Coreference Resolution and Entity Linking

24.1 Coreference phenomena: linguistic background

1. What are the four types of referring expressions?
   Indefinite NPs, definite NPs, pronouns (including demonstratives and zero anaphors), and names.
   • Jurafsky & Martin, Speech and Language Processing, 3rd ed. draft, ch. 24, §24.1, p. 542
2. What do indefinite NPs typically do in discourse?
   Introduce entities new to the hearer: a beautiful goose, some walnuts, colloquial this beautiful cauliflower.
   • Jurafsky & Martin, Speech and Language Processing, 3rd ed. draft, ch. 24, §24.1, p. 542
3. What makes the referent of a definite NP identifiable?
   Prior mention in the text (the white stallion), the hearer’s world knowledge (the New York Times), or uniqueness implied by the description (the car keys). Over half of newswire definite NPs are non-anaphoric.
   • Jurafsky & Martin, Speech and Language Processing, 3rd ed. draft, ch. 24, §24.1, p. 542
4. What are cataphora and bound pronouns?
   Cataphora: the pronoun comes before its referent (“Even before she saw it, Dorothy had been thinking about the Emerald City”). Bound: the pronoun acts like a variable bound by a quantifier (“Every dancer brought her left arm forward”).
   • Jurafsky & Martin, Speech and Language Processing, 3rd ed. draft, ch. 24, §24.1, p. 542
5. What is a zero anaphor?
   An anaphor with no lexical realization, possible in languages like Chinese, Japanese and Italian: “Giovanni andò a far visita a degli amici. Per via φ comprò del vino.” Zero anaphors complicate mention detection.
   • Jurafsky & Martin, Speech and Language Processing, 3rd ed. draft, ch. 24, §24.1, p. 543
6. What are the information status categories of Prince (1981)?

   type	discourse	hearer	example
   brand new	new	new	a fruit
   unused	new	old	Marie Curie
   old (evoked)	old	old	it after “a new restaurant”
   inferrable	new	new, but inferable	the chef of a restaurant

   • Jurafsky & Martin, Speech and Language Processing, 3rd ed. draft, ch. 24, §24.1, pp. 543–544
7. What is a bridging inference?
   Inferring that a new entity belongs to the discourse model through world knowledge about other entities: “I went to a superb restaurant. The chef had just opened it.”; “Mix flour, butter and water. Knead the dough.”
   • Jurafsky & Martin, Speech and Language Processing, 3rd ed. draft, ch. 24, §24.1, p. 544
8. How does an entity’s salience (accessibility) affect the form of its mentions?
   Very accessible referents can be referred to with less material (pronouns need high salience). First mentions tend to be long (full name, title, appositive: Victoria Chen, CFO of Megabucks Banking); later ones shorter (Ms. Chen, the 38-year-old, she), though not monotonically.
   • Jurafsky & Martin, Speech and Language Processing, 3rd ed. draft, ch. 24, §24.1, p. 544
9. Why can’t a car in “Janet doesn’t have a car” be referred back to?
   Under negation the NP doesn’t create a discourse referent (Karttunen 1969), so “*It is a Toyota” is bad.
   • Jurafsky & Martin, Speech and Language Processing, 3rd ed. draft, ch. 24, §24.1, p. 544
10. Which four structures look like NPs but are not mentions in coreference tasks?
    Appositives (United, a unit of UAL,), predicative/attributive NPs (became the company’s president, jump to $2.3 million), expletive/pleonastic pronouns (it is raining, clefts, extraposition), and generics (“I love mangos. They are very tasty”, generic you).
    • Jurafsky & Martin, Speech and Language Processing, 3rd ed. draft, ch. 24, §24.1, p. 545
11. Why can’t coreference systems enforce number agreement strictly?
    Semantically plural entities can be it or they (“IBM announced… They have been working on it”), and singular they refers to individuals.
    • Jurafsky & Martin, Speech and Language Processing, 3rd ed. draft, ch. 24, §24.1, pp. 545–546
12. What exception to person agreement does quotation cause?
    In “‘I voted for Nader because he was most aligned with my values,’ she said”, I, my and she corefer.
    • Jurafsky & Martin, Speech and Language Processing, 3rd ed. draft, ch. 24, §24.1, p. 546
13. What do binding theory constraints say about reflexive and non-reflexive pronouns?
    Reflexives corefer with the subject of the most immediate clause (“Janet bought herself…”, herself = Janet); non-reflexives cannot (“Janet bought her…”, her ≠ Janet).
    • Jurafsky & Martin, Speech and Language Processing, 3rd ed. draft, ch. 24, §24.1, p. 546
14. How do recency and grammatical role affect pronoun resolution?
    Recent entities are more salient; subjects are more salient than objects, which are more salient than obliques: “Billy Bones went to the bar with Jim Hawkins. He called for rum.” He = Billy; swap the subject and he = Jim.
    • Jurafsky & Martin, Speech and Language Processing, 3rd ed. draft, ch. 24, §24.1, p. 546
15. How does verb semantics bias pronoun interpretation in “John telephoned/criticized Bill. He lost the laptop”?
    With telephoned, he = John; with criticized, he = Bill. Implicit causality: the implicit cause of criticizing is its object, of telephoning its subject, and the implicit cause is more salient.
    • Jurafsky & Martin, Speech and Language Processing, 3rd ed. draft, ch. 24, §24.1, p. 547
16. How do selectional restrictions help in “I ate the soup in my new bowl after cooking it for hours”?
    Verbs restrict their arguments: food, not a bowl, is what one eats and cooks for hours, so it = the soup.
    • Jurafsky & Martin, Speech and Language Processing, 3rd ed. draft, ch. 24, §24.1, p. 547