Corpus Analysis of Natural Language Dialogue for Building ITSs

A Workshop of AIED 2011

The 15th International Conference on Artificial Intelligence in Education

(canceled)

Contents

• Important dates
• Call for papers
• Submissions
• Program committee
• For further information

Important dates

Call for papers

Computer engagement with natural language dialogue is an important aspect of intelligent tutoring technology and computer supported collaborative learning systems. In addition to being an innate capability of the human organism, natural language is flexible and extensible and does not require the user to learn a new GUI.

To form a scientific basis for machine processing of dialogue - either engaging in dialogue or processing dialogue data - many research groups have collected and analyzed corpora of human-human dialogues. Corpus analysis for system construction differs from corpus analysis for other purposes, including various types of linguistic, sociolinguistic or pedagogical analysis, for two reasons. First, technology researchers must focus on methods that can be implemented via computer in real time. Second, intelligent tutoring domains are often technical domains of discourse where a high degree of precision is required.

All researchers involved in natural language corpus analysis for pedagogical systems, whether for a current or a prospective system, are invited to participate. Linguistic, statistical, machine learning and hybrid approaches are all welcome.

The following topics are of special interest, but other topics are welcome.

• Input understanding: Parsing, chunking, information extraction, LSA, hybrid approaches
• Discourse analysis: Topic choice
• Dialogue management: Turn-taking
• Text generation: Corpus analysis for a wide spectrum of generation algorithms
• Spoken language processing: Speech understanding and generation
• Relation of language and affect: Recognition of affect in student or tutor language
• Correlation of natural language and other media, e.g., video
• Adaptation of generated text based on content, user model, or other factors
• Automatic authoring: Generation of ITS dialogues from attested examples
• Corpus collection: Issues involved in obtaining a representative sample of the desired text
• Text analysis for CSCL: Characterizing and adapting to individual student contributions

Submission information

• Short papers may range from 2-4 pages in length and include position papers, reports of work in progress, or a description of a proposed demo.

• Long papers may be up to 8 pages long and include reports of completed work or analyses of issues in the field.

All accepted papers will appear in the workshop proceedings.

Papers must be submitted through EasyChair by midnight (GMT) on March 29. Papers should be submitted in PDF format using the Springer LNCS format. Templates can be found in the conference author kit. To submit, go to the EasyChair workshop page.

Program Committee

• Reva Freedman (Chair), Northern Illinois University
• Charles Callaway, University of Haifa
• Barbara Di Eugenio, University of Illinois at Chicago
• Michael Glass, Valparaiso University
• Pamela W. Jordan, University of Pittsburgh
• Alistair Knott, University of Otago
• Bruce McLaren, Carnegie-Mellon University
• Andrew Olney, University of Memphis
• David R. Traum, USC Institute for Creative Technologies

For further information

Please contact Reva Freedman at rfreedman@niu.edu.