Natural language processing Exam
Natural language processing - answer Getting computers to process and produce
language
Ultimate goal of nlp - answer Getting computers to understand/converse with humans
Transitive - answer ‘John read the book'
Intransitive - answer ‘The book reads well'
Information Retrieval - answer Type in a word and get huge list of websites containing
that word
Websites ranked in order of...(information retrieval) - answer Keywords, morphological
forms, synonyms, and concepts related to key words
Information summarization - answer Relevant words, phrases, or sentences are
included to get best result
Spell checkers/grammar - answer Computer needs to recognize 'I'm' and 'Im' are
related
Automatic translation- source lang to target lang 1940's - answer Dictionary basis, word
for word translation (word morpheme)
Automatic translation problems 1940s - answer Word order e.g. The red house > la
class roja. Metaphors, idioms, target Lang may not have equivalent word
Google translate- machine learning algorithms - answer Detects millions of patterns in
documents to guess most likely translation. Users can validate transaction so computer
recognized it's correct
Text-to-Speech - answer Written text to speech. Text to tokenized & preprocessed to
detect names + numbers.
Useful applications - answer Navigation systems, automated announcements, reading
aids for blind people
Morphology - answer The structure of words
Automatic morphological analyzer (AMA) - answer Take a word in an language and
break down into its morphemes.
Natural language processing - answer Getting computers to process and produce
language
Ultimate goal of nlp - answer Getting computers to understand/converse with humans
Transitive - answer ‘John read the book'
Intransitive - answer ‘The book reads well'
Information Retrieval - answer Type in a word and get huge list of websites containing
that word
Websites ranked in order of...(information retrieval) - answer Keywords, morphological
forms, synonyms, and concepts related to key words
Information summarization - answer Relevant words, phrases, or sentences are
included to get best result
Spell checkers/grammar - answer Computer needs to recognize 'I'm' and 'Im' are
related
Automatic translation- source lang to target lang 1940's - answer Dictionary basis, word
for word translation (word morpheme)
Automatic translation problems 1940s - answer Word order e.g. The red house > la
class roja. Metaphors, idioms, target Lang may not have equivalent word
Google translate- machine learning algorithms - answer Detects millions of patterns in
documents to guess most likely translation. Users can validate transaction so computer
recognized it's correct
Text-to-Speech - answer Written text to speech. Text to tokenized & preprocessed to
detect names + numbers.
Useful applications - answer Navigation systems, automated announcements, reading
aids for blind people
Morphology - answer The structure of words
Automatic morphological analyzer (AMA) - answer Take a word in an language and
break down into its morphemes.
