Linguistics 110 Linguistic Analysis: Sentences & Dialects
Lecture Number Twenty Five
Speech Synthesis

  1. Computers and the Brain

    1. Both are based on electric toggles: either on or off
      1. Computers crunch numbers
      2. Brains talk

    2. Artificial Intelligence (AI) Hypothesis
      1. Weak Hypothesis: a computer can communicate with people in such a way that the people will not know that they are communicating with a machine.
      2. Strong Hypotheis: a computer can speak exactly like human beings speak

    3. Applications of Computational Linguistics
      1. Speech Synthesis
      2. Machine Translation
      3. Artificial Intelligence
      4. Voice identification

  2. Computational Linguistics

    1. Neural networks (Rumelhart & McClelland)
      1. Learn behaviors without being programmed differentially.
      2. If neural network can learn language without special programming, Innateness Hypothesis is undermined.
      3. If Innateness Hypothesis is valid, special program corresponding to UG will have to be coded into neural network for it to be able to learn language.

    2. Components of a Computational Speaker
      1. Phonetic interpreter (speech synthesis)
      2. Lexeme-Morpheme transducer
      3. Lexicon and Morphological Component
      4. Syntax transducer
      5. Semanic interpreter
      6. Cognitive processor

  3. Conclusion

    © 1996-2000, Inc.