At this week’s lab meeting, Jing Ji, will be presenting her recent work Efficiency of Top-Down Parsing of Recursive Adjunction for Tree Adjoining Grammar.
- Wednesday, December 2, 13:30–14:30 (Montreal time, UTC-4).
- Meetings are via Zoom. If you would like to attend the meeting but have not yet registered for this semester’s MCQLL meetings, please do so at this link. If you have already registered, please join using the link you received in your confirmation email.
CKY-type parser and Earley-type parser are two widely-used parsing algorithms for Tree Adjoining Grammar (TAG). In contrast, a standard top-down parser is not efficient since the looping problem occurs during both the left and right recursion of standard TAG derivation. Roark (2001) combines the top-down parser for CFG with a beam search, showing that the probabilistic top-down parser yields a perplexity improvement over previous results. In this paper, we define the stochastic tree adjoining grammar and apply the probabilistic top-down parser for CFG to TAG. Comparing the parsing efficiency of the standard and alternative TAG derivation of the recursive adjunction, we find that the alternative derivation is more efficient since it avoids the looping problem of the right recursion, increasing the parsing efficiency of our top-down parser. The result is consistent with the difference in process difficulty between left and right recursion from the psychological point of view.
Jing Ji is a first-year Ph.D. student from the Department of Linguistics interested in formal language theory and computational syntax.