Workshop on Mixed-Initiative Planning And Scheduling
To be held in conjunction with The 15th International Conference on
Automated Planning and Scheduling (ICAPS), Monterey, California, June,
New deadline for papers: March 9
The goals of mixed initiative planning and scheduling (MIPAS) systems
are to combine the strengths of humans and computers so as to produce
joint systems that exhibit better performance than either alone. A
wide variety of domains, including manufacturing, military logistics,
space, and transportation are finding both manual and automated
approaches to planning and scheduling inadequate. A manual approach,
involving large teams of highly trained experts, does not scale well
to increasingly complex scenarios, and remains expensive, time
consuming and error-prone. Automated planning and scheduling systems,
though long a focus of AI research, are difficult to integrate into
human-centric activities, for both technical and psychological
reasons. Thus there is increasing interest in mixed-initiative
planning and scheduling systems, where humans can apply their
expertise, intuition, and exercise an appropriate level of control,
while computing resources can be leveraged to assist, learn, verify,
generate options, and control resources, using ever more sophisticated
algorithms and more powerful hardware. However, few successful MIPAS
systems have been fielded to date. Challenges include identifying the
parts of a complex task where MIPAS systems might have the greatest
added value; developing planning and scheduling algorithms that can
simultaneously meet humans' requirements for speed, feasibility
and quality; presenting and communicating concepts effectively between
human and computer, gaining user acceptance, and assessing impact on
overall planning and scheduling performance.
This workshop will address the challenges of mixed-initiative planning
and scheduling, discussing past work, current needs and future
directions. Some of the topics that we plan to cover include:
- Methodologies and strategies for MIPAS system development.
- Human-computer Interaction (HCI) issues including communication of
users' constraints, presentation of computer-generated options,
and both collecting and providing rationale.
- Evaluation methodologies.
- Probabilistic and Machine Learning support for adaptive MIPAS.
- Uses of case-based reasoning in MIPAS.
- Intent recognition and intent reasoning for MIPAS.
- Models for controlling and sharing initiative in MIPAS
- Decision biasing
and over-reliance on automated systems.
- Knowledge representation for MIPAS.
- Experience with development, deployment, and evolution of MIPAS systems.
- Formal models of MIPAS components and approaches.
There are two types of submissions:
Attendance at the workshop will be by invitation, and invitations will
be based on submissions.
- A paper to be presented at the workshop. The submission should be
a maximum of 8 pages. The paper presentation will be short, with
emphasis on discussion.
- A one page statement of interest summarizing your degree of
interest in and familiarity with the topics outlined above.
The workshop will be one full day, with a mix of paper presentation,
expert panels, invited talks, and open discussion. Open and vigorous
discussion will be actively promoted, with correspondingly less time
spent on prepared presentations.
Paper or Statement of interest submission deadline:
March 9, 2005.
Notification date: March 23, 2005.
Camera ready copy due: April 18, 2005.
The proceedings will be published as a booklet for the conference
itself, and the program committee is pursuing wider publication
|John Bresina (NASA ARC)
||bresina at email.arc.nasa.gov
|Mark Burstein (BBN)
||burstein at bbn.com
|George Ferguson (co-chair) (U. Rochester)
||ferguson at cs.rochester.edu
|Caroline Hayes (co-chair) (UMN)
||hayes at me.umn.edu
|Karen Myers (SRI)
||myers at ai.sri.com
|Dana Nau (U. MD)
||nau at cs.umd.edu
|Debra Schreckenghost (NASA JSC)
||ghost at ieee.org
|Stephen Smith (CMU)
||sfs at cs.cmu.edu
|Greg Sullivan (co-chair) (BAE Systems)
||gregory.sullivan at baesystems.com