Topic 7: Intro to text and speech analytics

Question 1

NLP

Answer 1

this is about what goes into getting computers to perform useful and interesting task involving human languages

concerned with insights that such computational work gives us into human processing of language

Question 2

Why NLP?

Answer 2

• enourmous amount of knowledge is available in machine readable form as natural language text
• conversational agents are becoming important form of human-computer interaction
• much human-human communication is now mediated by computers

Question 3

Application of NLP

Answer 3

• Google translate
• web Q/A
• weblog analytics (product marketing info, political opinion tracking, social network analysis, buzz analysis)
• web analytics ( cambridge analytica mine data to analyse the strategic communication during electoral processes.

Question 4

Major topics

Answer 4

1. Words
2. Syntax
3. Meaning

Question 5

what makes application a language processing application?

Answer 5

• requires the use of knowledge about human language

word count yes
count line / bytes no

Question 6

Big application of NLP

Answer 6

• Q/A
• conversational agents
• summarization
• machine translation

Question 7

what is needed for NLP?

Answer 7

• speech recognition and synthesis
• knowledge of the english words
• how group of words clump
• dialog

Question 8

knowledge required in HAL

Answer 8

• determine what user is saying. (phonetics and phonology)
• recognizing variation of words (morphology)
• understand syntax / structure of sentence
• semantics meaning understanding
• implicit usage of languages . pragmatics
• discourse. managing larger linguistic units

Question 9

why NLP is considered part AI solution?

Answer 9

• dealing with ill-defined problems
• don’t have exact solutions/algorithm

aim of conversational agents is to have spontaneous dialogue

Question 10

categories of knowledge in NLP

Answer 10

phonology
morphology
syntax
semantics
pragmatics
discourse

Question 11

why NLP difficult?

Answer 11

• non-standard english
• segmentation issues
• idioms
• neologisms (newly coined words/expression)
• multilingual
• tricky entity names

Question 12

ambiguity

Answer 12

computational linguist are obsessed with ambiguity.
fundamental problem
resolving ambiguity is an important goal

example

i made her duck

POS
Word sense disambiguation
Syntactic disambiguation

Question 13

Text and speech application

Answer 13

small (standalone)

• spelling correction
• hyphenation

medium (enabler app)

• WSD
• name entity recognition
• information retrieval

large (funding/business plans)

• Q/A
• conversational agent
• machine translation
• speech-to-speech translation
• context to speech system