Call For Papers
------------------------------------------------------------------------ The AAAI-04 Workshop on Supervisory Control of Learning and Adaptive Systems
To be held at the Nineteenth National Conference on Artificial Intelligence (AAAI-04) July 25 or July 26, 2004 in San Jose, CA
------------------------------------------------------------------------
With supervisory control, a human operator intermittently takes control of a process that is otherwise controlled by a computer. Supervisory control involves both autonomy and intelligence, although the latter is normally attributed solely to the human operator. One goal of this workshop, therefore, is to bring together researchers in Robotics, Machine Learning, Human-Computer Interaction and other areas to explore the role of supervisory control for AI systems, especially for systems where both human and machine share the ability to learn and adapt to changing circumstances.
In the past, supervisory control has focused primarily on traditional applications of telerobotics, such as hazardous waste disposal, planetary and undersea exploration, and remote surveillance and repair. These applications remain important areas of research today, although in recent years supervisory control has become much more pervasive than we often realize. Assistive technology for the physically handicapped, software agents for electronic commerce, and education technology for the modern classroom are all examples where shared control by human and machine will have increasing societal impact. Thus, another goal of this workshop is to identify real-world applications where the combination of supervisory control and AI will have the most impact. Specific questions for discussion include the following:
- How should the user interface convey to the human operator the capabilities and learning progress of the AI system?
- What information is required for the human operator to develop a trust in the AI system?
- Is there a principled way for an AI system to infer the intentions of, and take control from its human supervisor?
- What heuristics or user models can be used to speed up what is essentially a transfer of knowledge from human to machine?
- Which learning algorithms are suitable for short time scales where data are abundant and for long time scales where data are relatively sparse?
- How should the learning system go about improving its capabilities beyond those of the human supervisor?
- Under what conditions can another AI system act as supervisor and for what conditions should a human always remain in the loop?
TOPICS
Areas of interest for this workshop include, but are not limited to the following:
- Adjustable autonomy/Mixed initiative control
- Co-learning
- Assistive technology
- Adaptive user interfaces
- Telerobotics
- Human-robot interaction
- Acquisition of transparent user models
- Fault-tolerant learning
- Learning from demonstration
- Evaluation methods
FORMAT
This one-day workshop will consist of a keynote talk of general interest, followed by several invited talks and paper presentations on more focused topics. Panels will be used when appropriate to facilitate discussion of clusters of closely related talks. Depending on the submissions, the workshop may include a poster session to encourage more in-depth discussions.
SUBMISSIONS
Participants are required to submit either a technical paper (roughly six pages in the conference format) or else an abstract (up to two pages) describing research relevant to the workshop. Submissions should be sent via email to one of the co-chairs and should be in Postscript, PDF, or MS Word format. If the papers are of sufficient quantity and quality, we will seek to publish them as an edited book or journal special issue.
CHAIRS
Mike Rosenstein Mohammad Ghavamzadeh Department of Computer Science Department of Computer Science Univ. of Massachusetts Amherst Univ. of Massachusetts Amherst
140 Governors Drive 140 Governors Drive Amherst, MA 01003 Amherst, MA 01003 Voice: 413-545-1876 Voice: 413-545-1596 Fax: 413-545-1249 Fax: 413-545-1249 Email: snipped-for-privacy@cs.umass.edu Email: snipped-for-privacy@cs.umass.edu WWW:COMMITTEE
Chuck Anderson, Colorado State University ( snipped-for-privacy@cs.colostate.edu) Mathias Bauer, German Research Center for AI ( snipped-for-privacy@dfki.de) Rob St. Amant, North Carolina State University ( snipped-for-privacy@csc.ncsu.edu) Holly Yanco, University of Massachusetts Lowell ( snipped-for-privacy@cs.uml.edu)
IMPORTANT DATES
- Technical submissions and abstracts due: March 12, 2004
- Notification of acceptance: April 16, 2004
- Camera ready copies due: May 25, 2004
- Workshop date: TBD, either July 25 or July 26, 2004
WEBSITE: