Accurate Unlexicalized Parsing

2 Vertical and Horizontal Markovization

1. What are the two basic deficiencies of the raw treebank PCFG?
   (1) Categories are too coarse to make expansions independent of context: a subject NP is 8.7 times more likely than an object NP to be just a pronoun. (2) Sparsity: many rules seen once are overestimated, and many rules in test sentences were never seen.
   • Klein & Manning, 2003, Accurate Unlexicalized Parsing (Proceedings of the 41st Annual Meeting of ACL), §2, pp. 425–426
2. What are vertical and horizontal markovization orders \(v\) and \(h\)?
   Vertical: how many ancestors condition an expansion (\(v = 2\) is parent annotation, e.g. NP^S vs NP^VP). Horizontal: how many previously generated siblings condition the next child, generating right-hand sides outward from the head. The raw treebank grammar is \(v = 1\), \(h = \infty\), a historical accident; increasing \(v\) and decreasing \(h\) both help.
   • Klein & Manning, 2003, Accurate Unlexicalized Parsing (Proceedings of the 41st Annual Meeting of ACL), §2, pp. 425–426
3. Show the \(v = 1\), \(h = 1\) markovization of VP → VBZ NP PP.
   Intermediate symbols remember the head and the last generated sibling: ⟨VP:[VBZ]⟩ → VBZ; ⟨VP:[VBZ]…NP⟩ → ⟨VP:[VBZ]⟩ NP; ⟨VP:[VBZ]…PP⟩ → ⟨VP:[VBZ]…NP⟩ PP; VP → ⟨VP:[VBZ]…PP⟩. These are ordinary PCFG rules.

   • Klein & Manning, 2003, Accurate Unlexicalized Parsing (Proceedings of the 41st Annual Meeting of ACL), §2, pp. 425–426
4. What did the markovization grid show?
   More vertical history helped, even exhaustive grandparent annotation, while for horizontal history a variable-order second-order model (\(h \le 2\)) was best. Best cell: \(v = 3\), \(h \le 2\), 79.74 F1 vs 72.62 for the raw grammar. Further work used \(v \le 2\), \(h \le 2\) (77.77 F1), the best compromise of grammar size and useful history.
   • Klein & Manning, 2003, Accurate Unlexicalized Parsing (Proceedings of the 41st Annual Meeting of ACL), §2, pp. 425–426