Kedis ku xidhan hagaha · Dhexdhexaad Heerka

AI oo ku jira Kediska Falanqaynta Hadalka Xarunta Wicitaanka

Tijaabi aqoonsiga hadalka, shubanka, iyo falanqaynta dareenka ee cilmi-baarista xarunta wicitaanka, oo ay ku jiraan xadka iyo baaxadda daraasad kasta oo la soo xigtay.

Wadooyinka haga ee la xidhiidhaFalanqaynta Hadalka Xarunta Wicitaanka AI
Su'aal 1 ee 8

What output can automatic speech recognition provide for speech analytics?

Horudhac su'aal kasta8 su'aalo

Jawaabaha iyo sharraxaadaha waxay soo baxayaan markaad qaadanayso kediska.

  1. What output can automatic speech recognition provide for speech analytics?
    • A machine-generated transcript for downstream text analysis
    • A guaranteed speaker identity
    • A final diagnosis of the caller’s emotional state
    • A payment authorization
  2. What does speaker diarization do in a call recording?
    • Wuxuu u tarjumaa wicitaanka luqad kale
    • Segments the recording into speaker turns and assigns speaker labels
    • Wuxuu cabbiraa sida uu dhawaaqu u dheeraa
    • Waxay siraysaa duubista
  3. In IBM’s two-stream call-center emotion-recognition study, which information streams are combined?
    • Video and keystroke timing
    • Customer purchase history and account balance
    • Acoustic and lexical information
    • Only transcript punctuation
  4. Sifadee acoustic ka caawisa in la ogaado xaaladda shucuureed macmiilka ka baxsan erayada la isticmaalo?
    • Cabbirka faylka duubista
    • Aqoonsiga shaqaalaha wakiilka
    • Siday telefoonka
    • Pitch, orodka, iyo mugga (prosody)
  5. In a two-stream call-center emotion system, what do acoustic features contribute?
    • Evidence from vocal delivery, such as how speech sounds
    • The caller’s verified identity
    • The account’s payment history
    • A guaranteed diagnosis of the caller’s emotion
  6. What task did Vidrascu and Devillers study for customer-service phone calls?
    • Predicting credit scores from call recordings
    • Speaker segmentation and emotion-related analysis of conversational turns
    • Generating live agent sales recommendations
    • Automating legal decisions about customer accounts
  7. Why might a call-center sentiment system combine lexical and acoustic features?
    • Acoustic features replace the transcript in every task
    • Lexical features reveal the phone carrier
    • The words and the way they are spoken provide complementary signals
    • Combining them guarantees the caller’s true emotion is known
  8. What does word-error rate compare in speech recognition evaluation?
    • Speaker identity against a customer account
    • Emotion labels against call duration
    • Call volume against agent schedule
    • Substitutions, deletions and insertions against the words in a reference transcript