G22.2590 - Natural Language Processing  -- Spring 2008 -- Prof. Grishman

Assignment #1

January 24, 2008

For each of 3 newsworthy events for which a 'central' fact can be captured in a single sentence:

  • find 3 or 4 news sources which report this fact
  • list the corresponding sentence from each source
  •  within each sentence, underline the contiguous portion which conveys this fact
  • state what sort of linguistic analysis would be needed to recognize that these are instances of the same fact or event:
  •  name recognition (to handle different forms of a name)
  • syntactic analysis (to identify modifiers present in one case but not another;  active vs. passive;  etc.)
  • lexical semantics (to recognize synonyms)
  • more complex semantic relationships
  • reference resolution (for pronouns, referring noun phrases)
  • Due January 31st

    Credit:  3 points

    (1 point for each event ...
      be sure you include 3 separate events, and at least 3 sentences (from different articles) expressing each event)

    Examples (all sources reached through finance.yahoo.com)

    Motley Fool
    1. Nestle Bags Ralston Purina
    2. Shares of pet food giant Ralston Purina jumped this morning on news that global comestibles leader Nestle, a Swiss company, agreed to buy the company for $37.50 per share in cash.

    3. Nestle Acquires Ralston Purina
    4. Nestle said yesterday that it would pay $10.3 billion for St. Louis-based Ralston Purina, maker of Puppy Chow, Dog Chow, Meow Mix, and other favorites of the four-leged set.

    AP Financial
    5. Nestle Buys Ralston Purina for $10B
    6. Food giant Nestle S.A. is buying pet food maker Ralston Purina for $10 billion in cash, linking such brands as Alpo, Purina, Friskies, and Meow Mix.

    some paraphrase relations:
    name:  Nestle vs. Nestle S.A. (#6)
    syntax:  ignore "the Swiss company" (#2);  "buys" (#5) vs. "is buying" (#6)
    lexical semantics:  "bag" (#1) vs. "acquire" (#3) vs. "buy" (#2, 5, 6)
    other semantic:  "agreed to buy" (#2) vs. "buy" (#5, 6)
    reference:  "the company" (#2), "it" (#4)