Kees van Deemter

Computational generation of referring expressions: A survey

This article offers a survey of computational research on referring expression generation (REG). It introduces the REG problem and describes early work in this area, discussing what basic assumptions lie behind it, and showing how its remit has widened in recent years. We discuss computational frameworks underlying REG, and demonstrate a recent trend that seeks to link REG algorithms with well-established Knowledge Representation techniques. Considerable attention is given to recent efforts at evaluating REG algorithms and the lessons that they allow us to learn. The article concludes with a discussion of the way forward in REG, focusing on references in larger and more realistic settings.

On coreferring: coreference in MUC and related annotation schemes

In this paper, it is argued that coreference annotations, as performed in the MUC community for example, go well beyond annotation of the relation of coreference proper. As a result, it is not always clear what semantic relation these annotations are encoding. The paper discusses a number of problems with these annotations and concludes that rethinking of the coreference task is needed before the task is expanded. In particular, it suggests a division of labor whereby annotation of the coreference relation proper is separated from other tasks such as annotation of bound anaphora and of the relation between a subject and a predicative NP.

Generating referring expressions: boolean extensions of the incremental algorithm

This paper brings a logical perspective to the generation of referring expressions, addressing the incompleteness of existing algorithms in this area. After studying references to individual objects, we discuss references to sets, including Boolean descriptions that make use of negated and disjoined properties. To guarantee that a distinguishing description is generated whenever such descriptions exist, the paper proposes generalizations and extensions of the Incremental Algorithm of Dale and Reiter (1995).

Real versus Template-Based Natural Language Generation: A False Opposition?

This article challenges the received wisdom that template-based approaches to the generation of language are necessarily inferior to other approaches as regards their maintainability, linguistic well-foundedness, and quality of output. Some recent NLG systems that call themselves template-based will illustrate our claims.

Generating Referring Expressions that Involve Gradable Properties

This article examines the role of gradable properties in referring expressions from the perspective of natural language generation. First, we propose a simple semantic analysis of vague descriptions (i.e., referring expressions that contain gradable adjectives) that reflects the context-dependent meaning of the adjectives in them. Second, we show how this type of analysis can inform algorithms for the generation of vague descriptions from numerical data. Third, we ask when such descriptions should be used. The article concludes with a discussion of salience and pointing, which are analyzed as if they were gradable adjectives.

Building a Semantically Transparent Corpus for the Generation of Referring Expressions.

This paper discusses the construction of a corpus for the evaluation of algorithms that generate referring expressions. It is argued that such an evaluation task requires a semantically transparent corpus, and controlled experiments are the best way to create such a resource. We address a number of issues that have arisen in an ongoing evaluation study, among which is the problem of judging the output of GRE algorithms against a human gold standard.

Generating Referring Expressions: Making Referents Easy to Identify

It is often desirable that referring expressions be chosen in such a way that their referents are easy to identify. This article focuses on referring expressions in hierarchically structured domains, exploring the hypothesis that referring expressions can be improved by including logically redundant information in them if this leads to a significant reduction in the amount of search that is needed to identify the referent. Generation algorithms are presented that implement this idea by including logically redundant information into the generated expression, in certain well-circumscribed situations. To test our hypotheses, and to assess the performance of our algorithms, two controlled experiments with human subjects were conducted. The first experiment confirms that human judges have a preference for logically redundant expressions in the cases where our model predicts this to be the case. The second experiment suggests that readers benefit from the kind of logical redundancy that our algorithms produce, as measured in terms of the effort needed to identify the referent of the expression.

Towards Competency Question-Driven Ontology Authoring

Ontology authoring is a non-trivial task for authors who are not proficient in logic. It is difficult to either specify the requirements for an ontology, or test their satisfaction. In this paper, we propose a novel approach to address this problem by leveraging the ideas of competency questions and test-before software development. We first analyse real-world competency questions collected from two different domains. Analysis shows that many of them can be categorised into patterns that differ along a set of features. Then we employ the linguistic notion of presupposition to describe the ontology requirements implied by competency questions, and show that these requirements can be tested automatically.

Utility and Language Generation: The Case of Vagueness

This paper asks why information should ever be expressed vaguely, re-assessing some previously proposed answers to this question and suggesting some new ones. Particular attention is paid to the benefits that vague expressions can have in situations where agreement over the meaning of an expression cannot be taken for granted. A distinction between two different versions of the above-mentioned question is advocated. The first asks why human languages contain vague expressions, the second question asks when and why a speaker should choose a vague expression when communicating with a hearer. While the former question is purely theoretical, the latter has practical implications for the computational generation of utterances in Natural Language Generation (NLG).

Graphs and booleans : On the generation of referring expressions

Generation of Referring Expressions (gre) is a key task of Natural Language Generation nlg systems (e.g., Reiter and Dale, 2000, section 5.4). The task of a gre algorithm is to find combinations of properties that allow the generator to refer uniquely to an object or set of objects, called the target of the algorithm. Older gre algorithms tend to be based on a number of strongly simplifying assumptions. For example, they assume that the target is always one object (rather than a set), and they assume that properties can always only be conjoined, never negated or disjoined. Thus, for example, they could refer to a target object as “the small violinist”, but not as “the musicians not holding an instrument”. As a result of such simplifications, many current gre algorithms are logically incomplete. That is, they sometimes fail to find an appropriate description where one does exists.1 To remedy such limitations, various new algorithms have been proposed in recent years, each of which removes one or more simplifying assumptions. They extend existing gre algorithms by allowing targets that are sets (Stone, 2000; van Deemter, 2000), gradable properties (van Deemter, 2000, 2006), salience (Krahmer and Theune, 2002), relations between objects (Dale & Haddock, 1991; Horacek, 1997), and Boolean properties (van Deemter, 2001, 2002). Recently a new formalism, based on labelled directed graphs, was proposed as a vehicle for expressing and implementing different gre algorithms (Krahmer et al., 2001, 2003). Although the formalism was primarily argued to support relatively simple descriptions (not involving negations or disjunctions, for example), we will show that it can be used beyond these confines. Far from claiming that this will solve all the problems in this area, we do believe that a common formalism