An algorithm for generating referential descriptions with flexible interfaces

Abstract

Most algorithms dedicated to the generation of referential descriptions widely suffer from a fundamental problem: they make too strong assumptions about adjacent processing components, resulting in a limited coordination with their perceptive and linguistics data, that is, the provider for object descriptors and the lexical expression by which the chosen descriptors is ultimately realized. Motivated by this deficit, we present a new algorithm that (1) allows for a widely unconstrained, incremental, and goal-driven selection of descriptors, (2) integrates linguistic constraints to ensure the expressibility of the chosen descriptors, and (3) provides means to control the appearance of the created referring expression. Hence, the main achievement of our approach lies in providing a core algorithm that makes few assumptions about other processing components and improves the flow of control between modules.

Alignment in Interactive Reference Production: Content Planning, Modifier Ordering, and Referential Overspecification

“…Linguistic realization of referring expressions has traditionally received less attention than content planning, and often it is assumed that this task can be performed by a separate realization algorithm, in a second phase (but see, e.g., Horacek, 1997; Krahmer & Theune, 2002; Stone & Webber, 1998, who argue for a tighter coupling between the two). An important complication for realizing referring expressions is that selected attribute–value pairs will often be realized as linguistic, prenominal modifiers, and ordering these may be difficult.…”

A meta-algorithm for the generation of referring expressions

“…0 Various important ingredients of other generation algorithms can be captured in the algorithm proposed here as well. For instance, Horacek (1997) points out that an algorithm should not collect a set of properties which cannot be realized given the constraints of the grammar. This problem can be solved, following Horacek's suggestion, by slightly modifying the algorithm in such a way that for each potential edge it is immediately investigated whether it can expressed by the realizer.…”

“…The generation of referring expressions is one of the most common tasks in natural language generation, and has been addressed by many researchers in the past two decades (including Appelt 1985, Dale 1992, Reiter 1990, Dale & Haddock 1991, Dale & Reiter 1995, Horacek 1997, Stone & Webber 1998, Krahmer & Theune 1999and van Deemter 2000. As a result, there are many different algorithms for the generation of referring expressions, each with its own objectives: some aim at producing the shortest possible description, others focus on efficiency or realistic output.…”

Computational Perspectives on Discourse and Dialog

“…Computational research attempts to specify not just what these additional candidates may be, but the specific search processes by which they will be found and the intended referent correctly identified, as in Bos et al (1995); Hahn et al (1996); Hobbs et al (1993); Markert and Hahn (1997). From the perspective of text generation, choosing whether to use a definite NP (and, if so, choosing one sufficient to refer uniquely to the intended referent) involves both search and inference for other entities in the context that block referential uniqueness, and search for properties that distinguish the intended referent from the remaining others, as in Dale and Haddock (1991); Dale and Reiter (1995); Horacek (1997); Stone and Doran (1997); Stone and Webber (1998).…”