Skip to main content

Conversational and Voice User Interfaces

Conversational UI/UX​

  • CUI = Conversational User Interface
  • Can refer to text-based chatbots
  • UX = User Experience
    • Broader term encompassing experience of the interaction, not just design of interfaces and information

VUI Example​

  • ASR = Automated Speech Recognition
  • wakeword detection + acoustic model-based transcription + confidence level
  • NLU/NLP
    • Keyword recognition + intent matching + variable/slot
  • Dialogue management + TTS vocalisation
    • Response generation

Challenges​

  • Voice is a profoundly personal trait
  • Reflects where you're from, mother tongue, dialect and accent etc
  • Issues around privacy and surveillance
    • Pertinent given the often intimate/home environment in which VUIs are placed

Socio-economic challenges​

  • Dialects, accents and pitch
  • VUIs struggle more with certain dialects, worst for the Scottish
  • Accuracy also lower for higher pitched voices
  • Could contain racial bias