Activation trigger processing

• US 10,629,204 B2
• Filed: 10/03/2018
• Issued: 04/21/2020
• Est. Priority Date: 04/23/2018
• Status: Active Grant

• Current Assignee: Spotify AB (Spotify Technology SA)
• Original Assignee: Spotify AB (Spotify Technology SA)
• Inventors: Mitic, Richard
• Primary Examiner(s): Colucci, Michael
• Application Number:
  US16/150,459
• Publication Number:
  US 20190325869A1
• Time in Patent Office: 566 Days
• Field of Search: 704235, 704 9, 704268, 704260, 704254, 704251, 704236, 704500, 705 1417, 705 39, 715221
• CPC Class Codes: G10L 15/08 Speech classification or search
• CPC Class Codes: G10L 15/22 Procedures used during a speech recognition process, e.g. man-machine dialogue
• CPC Class Codes: G10L 15/30 Distributed recognition, e.g. in client-server systems, for mobile phones or network applications
• CPC Class Codes: G10L 2015/088 Word spotting
• CPC Class Codes: G10L 2015/223 Execution procedure of a spoken command

Abstract

Utterance-based user interfaces can include activation trigger processing techniques for detecting activation triggers and causing execution of certain commands associated with particular command pattern activation triggers without waiting for output from a separate speech processing engine. The activation trigger processing techniques can also detect speech analysis patterns and selectively activate a speech processing engine.

Claims

1. A system, comprising:

one or more processors configured to operate as;

a speech analysis engine; and

an activation trigger engine configured to;

detect an activation trigger within an audio input;

responsive to the activation trigger matching a first an analysis-type activation trigger pattern, activating the speech analysis engine; and

responsive to the activation trigger matching a command-type activation trigger pattern, cause execution of a command associated with the command-type activation trigger pattern without activating the speech analysis engine,

wherein the analysis-type activation trigger pattern includes an analysis-type activation trigger set of two or more syllables; and

wherein the command-type activation trigger pattern includes a command-type activation trigger set formed from the union of a proper subset of the command-type activation trigger set and a command component set of one or more syllables.

2. The system of claim 1, wherein the speech analysis engine is configured to perform speech-to-text conversion and the activation trigger engine is configured to detect the activation trigger within the audio input without using speech-to-text conversion.

3. The system of claim 1, wherein the one or more processors include a first set of one or more processors configured to operate as the speech analysis engine and a second set of one or more processors configured to operate as the activation trigger engine;

and

wherein the system further comprises;

a first device having the first set of one or more processors configured to operate as the speech analysis engine; and

a second device remote from the first device and having the second set of one or more processors configured to operate as the activation trigger engine.

4. The system of claim 1, wherein the speech analysis engine is configured to perform speech recognition on audio input.

5. The system of claim 1,

wherein the speech analysis engine is configured to perform natural language processing on audio input;

and

wherein the activation trigger engine is not configured to perform natural language processing on audio input.

6. The system of claim 1,

wherein the speech analysis engine is configured to produce an intent based on natural language processing and cause performance of an associated action;

and

wherein the activation trigger engine is not configured to produce an intent based on natural language processing.

7. The system of claim 1, wherein selectively activating the speech analysis engine includes piping output of a recording device or process to the speech analysis engine.

8. The system of claim 1, wherein the activation trigger engine is configured to determine whether the activation trigger matches the analysis-type activation trigger pattern of the command-type activation trigger pattern based on features extracted from the audio input.

9. The system of claim 8, wherein the features include cepstral coefficient features.

10. A non-transitory computer readable medium storing instructions that, when executed by one or more processors, cause the one or more processors to:

operate a speech analysis engine configured to perform speech-to-text conversion;

operate an activation trigger engine configured to, without using speech-to-text conversion;

detect an activation trigger within an audio input;

responsive to the activation trigger matching an analysis-type activation trigger pattern, activate the speech analysis engine; and

responsive to the activation trigger matching a command-type activation trigger pattern, cause execution of a command associated with the command-type activation trigger pattern without activating the speech analysis engine,

wherein the analysis-type activation trigger pattern includes an analysis-type activation trigger set of two or more syllables; and

wherein the command-type activation trigger pattern includes a command-type activation trigger set formed from the union of a proper subset of the command-type activation trigger set and a command component set of one or more syllables.

11. A system, comprising:

a first device having a first set of one or more processors configured to operate as a speech analysis engine;

a second device having a second set of one or more processors configured to operate as an activation trigger engine, wherein the activation trigger engine is configured to, without using speech-to-text conversion;

detect an activation trigger within an audio input;

responsive to the activation trigger matching an analysis-type activation trigger pattern, remotely activate the speech analysis engine of the first device; and

responsive to the activation trigger matching a command-type activation trigger pattern, cause execution of a command associated with the command-type activation trigger pattern without activating the speech analysis engine,

wherein the analysis-type activation trigger pattern includes an analysis-type activation trigger set of two or more words; and

wherein the command-type activation trigger pattern includes a command-type activation trigger set formed from the union of a proper subset of the command-type activation trigger set and a command component set of one or more words.

12. The system of claim 11, wherein the speech analysis engine is configured to perform speech-to-text conversion.