Accepted demos and posters
Contact Email: email@example.com
Affiliation: Saarland university
Title: Language Phenomena in Tutorial Dialogs on Mathematical Proofs
Summary: Dialogs in formal domains, such as mathematics, are characterized by a mixture of telegraphic natural language text and embedded formal expressions. Due to the lack of empirical data for such environments, we have collected a corpus of dialogs with a simulated tutoring system for teaching proofs in naive set theory. The analysis of this corpus enabled us to identify genre-specific variants of linguistic phenomena which impose specific requirements on natural language dialog management.
Author(s): Johan Bos and Tetsushi Oka
Contact Email: firstname.lastname@example.org
Affiliation: ICCS, School of Informatics
Title: Building Spoken Dialogue Systems for Believable Characters
Summary: In this paper we report on integrating a three-dimensional talking head with a spoken dialogue system. Using the information-state approach to dialogue management, we investigate the use of the open agent architecture for modelling some of the requirements of believable agents, focussing on the non-linguistic aspects of conversation.
Author(s): Erica Costantini, Susanne Burger, Fabio Pianesi
Contact Email: email@example.com
Affiliation: Language Technologies Institute, Carnegie Mellon University
Title: Communicative Effectiveness in Multimodal and Multilingual Dialogues
Summary: Multilingual communication enabled by a speech-to-speech translation system and supported by a multimodal interface may differ from 'ordinary' monolingual conversation for example a) in the way gestures are integrated in speech and b) in form and structure of the conversation. This paper describes a user study addressing these aspects. The preliminary results confirm that multilingual communication obtained through an actual translation system strongly affects the form and structure of the conversation as well as the level of coordination of gestures and speech. We argue that analysis methods addressing dialogue structure can help understanding and clarifying the adaptive communication strategies on different communication contexts.
Author(s): Donna K. Byron and Sarah Brown-Schmidt
Contact Email: firstname.lastname@example.org
Affiliation: Dept of Computer and Info. Science, The Ohio State University
Title: Subdialogs and Attentional State
Summary: This paper reports on a pilot study using the visual world paradigm to compare online NP interpretation against the claims made by grosz & sidner '86. Participants heard short instruction sequences containing an embedded discourse segment, after the close of which an ambiguous anaphoric NP was heard. One candidate referent for the NP appeared in the outer dialog segment before the subdialog, and one within it. Eye-movements during this critical NP were recorded. Only a small fraction of subjects looked at the referent mention in the embedded discourse segment. However, this indicates that focus stack frames in the discourse 'participants' mental equivalent of the attentional state do not get cleaned up at the conclusion of a subdialog in the same manner as proposed in the attentional state model.
Author(s): A.Cavalluzzi, V.Carofiglio, G.Cellamare and G.Grassano
Contact Email: email@example.com
Affiliation: Department of Computer Science - Bari -Italy
Title: A Tool to Simulate Affective Dialogs with an ECA
Summary: We demonstrate a testbed for simulating affective dialogs with an Embodied Conversational Agent (ECA). We will show how the dialog evolves according to the social context and the emotional state of the agent and how the agent's emotional state is, in its turn, influenced by the dialog course.
Author(s): Christoph Müller and Michael Strube
Contact Email: firstname.lastname@example.org
Affiliation: European Media Laboratory GmbH
Title: A Tool for Multi-Level Annotation of Language Data
Summary: We present a simple, light-weight Java tool for the annotation of linguistic data on multiple levels. The tool is based on the reduction of annotations to sets of markables having attributes and standing in certain relations. Due to its simplicity, the tool is fast, robust and highly usable.
Author(s): John Bateman, Kerstin Fischer, Reinhard Moratz, Scott Farrar & Thora Tenbrink
Contact Email: email@example.com
Affiliation: FB10 Faculty of Linguistics and Literary Sciences, University of Bremen
Title: Project I1-OntoSpace: Ontologies for Spatial Communication
Summary: The poster will describe the essential ingredients of project I1-OntoSpace (SFB/TR8). In addition to presenting the project outline and main goals there will be some emphasis on the empirical studies in spatial linguistic human-robot interaction being carried out, including first results.
Contact Email: firstname.lastname@example.org
Affiliation: CLT Sprachtechnologie GmbH, Saarbrücken
Title: A Framework for Information-State based Dialogue
Summary: At CLT Sprachtechnologie, we have developed a tool based on information state update (ISU). The tool allows the user to model dialogues using an information state, update rules and static plans. The execution of these models is based on an interaction of input processing and plan execution.
Author(s): Tiit Hennoste, Mare Koit, Andriela Rääbis, Krista Strandson, Maret Valdisoo, Evely Vutt
Contact Email: email@example.com
Affiliation: University of Tartu, Estonia
Title: Developing a Typology of Dialogue Acts: Tagging Estonian Dialogue Corpus
Summary: Estonian dialogue corpus includes recordings of spoken conversations, among them 114 information dialogues. We have worked out a dialogue acts typology departing from principles of conversation analysis. The basis of classification is the notion of adjacency pair. Some problems of tagging the corpus will be treated in the paper. Our further aim is to describe the structure of information dialogues and to model natural dialogue on the computer.
Author(s): Amy Isard, Carsten Brockmann, Jon Oberlander, Michael White
Contact Email: Amy.Isard@ed.ac.uk
Affiliation: University of Edinburgh
Title: The Critical Agent Dialogue (CrAg) Project
Summary: The aim of this project is to build and evaluate a simple, adaptable natural language generation system which can generate textual dialogue incorporating relatively subtle linguistic features which reflect dimensions of personality such as extraversion and neuroticism. The system will then be evaluated to investigate the impact on user impressions of altering personality parameters.
Author(s): Anthony Jameson
Contact Email: Anthony.Jameson@dfki.de
Affiliation: DFKI, Saarbruecken
Title: Animated User Representatives in Support of Asynchronous Group Decision Making: New Challenges for Dialog Processing
Summary: We first briefly describe a recently developed system for the support of asynchronous group decision making that features the use of animated characters as representatives of group members who are not currently interacting with the system. We then discuss several challenges that this new application context raises for the natural language generation of the animated characters: (a) how best to combine natural language generation with graphical interface techniques in a complementary way; (b) how to deal with the natural tendency of group members in such contexts to generate insincere arguments; and (c) how to introduce sufficiently natural and appealing variety in the generated utterances of the animated characters.
Author(s): Elena Karagjosova
Contact Email: firstname.lastname@example.org
Affiliation: Computerlinguistik, Universitaet des Saarlandes
Title: Marked Informationally Redundant Utterances in Tutorial Dialogue
Summary: The paper presents the preliminary results of a corpus investigation of Informationally Redundant Utterances (IRUs) in tutorial dialogues. The focus is on IRUs that are linguistically marked as such. The study suggests that marked IRUs have the function to make non salient propositions salient or to maintain the saliency of propositions in dialogue. It is suggested that marked IRUs may have a greater learning effect on students, since students are able to recognize them as old information they are supposed to know, which may help activate related knowledge and thus promote active learning.
Author(s): Udo Kruschwitz
Contact Email: email@example.com
Affiliation: Department of Computer Science, University of Essex
Title: Information Seeking Dialogues with Automatically Acquired Domain Knowledge
Summary: We will demonstrate a simple dialogue system that assists a user who searches collections of partially structured documents, such as Web pages or classified advertisements. The system is based on a generic framework - a specialized dialogue system - that has access to automatically acquired domain knowledge.
Author(s): Bilyana Martinovski, David Traum, Susan Robinson, Saurabh Garg
Contact Email: firstname.lastname@example.org
Affiliation: Institute for Creative Technologies, University of Southern California
Title: Functions and Patterns of Speaker and Addressee Identifications in Distributed Complex Organizational Tasks Over Radio
Summary: The purpose of this paper is to contribute to the understanding and predictability of identifications of speaker and addressee in radio mediated organization of work and to increase the understanding of the main factors influencing multi-party contact and complex task management. The contact management style tends to be most predictable with respect to the relationships between the speakers, i.e. intra vs. inter team and with respect to function in an episode (or subactivity).
Author(s): Mary McGee Wood, Richard Craggs, Ian Fletcher, and Peter Maguire
Contact Email: email@example.com
Affiliation: Department of Computer Science, University of Manchester
Title: Initiative in Health Care Dialogues
Summary: "Initiative", or "control over the flow of conversation", can be superficially mapped to the distinction between "open" and "closed" questions developed by psychiatrists in the analysis of human-human cancer care dialogues. However we show that the dialogue functions of individual utterances are often ambiguous until seen in the context of a longer sequence.
Author(s): Morge, M. and Collins, S.
Contact Email: firstname.lastname@example.org
Affiliation: ENS Mines St Etienne (France)
Title: Software components for a dialogue multiagent system
Summary: This work proposes a software component approach to design a dialogue multiagent system.
Contact Email: email@example.com
Affiliation: CLT Sprachtechnologie GmbH, Saarbrücken
Title: DiaMant: A Tool for Rapidly Developing Spoken Dialogue Systems
Summary: At CLT Sprachtechnologie, we have developed DiaMant (Dialogue Management Tool), a tool that allows the rapid development of dialogue systems based on extended finite state dialogue models. The development is done inside a graphical user interface that intuitively shows the dialogue model as an automaton in a graph representation. The dialog flow can be edited without the need to write a single line of code.
Author(s): Hoá NGUYEN, Jean CAELEN
Contact Email: firstname.lastname@example.org
Affiliation: Laboratory CLIPS-IMAG, University of Joseph Fourier, France
Title: Generic manager for spoken dialogue systems
Summary: This paper presents an approach to management of the human-machine spoken dialogue system. Our management model of human-machine spoken dialogue is based on three principal criteria: dialogue acts, strategy, and dialogue goal. With the goal to enhance independence of dialogue manager with the task model, our dialogue manager has many advantages and reduces the charge of building spoken dialogue systems.
Author(s): Dennis Hofs, Rieks op den Akker, Anton Nijholt & Hendri Hondorp
Contact Email: email@example.com
Affiliation: University of Twente
Title: A multimodal interaction system for navigation
Summary: We present a system for multimodal interaction using a generic architecture and dialogue model. Architecture and model are transparent and have been used for different task domains. In this demonstration the emphasis is on their use for the navigation task in a virtual environment. The dialogue model is based on the information state approach and the recognition of dialogue acts. Pairs of backward and forward looking tags and the preference rules of the dialogue act determiner together determine the structure of the dialogues that can be handled by the system. Special emphasis has been given to the system's action selection mechanism and the method of reference resolution.
Author(s): Massimo Poesio, Hannes Rieser
Contact Email: firstname.lastname@example.org
Affiliation: University of Essex
Title: Coordination in a PTT Approach to Dialogue
Summary: This paper is part of a long-term project aimed at comparing and evaluating existing models of dialogue on the same set of data. In this paper, we focus on examples in which a dialogue act is collectively developed via the interaction of the participants to a dialogue, and discuss how this collaboration is modeled in PTT, a theory of information state update developed by Poesio, Traum and Muskens.
Author(s): C.J. Rupp
Contact Email: email@example.com
Affiliation: Universität des Saarlandes
Title: Prototyping Dialogues for Information Access
Summary: We present a protoype dialogue system for information access that exemplifies a strategy for elliciting data and building infrastructure in an area that is underpopulated with fully functional systems. We make use of tools for prototyping information state dialogue models and linguistic resource already available in our project.
Author(s): David Schlangen, Alex Lascarides and Jason Baldridge
Contact Email: firstname.lastname@example.org
Affiliation: School of Informatics University of Edinburgh
Title: A Tool for Semi-Automatic and Interactive Annotation of Dialogue-Utterances with Information-States
Summary: We describe a tool for semi-automatic, interactive annotation of dialogues with discourse semantic information. The novel feature of this tool is its use of a reasoning engine which implements a semantic theory of discourse interpretation to suggest annotations to the user. The semantic theory explores the use of underspecified discourse structures.
Author(s): Thora Tenbrink
Contact Email: email@example.com
Affiliation: FB10 Faculty of Linguistics and Literary Sciences; University of Bremen
Title: Conveying spatial information in linguistic human-robot interaction
Summary: The aim of this paper is to present an overview of the range of variability employed by speakers with regard to the specificity of information in spatial communication, focussing on the significance of the context. I will describe this variability in terms of two underlying dimensions, one of which reflects a dichotomy of underdeterminacy and redundancy, while the other concerns vagueness and precision. I will use relevant results of several HRI experiments to exemplify my findings.
Author(s): Claus Zinn, Johanna D. Moore, Mark G. Core, Sebastian Varges, Kaska Porayska-Pomsta
Contact Email: firstname.lastname@example.org
Affiliation: School of Informatics, University of Edinburgh
Title: The BE&E Tutorial Learning Environment (BEETLE)
Summary: We describe the architecture and the components of BEETLE, a dialogue-enhanced Basic Electricity and Electronics Tutorial Learning Environment.