Scholarlink

StutteringStuttering is speech that is characterized by repetition or prolongation of sounds, syllables, words, and hesitation or pauses that disrupt the rhythmic flow of speech. People who stutter want to use artificially intelligent automatic speech recognitionAutomatic speech recognition (AIArtificial Intelligence (AI)-ASR) systems but are frequently misunderstood and cut off because AIArtificial Intelligence (AI)-ASR models are optimized on data from people who do not stutter. A primary reason for the deficiency in current AIArtificial Intelligence (AI)-ASR models is the lack of large, diverse, and specified data on stuttered speech. To remedy this problem, this research proposes an AIArtificial Intelligence (AI) for Systems EngineeringSystems engineering (AI4SEArtificial Intelligence for Systems Engineering (AI4SE)) approach to data specification and modelingModeling of stuttered speech. While traditional SE lifecycle and principles have been successful in building heretofore complex systemsComplex systems, current AIArtificial Intelligence (AI)-enabled systems have introduced new paradigms that do not fit SE traditions. Despite the difficulty, this research advocates a refined AI4SEArtificial Intelligence for Systems Engineering (AI4SE) approach to establishing design integrity, artifacts, and configuration baselinesBaseline for such systems.

错误：搜索内容不能为空，请输入英文关键词
错误：关键词超出字数限制，请精简
高级检索

Integration of AI in Data Requirements for Stuttering-Aware Speech Recognition Systems

  • 会议论文

  • 发表日期：

    2026年

查看全文
  • Ibibia Altraide,
  • Steve Simske

摘要

StutteringStuttering is speech that is characterized by repetition or prolongation of sounds, syllables, words, and hesitation or pauses that disrupt the rhythmic flow of speech. People who stutter want to use artificially intelligent automatic speech recognitionAutomatic speech recognition (AIArtificial Intelligence (AI)-ASR) systems but are frequently misunderstood and cut off because AIArtificial Intelligence (AI)-ASR models are optimized on data from people who do not stutter. A primary reason for the deficiency in current AIArtificial Intelligence (AI)-ASR models is the lack of large, diverse, and specified data on stuttered speech. To remedy this problem, this research proposes an AIArtificial Intelligence (AI) for Systems EngineeringSystems engineering (AI4SEArtificial Intelligence for Systems Engineering (AI4SE)) approach to data specification and modelingModeling of stuttered speech. While traditional SE lifecycle and principles have been successful in building heretofore complex systemsComplex systems, current AIArtificial Intelligence (AI)-enabled systems have introduced new paradigms that do not fit SE traditions. Despite the difficulty, this research advocates a refined AI4SEArtificial Intelligence for Systems Engineering (AI4SE) approach to establishing design integrity, artifacts, and configuration baselinesBaseline for such systems.