We are pleased to introduce the technical programme of the Fourth International Natural Language Generation Conference (INLG), the Biennial Meeting of SIGGEN, the ACL Special Interest Group in Natural Language Generation. INLG is the leading international conference on research into natural language generation. It has been held at Brockenhurst (UK) in 2004, in Harriman (New York, USA) in 2002, and in Mitzpe Ramon (Israel) in 2000. Prior to 2000, the INLG meetings were International Workshops, running every other year since 1980. The INLG conference provides a forum for the discussion, dissemination and archiving of research topics and results in the field of text generation. This year, INLG is being held as a COLING/ACL 2006 workshop. It takes place on the weekend prior to the main COLING/ACL 2006 conference, on July 15-16th 2006, in Sydney, Australia.
The INLG programme consists of substantial, original, and previously unpublished results on all topics related to natural language generation. This year, as in previous years, each submission was reviewed as a full paper by at least three members of an international programme committee of leading researchers in the field, listed on the next page. We received 38 submissions (both long and short papers) from all over the world, from which we accepted 11 long papers (including two student papers) and five short papers. We would like to thank all who submitted papers and our programme committee for their hard work.
This year, the programme centers around a variety of research issues around the realisation component of a natural language system, including the use of statistical techniques. In particular, there are a substantial number of contributions on the generation of referring expressions. The programme also includes an invited talk by Professor Kathleen McKeown (Columbia University, New York, USA) entitled "Lessons Learned from Large Scale Evaluation from Systems that Produce Text: Nightmares and Pleasant Surprises" and a special session on "Sharing Data and Comparative Evaluation", organised by Dr Anja Belz (University of Brighton, UK) and Professor Robert Dale (Macquarie University, Australia).
Proceeding Downloads
Lessons learned from large scale evaluation of systems that produce text: nightmares and pleasant surprises
As the language generation community explores the possibility of an evaluation program for language generation, it behooves us to examine our experience in evaluation of other systems that produce text as output. Large scale evaluation of summarization ...
A generation-oriented workbench for performance grammar: capturing linear order variability in German and Dutch
We describe a generation-oriented workbench for the Performance Grammar (PG) formalism, highlighting the treatment of certain word order and movement constraints in Dutch and German. PG enables a simple and uniform treatment of a heterogeneous ...
CCG chart realization from disjunctive inputs
This paper presents a novel algorithm for efficiently generating paraphrases from disjunctive logical forms. The algorithm is couched in the framework of Combinatory Categorial Grammar (CCG) and has been implemented as an extension to the OpenCCG ...
Overgeneration and ranking for spoken dialogue systems
We describe an implemented generator for a spoken dialogue system that follows the 'overgeneration and ranking' approach. We find that overgeneration based on bottom-up chart generation is well-suited to a) model phenomena such as alignment and ...
Individuality and alignment in generated dialogues
It would be useful to enable dialogue agents to project, through linguistic means, their individuality or personality. Equally, each member of a pair of agents ought to adjust its language (to a greater or lesser extent) to match that of its ...
Using distributional similarity to identify individual verb choice
Human text is characterised by the individual lexical choices of a specific author. Significant variations exist between authors. In contrast, natural language generation systems normally produce uniform texts. In this paper we apply distributional ...
Adjective-to-verb paraphrasing in Japanese based on lexical constraints of verbs
This paper describes adjective-to-verb paraphrasing in Japanese. In this paraphrasing, generated verbs require additional suffixes according to their difference in meaning. To determine proper suffixes for a given adjective-verb pair, we have examined ...
Generating references to parts of recursively structured objects
Algorithms that generate expressions to identify a referent are mostly tailored towards objects which are in some sense conceived as holistic entities, describing them in terms of their properties and relations to other objects. This approach may prove ...
Overspecified reference in hierarchical domains: measuring the benefits for readers
It is often desirable that referring expressions be chosen in such a way that their referents are easy to identify. In this paper, we investigate to what extent identification becomes easier by the addition of logically redundant properties. We focus on ...
Algorithms for generating referring expressions: do they do what people do?
The natural language generation literature provides many algorithms for the generation of referring expressions. In this paper, we explore the question of whether these algorithms actually produce the kinds of expressions that people produce. We compare ...
Group-based generation of referring expressions
Past work of generating referring expressions mainly utilized attributes of objects and binary relations between objects in order to distinguish the target object from others. However, such an approach does not work well when there is no distinctive ...
Noun phrase generation for situated dialogs
We report on a study examining the generation of noun phrases within a spoken dialog agent for a navigation domain. The task is to provide real-time instructions that direct the user to move between a series of destinations within a large interior ...
The clarity-brevity trade-off in generating referring expressions
Existing algorithms for the Generation of Referring Expressions (GRE) aim at generating descriptions that allow a hearer to identify its intended referent uniquely; the length of the expression is also considered, usually as a secondary issue. We ...
Generic querying of relational databases using natural language generation techniques
This paper presents a method of querying databases by means of a natural language-like interface which offers the advantage of minimal configuration necessary for porting the system. The method allows us to first automatically infer the set of possible ...
Generating intelligent numerical answers in a question-answering system
In this paper, we present a question-answering system on the Web which aims at generating intelligent answers to numerical questions. These answers are generated in a cooperative way: besides a direct answer, comments are generated to explain to the ...
Generating multiple-choice test items from medical text: a pilot study
We report the results of a pilot study on generating Multiple-Choice Test Items from medical text and discuss the main tasks involved in this process and how our system was evaluated by domain experts.
Generation of biomedical arguments for lay readers
This paper presents the design of a discourse generator that plans the content and organization of lay-oriented genetic counseling documents containing arguments, and an experiment to evaluate the arguments. Due to the separation of domain, argument, ...
Introduction to the INLG'06 special session on sharing data and comparative evaluation
The idea for this special session had its origins in discussions with many different members of the NLG community at the 2005 Workshop on Using Corpora for Natural Language Generation (UCNLG'05, held in conjunction with the Corpus Linguistics 2005 ...
Evaluations of NLG systems: common corpus and tasks or common dimensions and metrics?
In this position paper, we argue that a common task and corpus are not the only ways to evaluate Natural Language Generation (NLG) systems. It might be, in fact, too narrow a view on evaluation and thus not be the best way to evaluate these systems. The ...
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 ...
Shared-task evaluations in HLT: lessons for NLG
While natural language generation (NLG) has a strong evaluation tradition, in particular in userbased and task-oriented evaluation, it has never evaluated different approaches and techniques by comparing their performance on the same tasks (shared-task ...
GENEVAL: a proposal for shared-task evaluation in NLG
We propose to organise a series of sharedtask NLG events, where participants are asked to build systems with similar input/output functionalities, and these systems are evaluated with a range of different evaluation techniques. The main purpose of these ...