Report on the International Workshop “Believing, Planning, Acting

advertisement
Report on the International Workshop
“Believing, Planning, Acting, Revising”
(Toulouse, June 26 and July 5, 2013)
Andreas Herzig
August 23, 2013
1. Background and aims of the workshop
Within the SINTELNET coordinated action, the workshop took place in the framework of the
activities of working group WG3 “Group attitudes”. The workshop continued the work in
Task 1.2 of WP1, whose aim is to set up a multi-disciplinary working group in area 3 and
which was started by the International Workshop on "The Cognitive Foundations of Group
Attitudes and Social Interaction" (May 1-June 1, 2012).
The very idea of SINTELENT is to revisit the basic concepts of philosophy, humanities, and
social sciences in the light of new forms of information technology-enabled social
environments. WG3 focusses on group attitudes such as group belief and group acceptance,
as well as joint goals and intentions. The study of such attitudes is fundamental for a theory of
social interaction, be it between human agents, between artificial agents, or in mixed
environments.
The aim of the workshop was to provide an overview over the state of the art in logics for
multiagent systems, in particular logics of knowledge and belief and logics of action. Most of
the talks presented papers that were to be presented at major conferences such as the 23rd Int.
Joint Conf. on AI (IJCAI 2013), the 27th Conf. on AI (AAAI 2013), the 14h Conf. on
Theoretical Aspects of Rationality and Knowledge (TARK 2013), the 4th Int. Workshop on
Logic, Rationality and Interaction (LORI 2013).
2. Organisation
The workshop was organised by the following persons: Sylvie Doutre, Andreas Herzig,
Laurent Perrussel, Valentin Goranko, and Frédéric Moisan. All but V. Goranko are affiliated
with the Institut de recherche en informatique de Toulouse IRIT. V. Goranko is affiliated with
Denmark Technical University and held an invited professor position at U. of Toulouse from
Feb. to July 2013 in the framework of the Excellence Lab CIMI.
The workshop was financially supported by the European Network for Social Intelligence
SINTELNET and (indirectly) by invited professor positions from U. Toulouse I. The
workshop also benefitted from organisational support from IRIT.
1
The workshop website is located on the SINTELNET pages
(http://www.sintelnet.eu/content/international-workshop-believing-planning-acting-revising).
A mirror page at IRIT was used during the preparation and setting-up of the workshop.
The workshop website contains the complete list of abstracts of the talks together with the
slides that were presented.
3. Abstracts of contributions
The list of contributions that were on the program of the workshop is below.
1. Hans van Ditmarsch (CNRS, LORIA)
"Awareness and knowledge"
ABSTRACT. Modal logics of knowledge model uncertainty. Logics of awareness
model incompleteness (as in vocabulary restriction) - a topic considered of great
interest in economics. I have been working on these matters with Tim French (Perth),
Fernando Velazquez (Sevilla), and Yi Wang (Bergen). We compare different
epistemic notions in the presence of awareness of propositional variables: the logics of
implicit knowledge (in which explicit knowledge is definable as implicit knowledge
plus awareness), explicit knowledge, and speculative knowledge. Speculative
knowledge goes back to the motivation in Levesque's 'A Logic of Implicit and Explicit
Belief': one can speculate over variables of which one is unaware, e.g. if you are
unaware of p, then p v ~p is still speculatively known by you. A cornerstone of this
framework is the notion of awareness bisimulation - this is the proper notion of
structural similarity on the structures enriched with awareness of propositional
variables proposed by Fagin and Halpern in 'Belief, awareness, and limited reasoning'.
A more 'standard' sort of bisimulation is also suitable for these logics. We provide
correspondence between bisimulation and modal equivalence on image-finite models
for these logics. The logic of speculative knowledge is equally expressive as the logic
of explicit knowledge, and the logic of implicit knowledge is more expressive than
both. The logics have complete axiomatizations. Dynamics can also be added: any
conceivable change of knowledge or awareness can be modelled in this setting. The
dynamic versions of all three logics are, surprising, equally expressive.
2. Florence Dupin de Saint-Cyr - Bannay (Université Paul Sabatier, IRIT,
Toulouse; joint work with Pierre Bisquert, Claudette Cayrol and MarieChristine Lagasquie)
"Enforcement in Argumentation is a kind of Update"
ABSTRACT. In the literature, enforcement consists in changing an argumentation
system in order to force it to accept a given set of arguments. In this paper, we extend
this notion by allowing incomplete information about the initial argumentation system. Generalized enforcement is an operation that maps a propositional formula
describing a system and a propositional formula that describes a goal, to a new
formula describing the possible resulting systems. This is done under some constraints
about the allowed changes. We give a set of postulates restraining the class of
enforcement operators and provide a representation theorem linking them to a family
of proximity relations on argumentation systems.
2
3. Jérôme Lang (CNRS, LAMSADE, Paris; joint work with Bruno Zanuttini)
"From knowledge-based programming to planning"
ABSTRACT. Knowledge-based programs (KBPs) are high-level protocols describing
the course of action an agent should perform as a function of its knowledge. The use
of KBPs for expressing action policies in AI planning has been surprisingly
overlooked. Given that to each KBP corresponds an equivalent plan and vice versa,
KBPs are typically more succinct than standard plans, but imply more on-line
computation time. Here we make this argument formal, and prove that there exists an
exponential succinctness gap between knowledge-based programs and standard plans.
Then we address the complexity of plan existence. Some results trivially follow from
results already known from the literature on planning under incomplete knowledge,
but many were unknown so far.
4. Yongmei Liu (Sun Yat-sen University, Guangzhou; joint work with Quan Yu,
Ximing Wen)
"Multi-agent Epistemic Explanatory Diagnosis via Reasoning about Actions"
ABSTRACT. The task of explanatory diagnosis conjectures actions to explain
observations. This is a common task in real life and an essential ability of intelligent
agents. It becomes more complicated in multi-agent scenarios, since agents’ actions
may be partially observable to other agents, and observations might involve agents’
knowledge about the world or other agents’ knowledge or even common knowledge of
a group of agents. For example, we might want to explain the observation that p does
not hold, but Ann believes p, or the observation that Ann, Bob, and Carl commonly
believe p. In this paper, we formalize the multi-agent explanatory diagnosis task in the
framework of dynamic epistemic logic, where Kripke models of actions are used to
represent agents’ partial observability of actions. Since this task is undecidable in
general, we identify important decidable fragments via techniques of reducing the
potentially infinite search spaces to finite ones of epistemic states or action sequences.
5. Yongmei Liu (Sun Yat-sen University, Guangzhou; joint work with Naiqi Li, Yi
Fan)
"Reasoning about State Constraints in the Situation Calculus"
ABSTRACT. In dynamic systems, state constraints are formulas that hold in every
reachable state. It has been shown that state constraints can be used to greatly reduce
the planning search space. They are also useful in program verification. In this paper,
we propose a sound but incomplete method for automatic verification and discovery of
state constraints for a class of action theories that include many planning benchmarks.
Our method is formulated in the situation calculus, theoretically based on
Skolemization and Herbrand Theorem, and implemented with SAT solvers. Basically,
we verify a state constraint by strengthening it in a novel and smart way so that it
becomes a state invariant. We experimented with the blocks world, logistics and
satellite domains, and the results showed that, almost all known state constraints can
be verified in a reasonable amount of time, and meanwhile succinct and intuitive
related state constraints are discovered.
3
6. Emiliano Lorini (CNRS, IRIT, Toulouse; joint work with Teddy Bouziat,
Andreas Herzig, Dominique Longin and Laurent Perrussel)
"Reasoning about moral agents"
ABSTRACT. The aim of this work is to provide a logical analysis of moral agency.
Although this concept has been extensively studied in moral philosophy and in
economics, it has been far less studied in the areas of logics of agents and multi-agent
systems. We discuss different aspects of moral agency such as the distinction between
desires and moral values and the concept of moral choice. All these concepts are
formalized in a logic of actions and agents’ mental attitudes including knowledge,
desires, moral values and preferences.
7. Faustine Maffre (Université Paul Sabatier, IRIT, Toulouse; joint work with
Andreas Herzig and Emiliano Lorini)
"Strategies and commitments"
ABSTRACT. ATL (Alternating-time Temporal Logic) is a modal logic which
expresses strategies of agents in a multi-agents system: can a set of agents ensure a
property by choosing specific actions? This logic provides a good basis but suffers
from some problems: one of them is it does not express explicitly actions composing
strategies. To fix this, we propose the logics ATLEA and ATLEP which adds actions
to the language of ATL to explicit them in addition to strategies. We also add
epistemic extension to ATLEA to reason about knowledge and uniform strategies.
8. Frederic Moisan (Université Paul Sabatier, IRIT, Toulouse; joint work with
Emiliano Lorini)
"When the group matters: an analysis of team reasoning and social ties"
ABSTRACT. In this paper, we investigate a central theory from the economics
literature whose aim is to clarify how agents reason and are able to cooperate when
acting as members of the same group: Bacharach's theory of team reasoning. After
discussing the various limitations of this theory in modeling various types of social
interactions, we introduce a new model of social ties which appears to provide a
simpler and more intuitive approach to modeling collaborative actions in the context
of complex strategic games where competing groups may coexist. We illustrate the
advantage of this model by performing a detailed comparative analysis of both
theories.
9. Laurent Perrussel (Université Toulouse 1 Capitole, IRIT, Toulouse; joint work
with Jerusa Marchi, Jean-Marc Thévenin and Dongmo Zhang)
"Relevant Minimal Change in Belief Update"
ABSTRACT. The notion of Relevance was introduced by Parikh in the belief revision
field for handling minimal change. It prevents the loss of beliefs that do not have
connections with the epistemic input. But, the problem of minimal change and
relevance is still an open issue in belief update. In this paper, a new framework for
4
handling minimal change and relevance in the context of belief update is introduced.
This framework goes beyond relevance in Parikh's sense and enforces minimal change
by first rewriting the Katzuno-Mendelzon postulates for belief update and second by
introducing a new relevance postulate. We show that relevant minimal change can be
characterized by setting agent's preferences on beliefs where preferences are indexed
by subsets of models of the belief set. Each subset represents a prime implicant of the
belief set and thus stresses the key propositional symbols for representing the belief
set.
10. Pilar Pozos Parra (Universidad Juarez Autonoma de Tabasco; joint work with
Kevin McAreavey and Weiru Liu)
"On the Merit of Selecting Different Belief Merging Operators"
ABSTRACT. Belief merging operators combine multiple belief bases (a profile) into a
collective one. When the conjunction of belief bases is consistent, all the operators
agree on the result. However, if the conjunction of belief bases is inconsistent, the
results vary between operators. There is no formal manner to measure the results and
decide on which operator to select. So, we propose to evaluate the result of merging
operators by using three ordering relations (fairness, satisfaction and strength) over
operators for a given profile. Moreover, a relation of conformity over operators is
introduced in order to classify how well the operator conforms to the de.nition of a
merging operator. By using the four proposed relations we provide a comparison of
some classical merging operators and evaluate the results for some specific pro.files.
11. François Schwarzentruber (Ecole Normale Supérieure Cachan, IRISA, Rennes;
joint work with Guillaume Aucher)
"On the Complexity of Dynamic Epistemic Logic"
ABSTRACT. In this talk, we present the language of DEL with event models and the
union construction for programs. We are interested in decision procedures for
automated reasoning. We prove that the model checking problem is PSPACEcomplete and that the satisfiability problem is NEXPTIME-complete.
12. Ezgi Iraz Su (Université Paul Sabatier, IRIT, Toulouse; joint work with Luis
Fariñas del Cerro and Andreas Herzig)
"Combining equilibrium logic and dynamic logic"
ABSTRACT. We extend the language of here-and-there logic by two kinds of atomic
programs allowing to minimally update the truth value of a propositional variable here
or there, if possible. These atomic programs are combined by the usual dynamic logic
program connectives. We investigate the mathematical properties of the resulting
extension of equilibrium logic: we prove that the problem of logical consequence in
equilibrium models is EXPTIME complete by relating equilibrium logic to dynamic
logic of propositional assignments.
5
Download