learn.ifs.name

Dependency Parsing

Jurafsky & Martin, Speech and Language Processing, 3rd ed. draft, ch. 20

12 cards in “20.1 Dependency relations” · whole deck

20.1 Dependency relations

  1. What are the two arguments of a dependency relation?
    A head, the central organizing word, and a dependent, a kind of modifier.
  2. Why is labeling grammatical functions especially important in flexible word-order languages?
    In English, functions like subject and object correlate with position and constituent type, so they are partly redundant with phrase structure; in freer word-order languages the relation labels carry information phrase structure barely provides.
  3. What is Universal Dependencies (UD)?
    An open community effort to annotate dependencies and other grammatical properties across more than 100 languages with a shared inventory of 37 relations.
  4. Into which two sets can the core UD relations be divided?
    Clausal relations, describing syntactic roles with respect to a predicate (often a verb), and modifier relations, describing how words modify their heads.
  5. Give core UD relations and their meanings.
    relationmeaning
    nsubjnominal subject
    objdirect object
    iobjindirect object
    ccompclausal complement
    nmodnominal modifier
    amodadjectival modifier
    apposappositional modifier
    detdeterminer
    casepreposition, postposition, case marker
    conjconjunct
    cccoordinating conjunction
  6. Identify the relations in “We booked her the first flight to Miami.”
    We nsubj of booked; her iobj; flight obj; the det and first amod of flight; Miami nmod of flight, with to as its case marker.
  7. How is a dependency structure represented formally?
    As a directed graph \(G = (V, A)\): vertices \(V\) (usually the words; possibly punctuation, or stems and affixes) and arcs \(A\) encoding head–dependent relations and their functions.
  8. What constraints make a dependency structure a dependency tree?
    (1) A single root node with no incoming arcs; (2) every other vertex has exactly one incoming arc; (3) a unique path from the root to every vertex. So each word has one head and the structure is connected.
  9. What is projectivity?
    An arc from head to dependent is projective if the head has a path to every word between them in the sentence; a tree is projective if all its arcs are, i.e. it can be drawn with no crossing arcs.
  10. Give an example of a non-projective dependency.
    In “JetBlue canceled our flight this morning which was already late”, the arc from flight to its modifier late must cross the arc linking morning to its head; flexible word order languages have many such cases.
    canceledourflightthismorningwhich…lateflight → latecanceled → morning
    The two arcs cross: the tree is non-projective.
  11. Why does projectivity matter for dependency parsing?
    English dependency treebanks converted from phrase-structure treebanks via head rules are always projective (so wrong for non-projective cases), and transition-based parsers can only produce projective trees.
  12. How are dependency treebanks created?
    By annotators directly building dependency structures, by hand-correcting parser output, or (for some early treebanks) by deterministically converting constituency treebanks.

Source