# US 2026/0310689 A1 · Processing queries using explicitly defined relationships and inferred relationships of data objects

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

- Number: US 2026/0310689 A1
- Document ID: US20260310689A1
- Status: Published application
- Kind: A1
- Application number: 19/098,578
- Filed: 2025-04-02
- Published: 2026-10-08
- Earliest claimed: 2025-04-02
- Assignees: Dell Products L.P.
- Applicants: Dell Products L.P.
- Inventors: XIAOCHEN CHEN; ZIJIA WANG; SCOTT E. JOYCE; NING LI
- Web page: https://patentgrep.com/patent/US20260310689A1

## CPC

- G06F 16/288: {Entity relationship models}
- G06F 16/2423: {Interactive query statement specification based on a database schema}
- G06F: ELECTRIC DIGITAL DATA PROCESSING (computer systems based on specific computational models G06N)

## Abstract

Methods and systems for managing use of data sources of a distributed system are disclosed. The use of the data sources may be managed by processing queries using a refined dependency graph. The refined dependency graph may include nodes, based on data objects in the data sources, and/or edges, based on explicitly defined relationships and/or inferred relationships. The explicitly defined relationships may be based on database keys (e.g., primary keys, foreign keys, key constraints, and/or any other database key). The inferred relationships may be based on any measures of semantic similarity of the data objects. The any measures of the semantic similarity may be computed using a trained machine learning model.

## Claims

### Claim 1

A method for managing use of data sources of a distributed system, the method comprising:
obtaining a data operation request;
identifying a template based on the data operation request;
generating a refined dependency graph for the data sources by at least:
extracting database schemas from databases of the data sources, each of the data sources being a computing device, the database schemas comprise data objects stored in the databases, and each of the data objects being a data structure that stores data;
extracting all of the data objects from the data schemas; and
generating inferred relationships between the data objects, the inferred relationships supplement explicitly defined relationships between the data objects that are already defined in the database schemas;
populating the template using the refined dependency graph to obtain a processable query, the refined dependency graph comprising:
a plurality of nodes based on the data objects
a first plurality of edges between the nodes, the first plurality of the edges being based on the explicitly defined relationships between the data objects, the first plurality of the edges having different edge types based on a similarity level between the nodes connected by the first plurality of edges, and
a second plurality of the edges between the nodes, the second plurality of the edges being based on the inferred relationships between the data objects; and
processing the processable query using a query servicing entity to provide one or more computer implemented services to complete the data operation request, a first level of accuracy of the one or more computer implemented services provided to complete the data operation request using both the inferred relationships and the explicitly defined relationships being higher than a second level of accuracy associated with using only the explicitly defined relationships without the generating of the refined dependency graph to generate the inferred relationships.

### Claim 2 (depends on 1)

The method of claim 1, wherein the data operation request comprises a database request.

### Claim 3

(canceled)

### Claim 4 (depends on 1)

The method of claim 1, wherein the inferred relationships are based on relationships between the data objects, the relationships between the data objects being based on embeddings derived from the data objects and machine learning model analysis.

### Claim 5 (depends on 4)

The method of claim 4, wherein the relationships between the data objects are based on semantic similarities between the embeddings.

### Claim 6 (depends on 5)

The method of claim 5, wherein the inferred relationships are based, at least in part, on semantic similarities between the embeddings.

### Claim 7 (depends on 1)

The method of claim 1, wherein the template is not processable by a query servicing engine due to presence of unpopulated placeholders in the template.

### Claim 8 (depends on 1)

The method of claim 1, wherein populating the template using the refined dependency graph to obtain the processable query comprises:
filtering the data objects, based on the data operation request, the first plurality of the edges and the second plurality of the edges, to identify at least one of the data objects; and
adding at least one identifier for the at least one of the data objects to the template to generate the processable query.

### Claim 9

A non-transitory machine-readable medium having instructions stored therein, which when executed by a processor, cause the processor to perform operations for managing use of data sources of a distributed system, the operations comprising:
obtaining a data operation request;
identifying a template based on the data operation request;
generating a refined dependency graph for the data sources by at least:
extracting database schemas from databases of the data sources, each of the data sources being a computing device, the database schemas comprise data objects stored in the databases, and each of the data objects being a data structure that stores data;
extracting all of the data objects from the data schemas; and
generating inferred relationships between the data objects, the inferred relationships supplement explicitly defined relationships between the data objects that are already defined in the database schemas;
populating the template using the refined dependency graph to obtain a processable query, the refined dependency graph comprising:
a plurality of nodes based on the data objects,
a first plurality of edges between the nodes, the first plurality of the edges being based on the explicitly defined relationships between the data objects, the first plurality of the edges having different edge types based on a similarity level between the nodes connected by the first plurality of edges, and
a second plurality of the edges between the nodes, the second plurality of the edges being based on the inferred relationships between the data objects; and
processing the processable query using a query servicing entity to provide one or more computer implemented services to complete the data operation request, a first level of accuracy of the one or more computer implemented services provided to complete the data operation request using both the inferred relationships and the explicitly defined relationships being higher than a second level of accuracy associated with using only the explicitly defined relationships without the generating of the refined dependency graph to generate the inferred relationships.

### Claim 10 (depends on 9)

The non-transitory machine-readable medium of claim 9, wherein the data operation request comprises a database request.

### Claim 11

(canceled)

### Claim 12 (depends on 9)

The non-transitory machine-readable medium of claim 9, wherein the inferred relationships are based on relationships between the data objects, the relationships between the data objects being based on embeddings derived from the data objects and machine learning model analysis.

### Claim 13 (depends on 12)

The non-transitory machine-readable medium of claim 12, wherein the relationships between the data objects are based on semantic similarities between the embeddings.

### Claim 14 (depends on 13)

The non-transitory machine-readable medium of claim 13, wherein the inferred relationships are based, at least in part, on semantic similarities between the embeddings.

### Claim 15

A data processing system, comprising:
a processor; and
a memory coupled to the processor to store instructions, which when executed by the processor, cause the processor to perform operations managing use of data sources of a distributed system, the operations comprising:
obtaining a data operation request;
identifying a template based on the data operation request;
generating a refined dependency graph for the data sources by at least:
extracting database schemas from databases of the data sources, each of the data sources being a computing device, the database schemas comprise data objects stored in the databases, and each of the data objects being a data structure that stores data;
extracting all of the data objects from the data schemas; and
generating inferred relationships between the data objects, the inferred relationships supplement explicitly defined relationships between the data objects that are already defined in the database schemas;
populating the template using the refined dependency graph to obtain a processable query, the refined dependency graph comprising:
a plurality of nodes based on the data objects,
a first plurality of edges between the nodes, the first plurality of the edges being based on the explicitly defined relationships between the data objects, the first plurality of the edges having different edge types based on a similarity level between the nodes connected by the first plurality of edges, and
a second plurality of the edges between the nodes, the second plurality of the edges being based on the inferred relationships between the data objects; and
processing the processable query using a query servicing entity to provide one or more computer implemented services to complete the data operation request, a first level of accuracy of the one or more computer implemented services provided to complete the data operation request using both the inferred relationships and the explicitly defined relationships being higher than a second level of accuracy associated with using only the explicitly defined relationships without the generating of the refined dependency graph to generate the inferred relationships.

### Claim 16 (depends on 15)

The data processing system of claim 15, wherein the data operation request comprises a database request.

### Claim 17

(canceled)

### Claim 18 (depends on 15)

The data processing system of claim 15, wherein the inferred relationships are based on relationships between the data objects, the relationships between the data objects being based on embeddings derived from the data objects and machine learning model analysis.

### Claim 19 (depends on 18)

The data processing system of claim 18, wherein the relationships between the data objects are based on semantic similarities between the embeddings.

### Claim 20 (depends on 19)

The data processing system of claim 19, wherein the inferred relationships are based, at least in part, on semantic similarities between the embeddings.

### Claim 21 (depends on 1)

The method of claim 1, wherein the edge types comprise a hierarchy to indicate the similarity level between the nodes connected by the first plurality of edges.

### Claim 22 (depends on 21)

The method of claim 21, wherein generating the refined dependency graph for the data sources further comprises, while generating the inferred relationships between the data objects, deleting one of the first plurality of edges between two nodes of the nodes, wherein the similarity level between the two nodes being determined to be lower than a predetermined similarity threshold.

### Claim 23 (depends on 21)

The method of claim 21, wherein generating the refined dependency graph for the data sources further comprises, while generating the inferred relationships between the data objects, keeping, instead of deleting, one of the first plurality of edges between two nodes of the nodes, wherein the similarity level between the two nodes being determined to be lower than a predetermined similarity threshold, and the one of the first plurality of edges being kept to allow for a future change to the data objects that would cause the similarity level between the two nodes to exceed the predetermined similarity threshold.
