Resolving PP attachment Ambiguities with Memory-Based Learning

Resolving PP attachment Ambiguities with Memory-Based Learning Jakub Zavrel, Walter Daelemans, Jorn Veenstra Computational Linguistics, Tilburg University PO Box 90153, 5000 LE Tilburg, The Netherlands zavrel, walter, veenstra©kub, nl Abstract In this paper we describe the application of Memory-Based Learning to the problem of Prepositional Phrase attachment disam- biguation. We compare Memory-Based Learning, which stores examples in mem- ory and generalizes by using intelligent sim- ilarity metrics, with a number of recently proposed statistical methods that are well suited to large numbers of features. We evaluate our methods on a common bench- mark dataset and show that our method compares favorably to previous methods, and is well-suited to incorporating vari- ous unconventional representations of word patterns such as value difference metrics and Lexical Space. 1 Introduction A central issue in natural language analysis is structural ambiguity resolution. A sentence is structurally ambiguous when it can be assigned more than one syntactic structure. The drosophila of structural ambiguity resolution is Prepositional Phrase (PP) attachment. Several sources of infor- mation can be used to resolve PP attachment am- biguity. Psycholinguistic theories have resulted in disambiguation strategies which use syntactic infor- mation only, i.e. structural properties of the parse tree are used to choose between different attachment sites. Two principles based on syntactic informa- tion are Minimal Attachment (MA) and Late Clo- sure (LC) (Frazier, 1979). MA tries to construct the parse tree that has the fewest nodes, whereas LC tries to attach new constituents as low in the parse tree as possible. These strategies always choose the same attachment regardless of the lexical content of the sentence. This results in a wrong attachment in one of the following sentences: 1 She cats pizza with a fork. 2 She cats pizza with anchovies. In sentence 1, the PP "with a fork" is attached to the verb "eats" (high attachment). Sentence 2 dif- fers only minimally from the first sentence; here, the PP "with anchovies" does not attach to the verb but to the NP "pizza" (low attachment). In languages like English and Dutch, in which there is very little overt case marking, syntactic information alone does not suffice to explain the difference in attachment sites between such sentences. The use of syntac- tic principles makes it necessary to re-analyse the sentence, using semantic or even pragmatic infor- mation, to reach the correct decision. In the exam- ple sentences 1 and 2, the meaning of the head of the object of 'with' determines low or high attach- ment. Several semantic criteria have been worked out to resolve structural ambiguities. However, pin- ning down the semantic properties of all the words is laborious and expensive, and is only feasible in a very restricted domain. The modeling of pragmatic inference seems to be even more difficult in a com- putational system. Due to the difficulties with the modeling of se- mantic strategies for ambiguity resolution, an at- tractive alternative is to look at the statistics of word patterns in annotated corpora. In such a cor- pus, …

Content truncated to 500 words