Wake Word Selection Assistance Engine for False Detection Reduction

Resolve Bottlenecks,
Find Innovative Solutions
Generate Solutions

Solution Overview

Problem

Selecting a suitable wake word for personal assistants is challenging due to the risk of false detections, as not all words or phrases are robust enough to avoid confusion with other sounds or words, leading to unnecessary activations.

Innovation Solution

The system provides assistance to users in selecting a robust wake word by using a wake word assistance engine that includes a pronunciation application to predict phonetic spelling, a uniqueness application to assess phonetic richness, and an accept application to determine false acceptance rates, allowing users to choose a wake word that minimizes false detections.

Engineering Contradictions & Design Principles

VSEngineering Contradiction Analysis

1Ease of operation

If a simple wake word selection process is used, then ease of operation is improved, but reliability deteriorates due to false detections

Engineering Contradiction:
Improvewake word selection processVSAvoidfalse detection rate
Core Design Contradiction:
Ease of operationVSReliability

Solution Approach 1:

The system performs preliminary phonetic analysis, confusion checks, and false detection rate calculations before the user finalizes their wake word selection. The wake word assistance engine evaluates candidate wake words against a confusion dictionary and computes false detection rates in advance, allowing users to make informed selections without understanding complex phonetic concepts.

Inventive Principle:
Principle #10Preliminary action

Solution Approach 2:

The wake word assistance engine acts as an intermediary between the user and the wake word recognition system. It translates user-selected wake words into phonetic representations, checks them against a confusion dictionary, and provides feedback on false detection rates, thereby mediating between simple user input and complex recognition reliability.

Inventive Principle:
Principle #24Intermediary (Mediator)

2Reliability

If expert knowledge is required for wake word selection, then reliability improves through better phonetic understanding, but ease of operation deteriorates

Engineering Contradiction:
Improvewake word robustnessVSAvoidselection complexity
Core Design Contradiction:
ReliabilityVSEase of operation

Solution Approach 1:

The system performs self-service by automatically evaluating wake word candidates against phonetic criteria and false detection metrics. The wake word assistance engine independently computes confusion scores and false detection rates, eliminating the need for users to possess expert phonetic knowledge while ensuring reliable wake word selection.

Inventive Principle:
Principle #25Self-service

Solution Approach 2:

The system provides feedback to users about the reliability of their selected wake words by displaying false detection rates and confusion scores. This feedback mechanism guides users toward selecting more robust wake words without requiring them to understand the underlying phonetic analysis.

Inventive Principle:
Principle #23Feedback

3Measurement precision

If comprehensive phonetic analysis is performed, then measurement precision improves for wake word robustness, but device complexity increases

Engineering Contradiction:
Improvephonetic evaluation accuracyVSAvoidwake word assistance engine
Core Design Contradiction:
Measurement precisionVSDevice complexity

Solution Approach 1:

The wake word assistance engine is segmented into distinct functional modules: a phonetic spelling application that converts text to phonetic representations, a confusion dictionary that stores pre-computed confusion data, and a false detection rate calculator. This segmentation allows comprehensive phonetic analysis while managing system complexity through modular design.

Inventive Principle:
Principle #1Segmentation

Data Source

PatentEP3966809B1Wake word selection assistance architectures and methods
Publication Date: 2024.05.15 MICROSOFT TECHNOLOGY LICENSING LLC
  • EP3966809B1 patent drawingFigure 1
  • EP3966809B1 patent drawingFigure 2~3
  • EP3966809B1 patent drawingFigure 4

AI summary

Generally discussed herein are devices, systems, and methods for custom wake word selection assistance. A method can include receiving, at a device, data indicating a custom wake word provided by a user, determining one or more characteristics of the custom wake word, determining that use of the custom wake word will cause more than a threshold rate of false detections based on the characteristics, rejecting the custom wake word as the wake word for accessing a personal assistant in response to determining that use of the custom wake word will cause more than a threshold rate of false detections, and setting the custom wake word as the wake word in response to determining that use of the custom wake word will not cause more than the threshold rate of false detections.