[Research | Projects | Publications | S&D | Education | Jobs | Staff]

From spoken dialogue to full natural interactive dialogue.
Theory, empirical analysis and evaluation.

29 May 2000 - immediately before LREC 2000, Athens, Greece

Call for Papers

Spoken dialogue systems have been in the marketplace since around 1990. Whereas the first systems only had single word recognition there has been a steady development towards increasingly natural spoken dialogue. The most advanced current systems still work within a limited task domain but some are capable of understanding and replying to fairly long user utterances, coping with various kinds of initiative, and taking a variety of contextual issues into account.

Naturalness, as perceived by the user, is closely connected to properties such as allowed user utterance length, grammar, vocabulary, style and initiative. Depending on task and situation, perceived naturalness is also connected to how the user can interact with the system. In human-human communication we normally do not restrict ourselves to using speech-only but also include gesture, facial expression, and bodily posture and we often draw on other information sources such as diagrams, maps and drawings.

Natural interactive dialogue and conversational systems are moving centre-stage because of increasing interest in adding other modalities to achieve a larger potential than speech alone can offer. Several recent conversational prototype systems include one or several natural interaction modalities in addition to speech. However, there are many open questions and unsolved or insufficiently explored problems related to extending spoken dialogue management and dialogue interfaces to enable increased natural interactivity.

The workshop aims to bring together researchers and developers in the area of natural interactive dialogue. The goal of the workshop is to highlight and evaluate empirically based theories and methods for natural multimodal conversational dialogue management and dialogue interfaces, and their evaluation. Focus will be on key issues such as dialogue initiative, reference, communicative acts, feedback, and cooperativity.

Topics

Topics of interest should fall within theory, empirical analysis, and evaluation of key issues in the transition from spoken dialogue to full natural interactive dialogue. Topics include but are not limited to:

• Cooperative natural interactive dialogue interfaces and usability

• Experience from natural interactive systems development and evaluation

• Integration of natural interactivity modalities

• Empirically based theories in support of natural interactive dialogue management and interfaces

• Communicative acts (beyond speech acts)

• Reference in a multimodal context

• Dialogue initiative

• Task management

• Feedback

• Methods for evaluation of natural interactive dialogue management and dialogue interfaces

Submission details

Extended abstracts should be around 4 pages in length. Final papers should not exceed six pages. Extended abstracts must be submitted electronically to laila@nis.sdu.dk and must be in postscript or rtf format. Please write "LREC2000 workshop paper submission" in the subject line.

Final paper style format is available at http://www.icp.grenet.fr/ELRA/lrec/kauthor.html and here.

Important dates

Deadline for submission of extended abstracts: 10 March 2000

Notification of acceptance/rejection: 27 March 2000

Deadline for submission of accepted papers: 10 April 2000

Workshop: 29 May 2000

Program Committee

• Laila Dybkjær, Natural Interactive Systems Laboratory, Odense, Denmark (chair)

• Niels Ole Bernsen, Natural Interactive Systems Laboratory, Odense, Denmark

• Justine Cassell, MIT Media Lab, USA

• Ronald Cole, Center for Spoken Language Understanding, University of Colorado at Boulder, USA

• Björn Granström, Dept. of Speech, Music and Hearing, KTH, Sweden

• Joseph Mariani, LIMSI-CNRS, France

• Dominic W. Massaro, Dept. of Psychology, University of California, USA

• David McNeill, Dept. of Psychology, University of Chicago, USA

• Sharon Oviatt, Center for Human-Computer Communication, Oregon Graduate Institute of Science & Technology, USA

• Oliviero Stock, IRST, Italy

• Jan van Kuppevelt, IMS, Universität Stuttgart, Germany