Automated planning is of central concern in high-level symbolic AI research, with applications in logistics, robotics and service composition. In the simple case of an agent operating in a known world, the output of a planner is just a sequence of actions to be performed to the effect that it achieves a desired goal state. Epistemic planning is the enrichment of automated planning with epistemic notions, including knowledge and beliefs, which not only refer to incomplete knowledge, but also beliefs about this knowledge. Epistemic planning has promising application potentials in all types of domains requiring artificial agents that have skills both in planning and in reasoning about knowledge and beliefs (of themselves and others). Such applications include domestic robots interacting with humans, non-player characters in video games, and autonomous robots interacting in a factory setting. It is a relatively recent area of research, and is inherently multi-disciplinary involving research from automated planning, epistemic logic, and knowledge representation & reasoning. In order to achieve formalisms and systems for epistemic planning that are both expressive and practically efficient, it is necessary to combine the state of the art from all three areas.
Topics relevant to the workshop include:
- Theory of mind, recursive reasoning
- Modeling teams/groups of agents
- Modeling other agents under partial observability
- Modeling belief change in other agents
- Modeling strategies and knowledge preconditions in games and knowledge-based programs
- Reasoning about utilities and preferences of other agents
- Formal/epistemic models of coordination and collaboration in multi-agent systems
- Planning formalisms for epistemic planning
- Action languages for epistemic planning
- Algorithms for epistemic planning
- Decidability and complexity results for epistemic planning variants
- Search heuristics for epistemic planning
- Epistemic planning benchmarks
- Learning epistemic action theories
- Methods for goal/intent/plan recognition
- Epistemic reasoning in social robotics
- (Dynamic) epistemic logic applied to goal recognition, diagnosis and/or planning
- Novel applications of epistemic planning
Submission: Friday, August 14, 2020, 23:59 UTC-12
Notification: Monday, September 14, 2020
Workshop: TBA (October 21-23, 2020)
Submissions should be formatted in AAAI style (see instructions in the Author Kit at https://www.aaai.org/Publications/Templates/AuthorKit20.zip) and be no longer than 8 pages (excluding references). Submissions will be double blind.
There will be no formal proceedings. Submissions sent to other conferences simultaneously are allowed. It is the responsibility of the authors to ensure that those venues allow for papers submitted to be already published in “informal” ways. Furthermore, we explicitly encourage the submission of relevant work that has already been published or accepted for publication elsewhere (in journals or conference proceedings). In that case, we do not require the authors to make an original submission, but instead allow submissions of the published/accepted paper with a reference to the publication venue.
At least one author of each accepted paper must attend the workshop in order to present the paper.
Submission link: https://easychair.org/conferences/?conf=epip20
Thomas Bolander (Technical University of Denmark)
Thorsten Engesser (University of Freiburg)
Andreas Herzig (CNRS, IRIT Toulouse)
Robert Mattmüller (University of Freiburg)
Sheila McIlraith (University of Toronto)
Bernhard Nebel (University of Freiburg)