[Ligncse256] More HW3 questions

[Roger Levy]

Just a quick note on question 3:

Ben Cipollini wrote:

>     3. Did we cover "Agenda-Based" parsers in class?  If so, did we call
>     them agenda-based parsers?  If not, what did we call them?

I didn't actually introduce the term "agenda-based" parser, but it 
simply means a type of parser in which every time you construct an item, 
it gets placed on an "agenda" (a list of items to be processed), and you 
process each item on the agenda exactly once, where "processing" the 
item means figuring out all the ways it can be combined with every other 
item that exists.  Usually, the item doesn't get put in the chart until 
it's been processed.

Actually, bottom-up CKY is not necessarily best thought of as an 
agenda-based parser, but the Earley algorithm is -- when an item is 
constructed, it goes on the agenda, and its predictions are all made 
before it goes in the chart.

Roger


-- 

Roger Levy                      Email: rlevy at ucsd.edu
Assistant Professor             Phone: 858-534-7219
Department of Linguistics       Fax:   858-534-4789
UC San Diego                    Web:   http://ling.ucsd.edu/~rlevy