Current Workshop |
To Be Announced! Synthesis of Epistemic Protocols
from 17 Aug 2015 through 21 Aug 2015
on five themes for Lorentz workshop To Be Announced
- Typically agents are primitive. They are names. Can
we define agents by their epistemic properties or their roles?
- What does agent A need to know about B to be able
to reason about B's knowledge ?
- How is agency related to control, to ensuring
- We assume a fixed set of agents typically. How do
we get beyond this, for example, can we assume an arbitrary number of
agents, an infinite number of agents, or a dynamically varying number of
- Our models view agents from the outside, how do we
incorporate agents' view of the world from inside agents ?
- How do we reconcile modelling communications as
formulas (such as public announcements), with a fixed meaning, and
modelling them as bit strings, that allow multiple meanings ?
- How do we understand group communication, in view
of different communication modes such as sending, copying, forwarding,
- How are protocols dependent on properties of
communication media ?
- How do we analyse
security implications, both in epistemic preconditions of communication
protocols and as epistemic goals ?
- What does it mean to know a communication protocol
- How do we distinguish concurrency from simultaneity
- How do we distinguish centralized solutions from
distributed solutions ? Is there a relation to the distinction between
outsider perspective and agent perspective ?
- How do we deal with unbounded concurrency ?
- Is there a knowledge / time tradeoff, whereby we
may know more by simply waiting ?
- Partial order models are standard in concurrency
theory. How is epistemic logic to make use of them ?
- How do we distinguish factual uncertainty from
epistemic uncertainty ?
- How do we relate epistemic update to probabilistic
- How do we construct epistemic protocols for (say)
mean payoff goals ?
- How do we relate quantitative probability with
qualitative likelihood ?
- Epistemic uncertainty is typically between full
descriptions, whereas incompleteness is about partial descriptions. What
is the trade-off between the two ?
- How do we distinguish plans, protocols, programs
and strategies ? Do we need to ?
- Plans are mostly considered to be sequences or
programs. Strategies usually lack structure, only their existence is
discussed. Do we need more structure in plans and strategies ?
- Game theory is almost all about randomized
('mixed') strategies and epistemic planning seems about deterministic
('pure') strategies. How do we explain this ?
- We usually describe the complexity of planning or
plan generation. What is the complexity of plans ?
- Strategy selection is often left implicit in game
theory and robot planning. Can strategy selection be explicit in the arena
of epistemic reasoning ?