Information Extraction: Relations, Events, and Time

21.7 Automatic temporal analysis

1. What are the three common steps in analyzing time in text?
   (1) Extract temporal expressions; (2) normalize them to a standard format; (3) link events to times and build time graphs or timelines.
   • Jurafsky & Martin, Speech and Language Processing, 3rd ed. draft, ch. 21, §21.7, p. 490
2. What are absolute, relative and duration temporal expressions?
   Absolute ones map directly to calendar dates or times (April 24, 1916, 10:15 AM); relative ones map via a reference point (yesterday, next semester, two weeks from yesterday); durations are spans (four hours, three weeks).
   • Jurafsky & Martin, Speech and Language Processing, 3rd ed. draft, ch. 21, §21.7, p. 490
3. What are temporal lexical triggers?
   Words heading temporal expressions: nouns (morning, winter), proper nouns (January, Monday, Ramadan), adjectives (recent, annual), adverbs (hourly, daily).
   • Jurafsky & Martin, Speech and Language Processing, 3rd ed. draft, ch. 21, §21.7, pp. 490–491
4. How are temporal expressions detected?
   Rule-based cascades of regular expressions over POS, trigger words and classes (e.g. SUTime), or IOB sequence labelers using embeddings, fine-tuned encoders or classic features.
   • Jurafsky & Martin, Speech and Language Processing, 3rd ed. draft, ch. 21, §21.7, p. 491
5. What causes false positives in temporal expression recognition?
   Lexical triggers used non-temporally: “1984 tells the story of Winston Smith”, “U2’s Sunday Bloody Sunday”.
   • Jurafsky & Martin, Speech and Language Processing, 3rd ed. draft, ch. 21, §21.7, p. 491
6. What is temporal normalization?
   Mapping a temporal expression to a point in time (calendar date, time of day) or a duration, represented in the ISO 8601 standard.
   • Jurafsky & Martin, Speech and Language Processing, 3rd ed. draft, ch. 21, §21.7, p. 491
7. Normalize “last week” and “the weekend” in an article dated July 2, 2007 (ISO week 27).
   last week → 2007-W26 (a date of type week); the weekend → duration P1WE, anchored to that week; two weeks → P2W. The dateline 2007-07-02 is the document’s creation time.
   • Jurafsky & Martin, Speech and Language Processing, 3rd ed. draft, ch. 21, §21.7, p. 492
8. What do ISO 8601 patterns like YYYY-Wnn and P3Y mean?
   YYYY-Wnn is week nn of the year (weeks start Monday; week 1 contains the year’s first Thursday). Pnx is a duration of n units x: P3Y three years, P2D two days, P1WE one weekend.
   • Jurafsky & Martin, Speech and Language Processing, 3rd ed. draft, ch. 21, §21.7, p. 492
9. How do rule-based normalizers work?
   Patterns matching temporal expressions are paired with semantic procedures, e.g. the pattern for “3 years old” maps to Duration(3, year).
   • Jurafsky & Martin, Speech and Language Processing, 3rd ed. draft, ch. 21, §21.7, p. 492
10. What is a document’s temporal anchor, and how is it used?
    Its dateline: relative expressions like today, yesterday, tomorrow are computed from it with (modular) temporal arithmetic.
    • Jurafsky & Martin, Speech and Language Processing, 3rd ed. draft, ch. 21, §21.7, pp. 492–493
11. Why is normalizing “the weekend” or “next Friday” hard?
    Their reference depends on context: the weekend may be the past or coming one depending on the governing verb’s tense; next Friday may mean the nearest Friday or the one after, depending on how close the document date is to a Friday.
    • Jurafsky & Martin, Speech and Language Processing, 3rd ed. draft, ch. 21, §21.7, p. 493
12. What is the temporal ordering task, and why does it need event detection?
    Imposing a (partial) order on the events and times in a text, as binary relation classification. Many times are anchored to events (“one week after the storm”), so events must be detected first.
    • Jurafsky & Martin, Speech and Language Processing, 3rd ed. draft, ch. 21, §21.7, p. 493
13. How are temporal links between events and times predicted?
    By classifiers trained on TimeBank gold TLINKs between pairs of events/times (and the document time), using words/embeddings, parse paths, tense and aspect; sieve architectures are also used.
    • Jurafsky & Martin, Speech and Language Processing, 3rd ed. draft, ch. 21, §21.7, pp. 493–494