Phoneme Database Sorting for Voice Interfaces

Resolve Bottlenecks,
Find Innovative Solutions
Generate Solutions

Solution Overview

Problem

Voice-controlled user interfaces face significant latency due to the computationally intensive process of re-phonetizing larger data lists, which hinders quick access to specific list entries in applications like car phones and contact lists, where sorting names by first and last names is common, leading to recognition ambiguities and user frustration.

Innovation Solution

Storing phonemes with separators between data fields in a phonetics database allows for rapid reassembly of phonetized data lists without re-phonetization, enabling fast response times by treating separators as special phoneme symbols and swapping phoneme sequences, thus avoiding the need for resource-intensive re-phonetization during sorting.

Engineering Contradictions & Design Principles

VSEngineering Contradiction Analysis

1Measurement precision

If re-phonetization is performed during sorting operations, then speech recognition accuracy is maintained, but system latency increases significantly

Engineering Contradiction:
Improvespeech recognition accuracyVSAvoidsystem latency
Core Design Contradiction:
Measurement precisionVSLoss of time

Solution Approach 1:

The patent applies preliminary action by pre-dividing list entries into data fields with embedded separators before phonetization. This allows the phonetized version to be generated once and stored, enabling rapid reassembly during sorting without re-performing the computationally intensive phonetization process. The separator symbols are inserted in advance at field boundaries, so that during sorting operations, only the phoneme sequences need to be rearranged and recombined, dramatically reducing latency while preserving recognition accuracy.

Inventive Principle:
Principle #10Preliminary action

2Reliability

If complete list entries are phonetized without field separators, then speech recognition works, but sorting and reassembly become computationally intensive

Engineering Contradiction:
Improvespeech recognition functionalityVSAvoidsorting speed
Core Design Contradiction:
ReliabilityVSProductivity

Solution Approach 1:

The patent applies segmentation by dividing each list entry into distinct data fields separated by special separator symbols before phonetization. This creates a structured phonetized representation where separator phonemes mark field boundaries. During sorting operations, the system can efficiently reassemble phonetized entries by rearranging field segments between separators rather than reprocessing entire entries, significantly improving sorting speed while maintaining the integrity needed for speech recognition.

Inventive Principle:
Principle #1Segmentation

3Manufacturing precision

If phonetized data lists are regenerated during sorting, then data accuracy is maintained, but resource consumption increases

Engineering Contradiction:
Improvedata accuracyVSAvoidcomputational resource consumption
Core Design Contradiction:
Manufacturing precisionVSUse of energy by moving object

Solution Approach 1:

The patent applies preliminary action by pre-inserting separator symbols and performing phonetization once during data preparation. The phonetized data with embedded separators is stored for reuse during sorting operations. This eliminates the need to regenerate phonetized data during sorting, significantly reducing computational resource consumption while maintaining data accuracy through the preserved separator structure that enables precise field reassembly.

Inventive Principle:
Principle #10Preliminary action

Data Source

PatentEP2815396B1Method for phoneticizing a data list and speech-controlled user interface
Publication Date: 2016.12.07 CONTINENTAL AUTOMOTIVE GMBH
  • EP2815396B1 patent drawingFigure 1
  • EP2815396B1 patent drawingFigure 2
  • EP2815396B1 patent drawingFigure 3

AI summary

The invention relates to a method for phonetising a data list (2) consisting of list entries containing text, each list entry in the data list (2) being subdivided into at least two data fields and a speech-controlled user interface (1). According to said method, a list entry from the text representation is phonetically converted and is saved as a phoneme in a phonetisied data list (6). According to the invention, a separator is introduced into the text of a list entry between the respective data fields of the list entry, is phonetically converted and is saved as a phoneme symbol, and the phonemes are saved in the phonetic data base (7), the phonetic data list (6) being produced from the phonemes saved in the phonetic data base (7). The invention also relates to a correspondingly designed speech-controlled user interface.