JMUI - Journal on Multimodal User Interfaces Special Issue


Important Dates

Submission Deadline: March 7 2013 ---- Extended to April 7, 2013 (23:59 UTC/GMT -11)

Notification of acceptance: June 7, 2013

Camera-ready deadline: September 7, 2013 (23.59 UTC?GMT-11)

Publication November 2013

                    From Multimodal Analysis to Real-Time Interactions with Virtual Agents

Extended Deadline for paper submission: April 7, 2013

Call for papers:

One of the aims in building multimodal user interfaces is to make the interaction between user and system as natural as possible. Possibly the most natural form of interaction we know is the way we interact with other humans. By building any kind of virtual agents, we aim to recreate this natural form of interaction in human machine interaction.

Development of virtual agent systems that are capable of recreating natural interactions with humans typically involves several steps: record and analyze natural interaction data, extract and recognize relevant multimodal features, craft or learn models from these features, generate the appropriate behavior in real-time based on these models and evaluate the system in a methodologically sound experiment. Each of these steps presents challenges.

The collection of multimodal natural corpora brings the challenges of collecting naturalistic behavior, enhancing the corpus with reliable ground truth annotations and dealing with the inter-personal differences in behavior.

Through multimodal analysis of the corpus, features that need to be extracted in real-time from the recording signals need to be identified. Such features include, but are not limited to, speech and its content, eye gaze, prosodic and paralinguistic features, facial expressions, body movements, or more advanced interpretations of such features such as the affective state, personality, mood or intentions of the user.

By combining these multimodal features and interpretations, the agent needs to reason about, plan and realize the actions it needs to perform in response with the correct timing. These responses include content and paralinguistic features of the speech, appropriate accompanying gestures and facial expressions. Continuous perception, interpretation, reasoning and generation is required to keep the interaction between the user and the virtual agent as natural and fluent as human-human interaction.

The development and evaluation of such complex real-time continuous virtual agent systems also requires new research methodologies in terms of experiment design. Many aspects can influence the evaluation of the system as a whole and individual aspects of the system may require continuous evaluation as well.

For this special issue, we welcome technical, empirical and overview papers that contribute towards the development, evaluation and understanding of real-time interactions with virtual agents.

Topics include:

  1. Multimodal analysis and annotation of human-human corpora

  2. Real-time perception and understanding of speech, facial expression and gestures

  3. Estimation of user affect, mood and personality

  4. Models for multimodal generation of speech, facial expressions and body movement

  5. Real-time dialog management

  6. Real-time interactions with humanoid robots

  7. Applications of real-time interactions between humans and virtual agents

  8. Tools, systems and evaluation methods for above

  9. Comparison of human-human and human-agent interactions

Authors Instructions:

Submissions should be 8-12 pages and must not have been previously published, with the exception that substantial extensions of conference and workshop papers (at least 30% new content) can be considered. The authors will be required to follow the Author’s Guide for manuscriptsubmissiontotheJournalofMultimodalUserInterfaces ( published by Springer.

During the submission process, please select “REAL-TIME special issue” as article type.

Guest editors:

Ronald Böck, Otto von Guericke University Magdeburg, Germany (

Francesca Bonin, Trinity College Dublin, Ireland (

Nick Campbell, Trinity College Dublin, Ireland

Jens Edlund, KTH Stockholm, Sweden

Iwan de Kok, University of Twente, The Netherlands (

Ronald Poppe, University of Twente, The Netherlands (

David Traum, USC Institute for Creative Technologies, USA