Cascaded Grammatical Relation Assignment

Cascaded Grammatical Relation Assignment Sabine Buchholz and Jorn Veenstra and Walter Daelemans ILK, Computational Linguistics, Tilburg University PO box 90153, 5000 LE Tilburg, The Netherlands [buchholz, veenstra, dae lemans] 0kub. nl Abstract In this paper we discuss cascaded Memory- Based grammatical relations assignment. In the first stages of the cascade, we find chunks of sev- eral types (NP,VP,ADJP,ADVP,PP) and label them with their adverbial function (e.g. local, temporal). In the last stage, we assign gram- matical relations to pairs of chunks. We stud- ied the effect of adding several levels to this cas- caded classifier and we found that even the less performing chunkers enhanced the performance of the relation finder. 1 Introduction When dealing with large amounts of text, find- ing structure in sentences is often a useful pre- processing step. Traditionally, full parsing is used to find structure in sentences. However, full parsing is a complex task and often pro- vides us with more information then we need. For many tasks detecting only shallow struc- tures in a sentence in a fast and reliable way is to be preferred over full parsing. For example, in information retrieval it can be enough to find only simple NPs and VPs in a sentence, for in- formation extraction we might also want to find relations between constituents as for example the subject and object of a verb. In this paper we discuss some Memory-Based (MB) shallow parsing techniques to find labeled chunks and grammatical relations in a sentence. Several MB modules have been developed in previous work, such as: a POS tagger (Daele- mans et al., 1996), a chunker (Veenstra, 1998; Tjong Kim Sang and Veenstra, 1999) and a grammatical relation (GR) assigner (Buchholz, 1998). The questions we will answer in this pa- per are: Can we reuse these modules in a cas- cade of classifiers? What is the effect of cascad- ing? Will errors at a lower level percolate to higher modules? Recently, many people have looked at cas- caded and/or shallow parsing and GR assign- ment. Abney (1991) is one of the first who pro- posed to split up parsing into several cascades. He suggests to first find the chunks and then the dependecies between these chunks. Grefen- stette (1996) describes a cascade of finite-state transducers, which first finds noun and verb groups, then their heads, and finally syntactic functions. Brants and Skut (1998) describe a partially automated annotation tool which con- structs a complete parse of a sentence by recur- sively adding levels to the tree. (Collins, 1997; Ratnaparkhi, 1997) use cascaded processing for full parsing with good results. Argamon et al. (1998) applied Memory-Based Sequence Learn- ing (MBSL) to NP chunking and subject/object identification. However, their subject and ob- ject finders are independent of their chunker (i.e. not cascaded). Drawing from this previous work we will explicitly study the effect of adding steps to the grammatical relations assignment cascade. Through experiments with cascading several classifiers, we will show that even using im- perfect classifiers can improve overall …

Content truncated to 500 words