Dialog Design 3

1
Dialog Design 3

• Speech and natural language

2
Agenda

• Video
• Speech generation
• Speech recognition
• Natural language interfaces
• Project part 3
• Mid-term exam preview

3
Speech Input

• Speech synthesis
• Speaker recognition
• Speech recognition
• Natural language understanding

4
English Speech

• Made up of 40 phonemes, 24 consonants and 16
  vowels

5
Speech Synthesis

• Often hear robotic voice
• Store tones, then put them together
• -gt The transition is the difficult thing to do

6
Speaker Recognition

• Tell which person it is (voice print)
• Could be important for monitoring meetings

7
Speech Recognition

• Primarily identifying words
• Improving all the time
• Commercial systems
• IBM ViaVoice, Dragon Dictate, ...

8
Recognition Dimensions

• Speaker dependent/independent
• Parametric patters are sensitive to speaker
• With training (dept) can get better
• Vocabulary
• Some are getting 50,000 words
• Isolated word vs. continuous speech
• Continuous where words stop begin
• Typically a pattern match, no context used

9
Recognition Systems

• Typical system has 5 components
• Speech capture device - Has analog -gt digital
  converter
• Digital Signal Processor - Gets word boundaries,
  scales, filters, cuts out extra stuff
• Preprocessed signal storage - Processed speech
  buffered for recognition algorithm
• Reference speech patterns - Stored templates or
  generative speech models for comparisons
• Pattern matching algorithm - Goodness of fit from
  templates/model to users speech

10
Errors

• Systems make four types of errors
• Substitution - one for another
• Rejection - detected, but not recognized
• Insertion - added
• Deletion - not detected
• Which is more common, dangerous?

11
Natural Language Understanding

• Putting meaning to the words
• Input might be speech or could be typed in
• Holy grail of Artificial Intelligence problems

12
NL Factors/Terms

• Syntactic
• Grammar or structure
• Prosodic
• Inflection, stress, pitch, timing
• Pragmatic
• Situated context of utterance, location, time
• Semantic
• Meaning of words

13
SR/NLU Assessment

• Advantages

• Disadvantages

14
SR/NLU Advantages

• Easy to learn and remember
• Powerful
• Fast, efficient (not always)
• Little screen real estate

15
SR/NLU Disadvantages

• Doesnt work good enough yet
• Assumes knowledge of problem domain
• Not prompted, like menus
• Requires typing skill (if keyboard)
• Enhancements are invisible
• Expensive to implement

16
Good in Situations

• Hands busy
• Mobility required
• Eyes occupied
• Conditions preclude use of keyboard
• Visual impairment
• Physical limitation

17
Project Part 3

• Decide on a design
• Implement a prototype
• (Develop evaluation plan)

18
Mid-term Exam Review

• Lectures
• Chapters 1-7, 11.1-11.4
• Short answer style questions