Capturing the Interaction between Aggregation and Text Planning in Two Generation Systems

In natural language generation, different generation tasks often interact with each other in a complex way. We think that how to resolve the complex interactions inside and between tasks is more important to the generation of a coherent text than how to model each individual factor. This paper focuses on the interaction between aggregation and text planning, and tries to explore what preferences exist among the features considered by the two tasks. The preferences are implemented in two generation systems, namely ILEX-TS and a text planner using a Genetic Algorithm. The evaluation emphasises the second implementation and shows that capturing these preferences properly can lead to coherent text. 1 D i s c o u r s e c o h e r e n c e a n d a g g r e g a t i o n hi NLG, theories based on domain-independent rhetorical relations, in particular, Rhetorical Structure Theory (Mann and Thompson, 1987), are often used in text planning, whose task is to select the relevant information to be expressed and organise it into a hierarchical structure which captures certain discourse preferences such as preferences for the use of rhetorical relations. In the theory of discourse structure developed by Grosz and Sidner (1986), each discourse segment exhibits two types of coherence: local coherence among utterances inside the segment, and global coherence between this segment and other discourse segments. Discourse segments are connected by either a dominaTzce relation or a satisfaction-precedence relation. There has been an effort to synthesise tile two accounts of discourse structure. X loser and Moore (1996) argue that the two theories have considerable common ground, which lies in the correspondence between the notion of dominance and nuclearity. It is possible to map between Grosz and Sidner's linguistic s t ructure and RST text structure, and relation-based coherence and global coherence capture similar discourse properties. Oberlander et al. (1999) propose a distinction between two types of discourse coherence: proposition-based coherence, which exists between text spans connected by RST relations except for object-attribute elaboration, and entity-based coherence, which exists between spans of text in virtue of shared entities. entity-based coherence captures the coherence among adjacent propositions, which resembles local coherence in Grosz and Sidner's theory. To generate a coherent text, the text planning process must try to achieve both local (entitybased) and global (relation-based) coherence. Since the task of aggregation is to combine sinlple representations together to form a complex one, which in the mean time leads to a shorter text as a whole, aggregation could affect the ordering of text plans and the length of the whole text.. Therefore, it is closely related to tile task of maintaining both types of coherence. Here we treat embedding as a type of aggregation. There is no consensus as to where aggregation should happen or how it is related to other generation processes (Wilkinson, 1995; Reape and Mellish, 1999). In many NLG systems, aggregation is a post planning process whose preferences are only partially taken into account by