ICML/UAI/COLT 2008 workshop: Context-Sensing and Inference for Ubiquitous Interaction
Traditionally wireless sensor networks have been used for monitoring different physical and environmental conditions. Today there are a whole new class of ubiquitous, possibly wearable sensors available, that can be used for monitoring and inferring not only physical processes but also social and other types of interaction. Also the coverage of these sensor networks is much wider. Combined these types of sensor networks create an extremely rich source of data. How should advanced statistical methods and machine learning techniques be used with this type of data?
Learning and inference from sensor data, when the types of sensors are not limited in any way, poses a challenging and almost unlimited terrain for novel applications. How can inferred and derived knowledge and context information be utilized in different application scenarios? How does consumer-oriented applications benefit from this? How about industrial applications? What are the methods one can/should use for learning and inference and how does one apply them?
The workshop encourages submissions both on methodological advances in statistical analysis of sensor data, and on novel applications enabled by machine learning techniques utilizing this type of data.
Topics of interest include (but are not limited to):
- Semantically Meaningful abstractions from data (e.g., place identification, social analysis, activity recognition/detection etc.)
- Distributed machine learning
- Novel methods for handling large volumes of (potentially meaningless) data
- Data segmentation
- Ubiquitous user models that merge sensed data with personalization
- Statistical methods for providing feedback on uncertain belief states to the user
- Mobile spatial interaction & sensor fusion
- Parametrized pseudophysical models for inferred belief states
- Use of machine learning for instrumented usability research (ie. not just online for end users, but to better understand interaction behavior)
- Use of mobile agent- and sensor networks for simulation and analyzing viral and social phenomena
- Paper submission deadline: April 25th
- Acceptance notification: May 16th
The manuscripts should be prepared following the ICML paper submission guidelines. The maximum length of submissions is 6 pages for research and position papers and 2 pages for videos and demonstrations.
Submissions should be sent using the EasyChair submission system.
All participants in the workshop need to register for the event. Further information will be available later. At least one author of the submitted paper is expected to participate and present the paper at the workshop.
- Péter Pál Boda, Nokia Research Center, Palo Alto, USA
- Wray Buntine, NICTA, Canberra, Australia
- Patrik Floréen, HIIT, Helsinki, Finland
- Mark Hansen, UCLA, Los Angeles, USA
- Jari Kangas, Nokia Research Center, Tampere, Finland
- Roderick Murray-Smith, Glasgow University, UK
- Petteri Nurmi, HIIT, Helsinki, Finland
- Jukka Perkiö, University of Helsinki, Finland
- Ákos Vétek, Nokia Research Center, Helsinki, Finland
The workshop takes place on Wednesday 9 July 2008.
The program consists of 1-2 invited talks together with presentations, and possibly videos and demonstrations, from the participants. Each paper presentation is allocated a time of 20-30 minutes (including Q/A time) followed by time for discussions at the end of each session. The invited speakers will be announced at a later stage.
If you have questions, please contact csiui08 (at) easychair (dot) org
Last updated on 28 Oct 2008 by WWW administrator - Page created on 3 Mar 2008 by Jukka Perkiö