# US 2026/0017271 A1 · Predicting relevance of resources to search queries

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

- Number: US 2026/0017271 A1
- Document ID: US20260017271A1
- Status: Published application
- Kind: A1
- Application number: 19/278,742
- Filed: 2025-07-23
- Published: 2026-01-15
- Earliest claimed: 2024-07-15
- Assignees: Capital One Services, LLC
- Applicants: Capital One Services, LLC
- Inventors: Lavanya Gowlla; Lavanya Raja; Puneet Goyal; Ajay Syamaladevi; Vineet Kumar Jha; Roshni Singh; Uday Singh; Shivansh Sethi
- Web page: https://patentgrep.com/patent/US20260017271A1

## CPC

- G06F 16/24578: {using ranking}
- G06N 3/08: Learning methods
- G06F: ELECTRIC DIGITAL DATA PROCESSING (computer systems based on specific computational models G06N)
- G06N: COMPUTING ARRANGEMENTS BASED ON SPECIFIC COMPUTATIONAL MODELS

## Abstract

Systems and methods for predicting relevance of resources to search queries. In some aspects, the system may receive a search query requesting resources from a database and may identify resources, including messages, relating to the search query. The system may extract, from the messages, hyperlinks specifying locations within the database. The system may input, into a model, the search query and the hyperlinks to cause the model to generate predictions of relevance of the hyperlinks to the search query. The system may then determine an overall relevance score for the resources in relation to the search query based on the predictions of relevance.

## Claims

### Claim 1

A system for predicting relevance of resources to search queries, the system comprising:
one or more processors; and
one or more non-transitory, computer-readable media having computer-executable instructions stored thereon that, when executed by the one or more processors, cause the system to perform operations comprising:
in response to obtaining a search query requesting resources, identifying a plurality of resources relating to the search query, wherein the plurality of resources comprises one or more messages;
extracting, from the one or more messages, one or more hyperlinks specifying one or more locations;
generating, via a machine learning model that is trained to predict relevance based on training hyperlinks used as learning model inputs, one or more relevance predictions for the one or more hyperlinks to the search query based on the search query and the one or more hyperlinks, wherein a first hyperlink specifying a first type of location has a lower relevance than a second hyperlink specifying a second type of location; and
determining a relevance score for the plurality of resources in relation to the search query based on the one or more relevance predictions.

### Claim 2

A method comprising:
obtaining, from a user, a search query requesting resources;
in response to obtaining the search query requesting resources, identifying a plurality of resources relating to the search query;
extracting, from the plurality of resources, one or more hyperlinks specifying one or more locations;
generating, via a machine learning model that is trained to predict hyperlink relevance, one or more relevance predictions for the one or more hyperlinks to the search query based on the search query, and a set of types of locations of the one or more hyperlinks; and
determining a relevance score for the plurality of resources in relation to the search query based on the one or more relevance predictions.

### Claim 3 (depends on 2)

The method of claim 2, wherein determining the relevance score further comprises:
identifying a connected message to a first message of the plurality of resources, wherein the first message comprises a first hyperlink;
determining that the connected message comprises a target keyword; and
increasing a first relevance score corresponding with the first hyperlink based on the determining that the connected message comprises the target keyword.

### Claim 4 (depends on 2)

The method of claim 2, wherein determining the relevance score further comprises:
identifying a connected message to a first message of the plurality of resources, wherein the first message comprises a first hyperlink;
determining that the connected message indicates at least one of gratitude, resolution, or conclusion; and
increasing a first relevance score corresponding with the first hyperlink based on the determining that the connected message indicates at least one of gratitude, resolution, or conclusion.

### Claim 5 (depends on 2)

The method of claim 2, further comprising detecting one or more emojis in the plurality of resources, wherein determining the relevance score further comprises determining the relevance score based on a sentiment associated with the one or more emojis.

### Claim 6 (depends on 2)

The method of claim 2, wherein:
a first hyperlink specifying a first type of location has a lower relevance than a second hyperlink specifying a second type of location; and
the first type of location comprises a video conference platform and the second type of location comprises a document viewing platform.

### Claim 7 (depends on 2)

The method of claim 2, wherein generating the one or more relevance predictions comprises generating the one or more relevance predictions based on a respective type of location specified by each respective hyperlink of the one or more hyperlinks, wherein a first hyperlink specifying a first type of location has a lower relevance than a second hyperlink specifying a second type of location.

### Claim 8 (depends on 2)

The method of claim 2, wherein determining the relevance score further comprises:
determining a plurality of subjectivity scores for the plurality of resources, wherein a first type of resource has a lower subjectivity than a second type of resource; and
determining the relevance score based on the one or more relevance predictions and the plurality of subjectivity scores.

### Claim 9 (depends on 2)

The method of claim 2, wherein determining the relevance score further comprises:
classifying the search query into a first category of a plurality of categories;
classifying the plurality of resources into one or more categories of the plurality of categories;
determining a plurality of relatedness scores for the plurality of resources, wherein a first resource belonging to the first category has a higher relatedness score than a second resource belonging to a different category of the plurality of categories; and
determining the relevance score based on the one or more relevance predictions and the plurality of relatedness scores.

### Claim 10 (depends on 2)

The method of claim 2, further comprising:
determining that the relevance score satisfies a relevance threshold; and
based on determining that the relevance score satisfies the relevance threshold, outputting the plurality of resources to the user.

### Claim 11 (depends on 2)

The method of claim 2, wherein determining the relevance score further comprises:
identifying, within the plurality of resources, one or more keywords; and
determining the relevance score based on the one or more relevance predictions and the one or more keywords indicating resolution.

### Claim 12

One or more non-transitory, computer-readable media storing instructions that, when executed by one or more processors, cause operations comprising:
in response to a search query requesting resources, identifying a plurality of resources relating to the search query, wherein the plurality of resources comprises one or more messages;
extracting, from the one or more messages, one or more hyperlinks;
generating, via a machine learning model, one or more relevance predictions for the one or more hyperlinks to the search query based on the search query and the one or more hyperlinks, wherein the machine learning model is trained to predict hyperlink relevance; and
determining a relevance score for the plurality of resources in relation to the search query based on the one or more relevance predictions.

### Claim 13 (depends on 12)

The one or more non-transitory, computer-readable media of claim 12, the operations further comprising:
determining that the relevance score does not satisfy a relevance threshold; and
based on determining that the relevance score does not satisfy the relevance threshold, presenting a subset of the plurality of resources to a user, wherein the subset of the plurality of resources comprises a first subset of hyperlinks, wherein each respective hyperlink of the first subset of hyperlinks is associated with a respective relevance score that satisfies the relevance threshold.

### Claim 14 (depends on 12)

The one or more non-transitory, computer-readable media of claim 12, wherein determining the relevance score further comprises:
identifying a connected message to a first message of the plurality of resources, wherein the first message comprises a first hyperlink;
determining that the connected message comprises a target keyword; and
increasing a first relevance score corresponding with the first hyperlink based on the determining that the connected message comprises the target keyword.

### Claim 15 (depends on 12)

The one or more non-transitory, computer-readable media of claim 12, wherein determining the relevance score further comprises:
identifying a connected message to a first message of the plurality of resources, wherein the first message comprises a first hyperlink;
determining that the connected message indicates at least one of gratitude, resolution, or conclusion; and
increasing a first relevance score corresponding with the first hyperlink based on the determining that the connected message indicates at least one of gratitude, resolution, or conclusion.

### Claim 16 (depends on 12)

The one or more non-transitory, computer-readable media of claim 12, wherein generating the one or more relevance predictions comprises causing the machine learning model to generate the one or more relevance predictions based on a type of location specified by at least one hyperlink.

### Claim 17 (depends on 12)

The one or more non-transitory, computer-readable media of claim 12, wherein determining the relevance score further comprises:
determining a plurality of subjectivity scores for the plurality of resources, wherein a first type of resource has a lower subjectivity than a second type of resource; and
determining the relevance score based on the one or more relevance predictions and the plurality of subjectivity scores.

### Claim 18 (depends on 12)

The one or more non-transitory, computer-readable media of claim 12, wherein determining the relevance score further comprises:
classifying the search query into a first category of a plurality of categories;
classifying the plurality of resources into one or more categories of the plurality of categories;
determining a plurality of relatedness scores for the plurality of resources, wherein a first resource belonging to the first category has a higher relatedness score than a second resource belonging to a different category of the plurality of categories; and
determining the relevance score based on the one or more relevance predictions and the plurality of relatedness scores.

### Claim 19 (depends on 12)

The one or more non-transitory, computer-readable media of claim 12, the operations further comprising detecting one or more emojis in the plurality of resources, wherein determining the relevance score further comprises determining the relevance score based on a sentiment associated with the one or more emojis.

### Claim 20 (depends on 12)

The one or more non-transitory, computer-readable media of claim 12, the operations further comprising:
determining that the relevance score satisfies a relevance threshold; and
based on determining that the relevance score satisfies the relevance threshold, outputting the plurality of resources to a user.
