12/3/2013
Unit 9: Human
Language Processing
3 Dec 2013
How does it look like in the real
world?
• SIRI
• Watson
• Language Analysis
• Vocaloid
• Google Translate
• Spam Detection
• Voice Automated services
• Robots
• Chat bots (ELIZA)
The Speech Chain: The Brain-to-Brain Linking
1
12/3/2013
Computers that talk and listen
Computational Linguistics
• Interaction between language and computers in all
dimensions,
• from phonetics to pragmatics
• from producing speech to comprehending speech
• from spoken (or signed) utterances to written forms.
Computational Phonetics &
Phonology
Speech Recognition Speech Synthesis
Creation of electronic signals that
Analysis of speech signals
simulate phones & Prosodic features
Output: Electronic speaker, and
Product: Phonetic Transcription of
further processing (speech
speech
generation application)
2
12/3/2013
Speech Recognition
• Earlier versions of speech recognisers do not hear each sound,
but they are programmed to store acoustic patterns of entire
words or phrases
• They search for patterns, and identify the meanings intended.
• Currently, even phone systems have speech recognition (eg
telebanking services).
Speech Synthesis
URL: [Link]
• Produces machine-like speeches
• Combines knowledge from acoustic theory of speech & technology
developments in electronics
• Able to imitate process of producing sound signals physically through
use of electronic tools
• Still have weaknesses in simulating changing intonations and stress
patterns in speech
Computational Syntax
• ELIZA: 1st human-machine communication invented by J
Weizenbaum
• using syntax (print) simulating a psychiatric session
• Circuit-Fix-It-Shop: NCSU & DU repair tech programmed
speech
• Capable of understanding & speaking complex utterances
• Computer parser
3
12/3/2013
Circuit-Fix-It-Shop
• Constructed at Duke University, USA
• Assisted users in repairing an electronic circuit
• 125 word vocabulary, speaker-dependent connected speech
• Varying initiative fixed per interaction
• 141 dialogs, 2840 user utterances
• 114 or 84% of the dialogs were successfully completed
• 50% utterance recognition rate
• 81.5% correct interpretation rate
Sample Interaction – Part 1
C: This is the Circuit Fix-It Shop. How may I help you?
U: This RS111 circuit is not working.
C: I am familiar with that circuit. The LED is supposed to be displaying
alternately flashing one and seven.
U: The LED is flashing seven.
C: Did you mean to say that the LED is flashing?
U: The LED is displaying a flashing seven.
C: Is the one on the LED displaying?
U: No.
C: Is anything else on the LED displaying other than the NOT** flashing seven?
U: The seven is flashing.
C: Is anything else on the LED displaying other than the flashing seven?
U: No.
C: What is the switch at when the LED is displaying only a flashing seven?
U: It is up.
C: What is the knob at when the LED is displaying only a flashing seven?
U: The knob is at ten.
Sample Interaction – Part 2
C: Is there a wire between connector one eight and connector four
one?
U: Where is connector four one?
C: Below the white knob with the orange circle is a pink box, called
transistors. Now, in the upper right of this pink box, called
transistors is connector four one.
U: The wire is not there.
C: Did you mean to say that there IS** a wire between connectors one
eight and four one?
U: There is no wire.
C: Add a wire between connector one eight and connector four one.
4
12/3/2013
Let’s try Eliza
[Link]
Computational
Morphology
• Computational morphology deals with the processing of
words and word forms, in both their graphemic, i.e., written
form, and their phonemic, i.e., spoken form. It has a wide
range of practical applications.
• Computers need to understand the inter-weaving of rules,
exceptions & morpheme & word structure
• Computer’s dictionary: morphological forms – needs
continual updating
Computational Morphology
• Form predictability: impossible for compounding –
sky+box= skybox
• Component morpheme
• Monomorpheme or not – [reZENT] or [Resent]
• Heteronyms - lead [leed] & [led]
• Examples in use:
• Auto-correct on your mobile phones
• Hyphenations in printed newspapers
5
12/3/2013
NLP – natural language processing
• Concerned with interactions between computers and natural
human languages, specifically in extracting meaningful
information from natural language or producing meaningful
natural language
• Related to artificial intelligence, machine learning, machine
translations etc
