Multimodal Analyses enabling Artificial Agents in Human-Machine Interaction                                                          MA3


International Workshop on Multimodal Analyses for Human Machine Interaction


The deadline for submitting at the Special Issue of the Journal of Multimodal interfaces has been extended to the 7 of April (23.59 UTC -11).

The call for paper for the special issue of the Springer Journal on Multimodal User Interfaces :

This is an open call, we accept :

  1. -completely new papers

  2. - MA3 authors who wish to submit an extended version of their workshop paper need to add  at leas 70% novelty.

For any further information visit JMUI Special Issue


Thank you for your participation and for your papers making this workshop a fruitful scientific event. The proceedings will be published together with the joint RCVA workshop.

The workshop will be held on September, 15th at the 12th International Conference on Intelligent Virtual Agents, University of California, Santa Cruz (IVA 2012)

Exploring how human beings interact with the world and with each other is of outstanding importance for creating what is commonly named intelligent behaviour in Human-Machine Interaction. In the scientific community there is a growing interest in new techniques aimed at obtaining a more natural and human like interaction with machines. It is important, in the framework of intelligent virtual agent, to find the space for a socially intelligent virtual agent.

An appropriate system whether it is a virtual agent or a companion-like device has to analyse multiple sources of information to interact with humans in a proper and socially aware way. Such modalities are in general speech and its content, mimicry, gesture, biological signals, etc., and in particular it is the prosodic and paralinguistic features of the speech, facial mimicry, and heart rate, etc.

The aim of this workshop is to discuss techniques for gathering and analysing information on multimodal signals, combining methods of analysis in order to understand the correlation between different sources of information (gesture, speech, motion, etc.), as well as understanding how the environment and the context can influence how participants react.

Agents and smart devices are already more than a pure system. They become companions and a kind of friend. Thus, a Human- Machine Interaction will be no longer fixed on a task solution; it will be more human like, and include elements of chat and non- fixed gestures. The workshop will discuss the combination of these different signals, to foster more human-like interactions with virtual agents.

Workshop topics include, but are not limited to:

  1. Multimodal recognition and identification of user behaviour

  2. Awareness of the user's/agent's situation based on multimodal data

  3. Analysis of Human-Machine conversations

  4. Comparison of Human-Human vs. Human-Machine conversations

  5. Strategies of multimodal Human-Machine Interaction

  6. Analyses of speech and prosody with respect to paralinguistic features

  7. Mimicry and gesture analyses

  8. Using multimodal data sets for Human-Machine Interaction

  9. Annotation and processing of multimodal data sets

  10. Developments and combinations of analysis methods

  11. Applications of agents in Human-Machine Interaction.

University of California, Santa Cruz

September 15th, 2012

Workshop at IVA 2012 Twelfth

International Conference on Intelligent Virtual Agents

Organized by

One organizer is sponsored by