Lexical Semantics

  • Lemma: basic word form (paired with pos)

    • e.g break, breaks, breaking, broke, broken

  • Word Form: word how it actually occurs

    • e.g bases is the plural of basis and base

  • Sense: refer to difference meaning of a word in a given context

    • Wordnet:

      • Sense is defined by synset, a set of words that share a same meaning

      • Senses are hierachical

        • furniture -> seat -> chair -> recliner

  • Pointwise Mutual Information

    • word word matrix (word embedding)

      • rows is word

      • columns is word co-occuring with word

    • Scores in matrix

      • how related is word1 and word2 represented by column c

    • The Curse of Dimensionality

      • as dimension rises, the sparse rises too

      • computationally heavy