Flashcards
NAturalLAnguageProcessingAMIA22
Cards in this deck (18)
- 1define Natural Language Processing?Automated methods containing linguistic knowledge that aim to improve the management of information in Text - derive "meaning" from text
- 2Define Automatic Speech RecognitionUse of computer hardware /software based techniques to identify and process human voice sound --> text
- 3What are the 2 main broad categories of use case for NLP?1) Named -Entity Recognition -ie. problems, Meds, ADEs
2) Relation Extraction - i.e. med attribute relations (dose, sig, route) , Drug-drug interactions - 4Name lower level NLP Tasks?Sentence boundary detection
Tokenization
Part of Speech assignment
Morphologic decomposition of compound words
Shallow parsing (chunking)
Problem specific segmentation - 5What is Tokenization?identifying individual tokens (words, punctuation) within a sentence
- 6What is Sentence Boundary detection?finding the beginning and end of a sentence
- 7What detects tokens in text?A Lexer-
- 8what complicates part of speech assignment?homographs (i.e. cold, set, etc)
Gerunds ( versb ending in "ing" convert verbs into nouns for example) - 9What is "chunking" in NLP?identifying phrases from constituent Part of Speech tagged tokens - i.e. a noun phrase = adjective sequence followed by a noun
- 10What are "high" level tasks of NLP?spelling/grammatical error identification
Named entity recognition
word sense disambiguation
negation and uncertainty detection
relationship extraction
Temporal inferences/relationship
Information extraction - 11What is Named Entity recognition?identifying specific words/phrases (entities) and categorizing them (i..e into people, locations, diseases, genes, medications etc)
- 12What is Word Sense Disambiguation?determine a Homograph's correct meaning
- 13What is temporal inferencing?determine time of events i.e. med X prescribe after symptoms begin.
- 14What is NLP information extractionidentification of problem specific information and its transformation into structured form
- 15what is co-referencewhen you first refer to an entity such as bob and then later say "he" and to understand that he refers to bob
- 16what are challenges to Named Entity recognition NLP tasks?Linguistic variation - different words, same meaning
Polysemy - 1 word, multiple meanings
Finding validation
implication - 17What is an example of Contextual Attribute assignment problems?negation, uncertainty and temporality
- 18what are challenges to discourse processing?formating /report structures, co-reference
Information
- Description
- informaticsBoards DAtaAnalytics Flashcards on NAturalLAnguageProcessingAMIA22, created by Michael Riben on 12/10/2022.
- Created by
- Michael Riben
- Views
- 0
- Rating