Speaker-Dependent Warm Words for Secure Voice Control

Resolve Bottlenecks,
Find Innovative Solutions
Generate Solutions

Solution Overview

Problem

In speech-enabled environments, the existing systems face challenges in accurately and securely controlling long-standing operations using hotwords, as they often require repeated hotword invocation and are susceptible to false positives and unauthorized access by other users.

Innovation Solution

The method involves activating a set of speaker-dependent warm words associated with a user who initiated a long-standing operation, allowing only that user to control the operation by speaking specific warm words, with speaker verification ensuring authenticity and reducing false positives.

Engineering Contradictions & Design Principles

VSEngineering Contradiction Analysis

1Ease of operation

If hotwords are used to control long-standing operations, then operations can be initiated easily, but repeated hotword invocation is required and false positives occur

Engineering Contradiction:
Improveoperation initiationVSAvoidfalse positive rate
Core Design Contradiction:
Ease of operationVSReliability

Solution Approach 1:

The system segments the control mechanism into two distinct parts: hotwords for initiating operations and warmwords for controlling ongoing operations. This segmentation allows hotwords to remain simple and easily recognizable for starting operations, while warmwords provide more specific control commands that reduce false positives during operation continuation and modification.

Inventive Principle:
Principle #1Segmentation

Solution Approach 2:

The system performs preliminary action by activating a set of warmwords specifically associated with a long-standing operation when that operation is initiated. These warmwords are pre-configured and activated in advance, allowing the system to respond to specific control commands without requiring repeated hotword invocation, thereby improving both ease of operation and reliability.

Inventive Principle:
Principle #10Preliminary action

2Ease of operation

If hotwords are used for controlling operations, then any user can trigger actions, but unauthorized access by other users occurs

Engineering Contradiction:
Improvecontrol accessibilityVSAvoidauthorization security
Core Design Contradiction:
Ease of operationVSReliability

Solution Approach 1:

The system applies local quality by associating specific warmwords with specific users based on their voice characteristics. When a user initiates a long-standing operation, the system activates warmwords that are locally tailored to that user's voice profile. This ensures that only the authorized user can effectively control their operations, maintaining security while preserving ease of operation for the legitimate user.

Inventive Principle:
Principle #3Local quality

Solution Approach 2:

The system implements feedback through speaker verification that confirms whether the person speaking a warmword is the authorized user. The system continuously monitors and verifies speaker identity, providing feedback control that allows operation modification only when the authorized user is detected, thereby preventing unauthorized access while maintaining user-friendly control.

Inventive Principle:
Principle #23Feedback

3Ease of operation

If warm words are activated for all users, then any user can control operations, but false positives increase and security decreases

Engineering Contradiction:
Improvecontrol convenienceVSAvoidfalse positive rate
Core Design Contradiction:
Ease of operationVSReliability

Solution Approach 1:

The system makes warmword recognition speaker-dependent by tailoring the recognition parameters to each specific user's voice characteristics. This local quality approach ensures that warmwords are recognized accurately for the intended user while rejecting similar-sounding commands from other users, thereby reducing false positives and maintaining security without sacrificing control convenience for the authorized user.

Inventive Principle:
Principle #3Local quality

4Device complexity

If repeated hotword invocation is required, then operation control is simple, but user experience deteriorates and time is lost

Engineering Contradiction:
Improvecontrol mechanism simplicityVSAvoidoperation control time
Core Design Contradiction:
Device complexityVSLoss of time

Solution Approach 1:

The system performs preliminary action by pre-activating a set of warmwords associated with a long-standing operation when it is initiated. These warmwords remain active and ready for use, allowing the user to control the operation with simple follow-up commands rather than repeating the hotword. This eliminates time loss while maintaining control simplicity, as the warmwords are specifically tailored to the ongoing operation.

Inventive Principle:
Principle #10Preliminary action

Solution Approach 2:

The system maintains continuity of useful action by keeping the activated warmwords available throughout the duration of the long-standing operation. This allows the user to continuously control the operation using the same set of warmwords without interruption or need to re-invite the hotword, thereby improving user experience and reducing time loss while maintaining operational simplicity.

Inventive Principle:
Principle #20Continuity of useful action

Data Source

PatentEP4248440B1Speaker dependent follow up actions and warm words
Publication Date: 2024.09.04 GOOGLE LLC
  • EP4248440B1 patent drawingFigure 1A
  • EP4248440B1 patent drawingFigure 1B
  • EP4248440B1 patent drawingFigure 1C

AI summary

A method (500) includes receiving audio data (402) corresponding to an utterance (106) spoken by a user (102) that includes a command (118) for a digital assistant (105) to perform a long-standing operation, activating a set of one or more warm words (112) associated with a respective action for controlling the long-standing operation, and associating the activated set of one or more warm words with only the user. While the digital assistant is performing the long-standing operation, the method includes receiving additional audio data corresponding to an additional utterance, identifying one of the warm words from the activated set of warm words, and performing speaker verification (400) on the additional audio data. The method further includes performing the respective action associated with the identified one of the warm words for controlling the long-standing operation when the additional utterance was spoken by the same user that is associated with the activated set of one or more warm words.