# US 2025/0094490 A1 · Systems and methods for detecting non-narrative regions of texts

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

- Number: US 2025/0094490 A1
- Document ID: US20250094490A1
- Status: Published application
- Kind: A1
- Application number: 18/916,472
- Filed: 2024-10-15
- Published: 2025-03-20
- Earliest claimed: 2021-03-22
- Assignees: Spotify AB
- Applicants: Spotify AB
- Inventors: Yongze YU; Sravana Kukati REDDY; Jussi Jerker KARLGREN; Rosemary Ellen JONES
- Web page: https://patentgrep.com/patent/US20250094490A1

## CPC

- G06F 16/685: {using automatically derived transcript of audio data, e.g. lyrics (speech recognition G10L15/00)}
- G06F 40/30: Semantic analysis
- G06F 16/635: Filtering based on additional data, e.g. user or group profiles
- G06F 16/65: Clustering; Classification
- G06F 40/166: Editing, e.g. inserting or deleting
- G06F 40/169: Annotation, e.g. comment data or footnotes
- G06F 40/194: Calculation of difference between files
- G06F 40/216: using statistical methods
- G06F 40/253: Grammatical analysis; Style critique
- G06F: ELECTRIC DIGITAL DATA PROCESSING (computer systems based on specific computational models G06N)

## Abstract

A method includes retrieving a text from a database. The text corresponds to audio from a media content item that is provided by a media providing service, and the text includes a plurality of segments. The method also includes assigning a score for each segment in the text by applying the text to a trained computational model. The score corresponds to a predicted relevance of the respective segment to a narrative of the media content item. The method further includes identifying a non-narrative segment within the text using the assigned scores.

## Claims

### Claim 1

(canceled)

### Claim 2

A method, comprising:
retrieving, from a database, a text that corresponds to audio from a media content item that is provided by a media providing service, the text comprising a plurality of segments;
determining, by applying the text to a trained language model, whether each segment in the text is topical or non-topical relative to a narrative of the media content item;
identifying a position within the text that changes from topical to non-topical, wherein two consecutive segments;
identifying a non-narrative portion within the text using the identified position; and
generating a clean text, including removing the non-narrative portion from the text based at least in part on the identified position within the text.

### Claim 3 (depends on 2)

The method of claim 2, wherein determining whether each segment in the text is topical or non-topical relative to the narrative of the media content item comprises classifying each segment in the text as topical or non-topical relative to the narrative of the media content item.

### Claim 4 (depends on 2)

The method of claim 2, wherein the text includes a transcript of audio content of the media content item, and the method further comprises:
generating a description of the media content item based on the clean text.

### Claim 5 (depends on 2)

The method of claim 2, further comprising:
providing the media content item associated with the text to a user of the media providing service based at least in part on the generated clean text.

### Claim 6 (depends on 2)

The method of claim 2, wherein:
the trained language model is trained on a plurality of annotated texts; and
each of the annotated texts in the plurality of annotated texts includes text corresponding to audio from a media content item of a plurality of media content items and a plurality of annotations.

### Claim 7 (depends on 6)

The method of claim 6, wherein:
each annotated text of the plurality of annotated texts includes an annotation for each segment in the text corresponding to audio from a media content item of the plurality of media content items.

### Claim 8 (depends on 6)

The method of claim 6, further comprising:
for a respective annotated text, generating a label for each segment in the respective annotated text based on at least a portion of the plurality of annotations.

### Claim 9 (depends on 2)

The method of claim 2, wherein:
determining a segment in the text is topical or non-topical relative to a narrative of the media content item includes analyzing content of the segment.

### Claim 10 (depends on 2)

The method of claim 2, wherein:
determining a segment in the text is topical or non-topical relative to a narrative of the media content item includes analyzing content of the segment and content of a segment preceding the segment.

### Claim 11

An electronic device associated with a media providing service, comprising:
one or more processors; and
memory storing one or more programs, the one or more programs including instructions, which when executed by the one or more processors, cause the electronic device to perform a set of operations, comprising:
retrieving, from a database, a text that corresponds to audio from a media content item that is provided by a media providing service, the text comprising a plurality of segments;
determining, by applying the text to a trained language model, whether each segment in the text is topical or non-topical relative to a narrative of the media content item;
identifying a position within the text that changes from topical to non-topical, wherein two consecutive segments;
identifying a non-narrative portion within the text using the identified position; and
using the identification of the non-narrative portions, providing the media content item to a user as a recommendation.

### Claim 12 (depends on 11)

The electronic device of claim 11, wherein determining whether each segment in the text is topical or non-topical relative to the narrative of the media content item comprises classifying each segment in the text as topical or non-topical relative to the narrative of the media content item.

### Claim 13 (depends on 11)

The electronic device of claim 11, wherein the text includes a transcript of audio content of the media content item, and the set of operations further comprises:
generating a description of the media content item based on the identification of the non-narrative portions.

### Claim 14 (depends on 11)

The electronic device of claim 11, wherein:
the trained language model is trained on a plurality of annotated texts; and
each of the annotated texts in the plurality of annotated texts includes text corresponding to audio from a media content item of a plurality of media content items and a plurality of annotations.

### Claim 15 (depends on 14)

The electronic device of claim 14, wherein:
each annotated text of the plurality of annotated texts includes an annotation for each segment in the text corresponding to audio from a media content item of the plurality of media content items.

### Claim 16 (depends on 14)

The electronic device of claim 14, wherein the set of operations further comprises:
for a respective annotated text, generating a label for each segment in the respective annotated text based on at least a portion of the plurality of annotations.

### Claim 17 (depends on 11)

The electronic device of claim 11, wherein:
determining a segment in the text is topical or non-topical relative to a narrative of the media content item includes analyzing content of the segment.

### Claim 18 (depends on 11)

The electronic device of claim 11, wherein:
determining a segment in the text is topical or non-topical relative to a narrative of the media content item includes analyzing content of the segment and content of a segment preceding the segment.

### Claim 19

A non-transitory computer-readable storage medium storing one or more programs configured for execution by an electronic device associated with a media providing service, the electronic device having one or more processors, the one or more programs including instructions, which when executed by the one or more processors, cause the electronic device to perform a set of operations, comprising:
retrieving, from a database, a text that corresponds to audio from a media content item that is provided by a media providing service, the text comprising a plurality of segments;
determining, by applying the text to a trained language model, whether each segment in the text is topical or non-topical relative to a narrative of the media content item;
identifying a position within the text that changes from topical to non-topical, wherein two consecutive segments;
identifying a non-narrative portion within the text using the identified position; and
using the identification of the non-narrative portions, providing a description of the media content item to a user.

### Claim 20 (depends on 19)

The non-transitory computer-readable storage medium of claim 19, wherein determining whether each segment in the text is topical or non-topical relative to the narrative of the media content item comprises classifying each segment in the text as topical or non-topical relative to the narrative of the media content item.

### Claim 21 (depends on 20)

The non-transitory computer-readable storage medium of claim 20, wherein:
determining a segment in the text is topical or non-topical relative to a narrative of the media content item includes analyzing content of the segment.
