An Algorithm to Co-Ordinate Anaphora Resolution and PPS Disambiguation Process

An Algorithm to Co-Ordinate Anaphora Resolution and PPS Disambiguation Process Saliha Azzam CRIL Ing6nierie - CAMS La Sorbonne 174, Rue de la RSpubhque 92817 Puteaux France azzam@cril-ing.fr Abstract Both anaphora resolution and preposi- tional phrase (PP) attachment are the most frequent ambiguities in natural lan- guage processing. Several methods have been proposed to deal with each phe- nomenon separately, however none of proposed systems has considered the way of dealing both phenomena We tackle this issue here, proposing an algorithm to co-ordinate the treatment of these two problems efficiently, i.e., the aim is also to exploit at each step all the results that each component can provide. 1 Introduction Several methods have been proposed to deal with anaphora resolution and prepositional phrase (PP) attachment phenomenon and separately, so that the literature is very abundant : for PPs see e.g., (Frazier and Fodor, 1979; Hobbs, 1990), and (Wilks and Huang, 1985), and for anaphora see e.g., (Carter, 1986; Reinhart, 1983) and (Sidner, 1983). However none of these methods has con- sidered the way of dealing both phenomena in the same concrete system. We propose in this paper an algorithm that deals with both phenomena, in the same anal- yser. The anaphora module pertains to the recent methods, uses a set of resolution rules based on the focusing approach, see (Sidner, 1983). These rules are applied to the conceptual representation and their output is a set of candidate antecedents. Concerning the PPs, unattached prepositions in- volve empty or unfilled roles in the Conceptual Structures (CSs), expressed in a frame-based lan- guage (Zarri, 1999). The disambiguation proce- dure alms at filling the empty roles using attach- ment rules. This work was accomplished in the context of COBALT project (LRE 61-01 ), dealing with fi- nancial news. A detailed discussion about both procedures of anaphora resolution and PP attach- ment is largely developed in (Azzam, 1994). 2 The algorithm Two of the main principles of the algorithm are : a) The algorithm is applied on the text sentence by sentence, i.e. the ambiguities of the previous sentences have already been considered (resolved or not). b) The anaphora procedure skips the resolution of a given anaphor when this anaphor is preceded by an unattached preposition. This is because the resolution rules may have an empty role as a pa- rameter, due to this unattached preposition. The resolution of the anaphor is then postponed to the second phase of anaphora resolution. The proposed procedure is based on successive calls to the anaphora module and to the PP at- tachment module. The output of each call is a set of CSs that represent the intermediate re- suits exchanged between each call and on which both modules operate in turn. The aim is to fill the unfilled roles in the CSs, due to anaphora or unattached PPs. To summarize the algorithm is: 1) Apply the anaphora module first. 2) Apply the PP attachment procedure. 3) If some anaphora.are left unresolved, apply the anaphora module again. 4) If there …

Content truncated to 500 words