NLP Top Exam Review, 100% Accurate answers. Verified.

NLP Top Exam Review, 100% Accurate answers. Verified. grammar - -set of rules that describes what's allowable in a language classical grammar - -meant for humans; cannot be programmed explicit grammars - -can be programmed and tested on texts and data phonology - -study of the sounds that make up words orthography - -study of the letters that make up (written) words syntax - -the way words are grouped together into larger constituents and the way these phrases can be ordered semantics - -context-independent meaning of utterances pragmatics - -context-dependent meaning of utterances; knowledge of the relationship of meaning to the goals/intentions of the speakers lemma - -the citation form or dictionary form of a word phoneme - -smallest unit of sound that can stand on its own; abstractions rather than physical segments morpheme - -parts of a word, e.g. delive-rable clitic - -phrasal affix, e.g. `'s` phone - -instances of phonemes, e.g. madder and matter in English inflection - -modification of a root form to indicate grammatical distinctions (gener, number, case, tense. etc) derivation - -change of syntactic category, e.g. adding -ly to make an adverb anaphor - -something that refers to something close by in the text, e.g. "each other" hypernym - -word with a more general sense, e.g. animal is a hypernym of cat hyponym - -word with a more specific sense, e.g. cat is a hyponym of animal meronym - -a compositional part of a whole, e.g. tire makes up a car homonym - -words written the same way with different meanings homophones - -subset of homonyms that are also pronounced the same way holonym - -whole corresponding to the parts, e.g. car is a holonym of tire collocation - -pairing of two words that gives different meanings: white wine vs. white hair phonetics and phonology - -are concerned with how words are pronounced as sequences of sounds compositional semantics - -meaning of a word based on what it's composed of, e.g. "Western Europe" Turing test - -the canonincal benchmark for evaluating AI systems: can a human tell whether he/she is conversing with another person or a machine? ? operator in regex - -makes preceding sequence optional, i.e. matches "zero or one instances"