidate expansion query based on a knowledge structure relation between different entity words in the knowledge graph, to obtain an expanded entity word of the target query and an expanded entity word of the each candidate expansion query; and
matching the qualifier of the target query and the qualifier of the each candidate expansion query with an associated qualifier set constructed based on the knowledge structure relation in the knowledge graph, to obtain an expanded qualifier of the target query and an expanded qualifier of the each candidate expansion query.

### Claim 13 (depends on 12)

The apparatus according to claim 12, wherein the knowledge graph comprises knowledge graphs of a plurality of domain categories;
the expanding the entity word and the qualifier of the target query, and the entity word and the qualifier of the each candidate expansion query in combination with the knowledge graph further comprises:
determining a target domain category of the knowledge graph matching the entity word of the target query and the entity word of the each candidate expansion query,
wherein the expanding the entity word of the target query and the entity word of the each candidate expansion query based on a knowledge structure relation between different entity words in the knowledge graph comprises:
expanding the entity word of the target query and the entity word of the each candidate expansion query based on a knowledge structure relation between different entity words in the knowledge graph of the target domain category; and
the matching the qualifier of the target query and the qualifier of the each candidate expansion query with an associated qualifier set constructed based on the knowledge structure relation in the knowledge graph comprises:
matching the qualifier of the target query and the qualifier of the each candidate expansion query with the associated qualifier set constructed based on the knowledge structure relation in the knowledge graph of the target domain category.

### Claim 14 (depends on 12)

The apparatus according to claim 12, wherein the matching using expansion results of the entity word and the qualifier of the target query and the entity word and the qualifier of the each candidate expansion query to determine an expanded query of the target query from the candidate expansion query set comprises:
using a candidate expansion query having an entity word identical to the expanded entity word of the target query and/or a qualifier identical to the expanded qualifier of the target query in the candidate expansion query set as the expanded query of the target query.

### Claim 15

A non-transitory computer storage medium storing a computer program, the computer program when executed by one or more processors, causes the one or more processors to perform operations, the operations comprising:
calculating an image click characteristic vector of a target query based on an acquired image click log associated with the target query;
finding, in a preset query set, a similar query of the target query based on the image click characteristic vector of the target query, to obtain a candidate expansion query set of the target query;
matching the target query and each candidate expansion query in the candidate expansion query set with an entity word and a qualifier in a knowledge graph, to extract an entity word and a qualifier of the target query and an entity word and a qualifier of each candidate expansion query;
expanding the entity word and the qualifier of the target query, and the entity word and the qualifier of each candidate expansion query in combination with the knowledge graph; and
matching using expansion results of the entity word and the qualifier of the target query and the entity word and the qualifier of each candidate expansion query, to determine an expanded query of the target query from the candidate expansion query set.
