C. Ghidini and F. Giunchiglia. Local models semantics, or contextual reasoning = locality + compatibility. Articial Intelligence, 127(2):221-259, April 2001.  Home/Search   Document Not in Database   Summary   Related Articles   Check

....the potential applications to security are shown. In this work we show how to extend the framework in terms of symbolic model checking. The notions of multiple views and multiple languages are mostly inspired by the works by Giunchiglia and his collaborators in the field of Multi Language Systems [10, 9]. Other related works use combinations of modal temporal logics [14] although the mechanization is not based on symbolic model checking techniques. 12] present an automata theoretic approach to temporal modal logic (restricted to the case of single nesting of beliefs) applied to the ....

C. Ghidini and F. Giunchiglia. Local Models Semantics, or Contextual Reasoning = Locality + Compatibility. Artificial Intelligence, 127(2):221--259, April 2001.

....or more knowledge bases, the information in a knowledge base is constrained to the information in a di erent knowledge base) Our approach is based on the theory of context. Contexts formalisms have been proposed in AI by McCarthy, Buva c and Mason [3, 4] and by Giunchiglia, Ghidini, and Sera ni [8, 6, 18]. Furthermore, in [17, 15] contexts have been proposed for the representation of distributed knowledge in presence of heterogeneity. The theory so far, however, does not provide any support for revising the contents of contexts. In this paper, we give a formal de nition of contextual update, and ....

....operation. In Section 4, we provide an algorithm that computes this update operator. We conclude the paper by a comparison with similar approaches (Section 5) 2 Logic of contexts To represent distributed knowledge we adopt the logic of contexts developed by Giunchiglia, Sera ni, and Ghidini [8, 6]. In this section we recall the key concepts of this logic. 2 Logic of context represents knowledge in a set of theories (contexts) each of them constituting a partial, approximate, and perspectival description of the world [7, 2] Each context partially describes a portion of the world from a ....

[Article contains additional citation context not shown here]

C. Ghidini and F. Giunchiglia. Local models semantics, or contextual reasoning = locality + compatibility. Articial Intelligence, 127(2):221-259, April 2001.

....the idea of distributed context aware applications. The main contributions of the paper are two. First, we discuss the problem of what notion of context is adequate to design distributed context aware applications, and propose an answer which is inspired to the framework of Local Models Semantic [8] and MultiContext Systems [9] second, as an illustration, we describe the architecture of a multi agent system for semantic based information and knowledge management in distributed environments, such as the Internet or a big corporate Intranet. 2 Context in distributed systems Context is often ....

....provide a systematic model of the relations between partial and perspectival representations of an environment that autonomous agents use to achieve their goals and to interact with other agents. In what follows, we propose such a formal model, largely inspired to Local Models Semantics (LMS) [8] and its proof theoretical counterpart called MultiContext Systems (MCS) 9] A technical presentation of LMS is outside the scope of this paper. Therefore we only present the main intuitions through a simple example, and refer the interested reader to the technical papers in the bibliography at ....

[Article contains additional citation context not shown here]

F. Giunchiglia and C. Ghidini. Local Models Semantics, or Contextual Reasoning = Locality + Compatibility. In Proceedings of the Sixth International Conference on Principles of Knowledge Representation and Reasoning (KR'98), pages 282-289, Trento,

....we may say that the work on context provides two basic intuitions, namely that (i) reasoning processes are always local (namely, happen in a context) and that (ii) local reasoning processes may be related one another by some constraints that guarantee their compatibility. This is what [9] presents as the two fundamental principles of contextual reasoning: locality and compatibility. These ideas have been developed into many di erent directions, and the papers we have selected for this course present some of these directions. 1 More detailed information can be found at: ....

....P. Introduction to contextual reasoning. An Arti cial Intelligence Perspective [8] 2. Benerecetti M. Bouquet P. Ghidini C. Contextual Reasoning Distilled [2] II Logics 1. Ghidini C. Giunchiglia F. Local Models Semantics, or Contextual reasoning = Locality Compatibility [9]. 2. Sera ni L. Giunchiglia F. ML systems: A Proof Theory for Contexts [13] 3. Ghidini C. Sera ni L. Distributed First Order Logic [7] 4. Giunchiglia F. Sera ni L. Multilanguage Hierarchical Logics (or: how we can do without modal logics) 10] III Applications 1. Bouquet ....

F. Giunchiglia and C. Ghidini. Local Models Semantics, or Contextual Reasoning = Locality + Compatibility. In Proceedings of the Sixth International Conference on Principles of Knowledge Representation and Reasoning (KR'98), pages 282-289, Trento,

....and Sera ni proposed MultiContext Systems (MCS) as a proof theoretical framework for contextual reasoning. Recently, Ghidini and Giunchiglia proposed Local Models Semantics (LMS) as a model theoretic framework for contextual reasoning, and used MCS to axiomatize many important classes of LMS [18]. From a conceptual point of view, Ghidini and Giunchiglia argued that contextual reasoning can be analyzed as the result of the interaction of two very general principles: the principle of locality (reasoning always happens in a context) and the principle of compatibility (there can be ....

....to do with physical location (e.g. when an agent a facing another agent b says Open the window to the right , b must be able to transform this request in a request to open the window which is to its left) Also the Magic Box scenario (see section 3. 2 below) discussed by Ghidini and Giunchiglia [18] requires the ability to capture the relationship between di erent viewpoints, and also to reason about the beliefs of agents with a di erent viewpoint to integrate what they say in a single, coherent reasoning process. However, the most typical example of contextual reasoning as relate to ....

[Article contains additional citation context not shown here]

C. Ghidini and F. Giunchiglia. Local models semantics, or contextual reasoning = locality + compatibility. Articial Intelligence, 127(2):221{ 259, April 2001.

....modeling such kind of systems. Answers to feasibility questions along this way will depend on the kind of axioms that form the DB knowledge as well as the way an agent s knowledge is represented. In this perspective, an integration of the theory of contexts [Giu93, GS94, GB97, GS98, GS99, BBG00, GG01, SG00] to the present setting might be a productive direction of research, as inductive paradigms eventually give a dynamics oriented plus value to the static view of the compatibility of contexts presented e.g. in [GG01] First order logic provides a rich class of database queries. However, some ....

.... integration of the theory of contexts [Giu93, GS94, GB97, GS98, GS99, BBG00, GG01, SG00] to the present setting might be a productive direction of research, as inductive paradigms eventually give a dynamics oriented plus value to the static view of the compatibility of contexts presented e.g. in [GG01] First order logic provides a rich class of database queries. However, some plausible queries are not rst order expressible, so that we may be led to ask for stronger logics. 7.4 Incomplete, noisy , imperfect, recursive environments A game between two agents may be seen as concerning ....

C. Ghidini and F. Giunchiglia. Local Models Semantics, or contextual reasoning = locality + compatibility. Articial Intelligence, 127(2):221-259, 2001.

....since we suppose that whenever an agent desires to change the hearer s mental state and it believes that this is possible by means of a communicative act, then it intends to do it; and vice versa, if the agent intends to change the hearer s mental state, then it also desires to do it. Following [4, 13, 14, 15], we use propositional contexts to formalize agents mental states. A context is de ned as a set of formulae closed under a set of inference rules (a theory) For any agent i, its sets of beliefs and intentions are represented by the contexts B i and I i , respectively. A formula in the context ....

F. Giunchiglia and C. Ghidini. Local Models Semantics, or Contextual Reasoning = Locality + Compatibility. In Proceedings of the Sixth International Conference on Principles of Knowledge Representation 24 and Reasoning (KR'98), pages 282-289. Morgan Kaufmann, 1998. Also IRST-Technical Report 9701-07, IRST, Trento, Italy.

.... Istituto per la Ricerca Scienti ca e Tecnologica Via Sommarive 38050 Trento (Italy) bouquet cs.unitn.it serafini itc.it Abstract. We investigate the relationship between two well known formalizations of context: Propositional Logic of Context (PLC) 4] and Local Models Semantics (LMS) [13]. We start with a summary of the desiderata for a logic of context, mainly inspired by McCarthy s paper on generality in AI [15] and his notes on formalizing context [16] We brie y present LMS, and its axiomatization using MultiContext Systems (MCS) 14] Then we present a revised (and simpli ....

....and discuss in detail how the two formalisms ful ll (or do not ful ll) each of them. 1 Introduction This paper is an investigation on the relationship between two well known formalizations of context, namely the Propositional Logic of Context (PLC) 4] and Local Models Semantics (LMS) [13], axiomatized via Multi Context Systems [14, 12] MCS) 1 . Both PLC and LMS MCS address issues that were raised by McCarthy in his papers on generality in AI [15] and in his notes on formalizing context [16] These issues can be summarized as a list of general desiderata for an adequate logic ....

[Article contains additional citation context not shown here]

F. Giunchiglia and C. Ghidini. Local Models Semantics, or Contextual Reasoning = Locality + Compatibility. In Proceedings of the Sixth International Conference on Principles of Knowledge Representation and Reasoning (KR'98), pages 282-289. Morgan Kaufmann, 1998. Also IRST-Technical Report 9701-07, IRST, Trento, Italy.

....of each single context, and that any PLC structure with partial vocabulary is equivalent to a PLC structure where the universal vocabulary is associated to each context. 2 Logics for Contexts In this section we summarize the two formalisms for contexts we will compare. LMS has been presented in [13, 10], while [14, 12] propose a class of proof systems for LMS called Multi Context Systems (MCS) PLC is presented in [4] 2.1 Local Model Semantics and Multi Context Systems Let fL i g i2I be a family of languages de ned over a set of indexes I (in the following we drop the index i 2 I) ....

C. Ghidini and F. Giunchiglia. Local Models Semantics, or Contextual Reasoning = Locality + Compatibility. Articial Intelligence, 2001. To appear.

.... Istituto per la Ricerca Scienti ca e Tecnologica Via Sommarive 38050 Trento (Italy) bouquet cs.unitn.it serafini itc.it Abstract. We investigate the relationship between two well known formalizations of context: Propositional Logic of Context (PLC) 4] and Local Models Semantics (LMS) [11]. We start with a summary of the desiderata for a logic of context, mainly inspired by McCarthy s paper on generality in AI [15] and his notes on formalizing context [16] We brie y present LMS, and its axiomatization using MultiContext Systems (MCS) 14] Then we present a revised (and simpli ....

....and discuss in detail how the two formalisms ful ll (or do not ful ll) each of them. 1 Introduction This paper is an investigation on the relationship between two well known formalizations of context, namely the Propositional Logic of Context (PLC) 4] and Local Models Semantics (LMS) [11], axiomatized via Multi Context Systems [14, 13] MCS) 1 . Both PLC and LMS MCS address issues that were raised by McCarthy in his papers on generality in AI [15] and in his notes on formalizing context [16] These issues can be summarized as a list of general desiderata for an adequate logic ....

[Article contains additional citation context not shown here]

C. Ghidini and F. Giunchiglia. Local Models Semantics, or Contextual Reasoning = Locality + Compatibility. Articial Intelligence, 127(2):221-259, 2001.

....logical agents with more and more sophisticated forms of rationality. Logical agents can represent (and communicate) their own as well as other agents thinking processes, and are able to build arguments by means of their own inference rules where: either each agent has a logic associated with it, [12], 16] with rules of inference which relate the different logics of the different agents; or, each agent is defined together with its specific inference rules and communication modalities [3] Logical agents are able to combine belief updating, goal updating and practical reasoning in a BDI ....

F. Giunchiglia and C. Ghidini. Local models semantics, or contextual reasoning = locality+compatibility. In Working papers of Common Sense 98 (the fourth Symp. On Logical Formalization of Commonsense Reasoning), London, 1998.

....the LRM semantically in terms of relational spaces. A relational space is a pair consisting of a set of databases (the peers) and a domain relation which makes explicit the relations among the domains of the databases. The LRM semantics are an extension of the Local Models Semantics, introduced in [2]. Section 3 introduces coordination formulas that relate the contents of the peer databases and de ne what it means for a coordination formula to be satis ed (with respect to a relational space) The crucial step is the quanti cation across the distinct domains of di erent databases. Section 3 ....

....by induction by showing that each rule preserves satis ability under assignments. Completeness is proved by showing that an i consistent set of formulas , i.e. i : is not derivable from , has a canonical relational space. A similar construction, restricted to the propositional case is given in [2]. Details of the proof will be provided in the full paper. The completeness result given above allows us to generalize Reiter s syntactic characterization of relational databases to relational spaces. We start by recalling Reiter s result (in a slightly di erent, but equivalent, formulation) De ....

C. Ghidini and F. Giunchiglia. Local models semantics, or contextual reasoning = locality + compatibility. Articial Intelligence, 127(2):221-259, April 2001.

.... Logic of Context (LoC) 13] which illustrates the divide and conquer type (in fact, we will review also Dinsmore s theory of partitioned representations, as it provides an interesting variation of a divide and conquer theory) and Local Models Semantics Multi context systems (LMS MCS) [24, 32], which illustrates the compose and conquer type. Of course, there are several other (formal) theories of context in KRR, for example structured contexts with bred semantics [22] and the type theoretic foundation for context [51] However, we decided to focus on LoC and LMS MCS for two reasons: ....

....hierarchical relations in compose and conquer theories are assimilated with any other relation, and does not require to hardwire in the logic any order over contexts. The neatest formalization of a compose and conquer theory of context is Ghidini and Giunchiglia s Local Models Semantics (LMS) [24], together with its proof theoretical counterpart, namely Giunchiglia and Sera ni s MultiContext Systems (MCS) 32] LMS is based on two very general principles, that for our purposes we restate as follows: principle of locality: reasoning always happen in a local theory (a context) ....

[Article contains additional citation context not shown here]

C. Ghidini and F. Giunchiglia. Local Models Semantics, or Contextual Reasoning = Locality + Compatibility. Articial Intelligence, 127(2):221-259, April 2001. 38

....are called formulae. Given a set of labeled formulae Gamma, Gamma j denotes the set of formulae ffl jhfl; ji 2 Gammag. From now on we say that OE is a i formula to specify that OE is a formula of the language L i . The semantics for fL i g is an extension of Local Models Semantics defined in [12]. The structure on which the languages fL i g are interpreted is constituted of two components: i) the local interpretation for each L i , and (ii) for each pair of languages L i and L j a binary relation between their interpretation domains. Let M i be the set of all the first order models of L ....

F. Giunchiglia and C. Ghidini. Local Models Semantics, or Contextual Reasoning = Locality + Compatibility. In Proceedings of the Sixth International Conference on Principles of Knowledge Representation and Reasoning (KR'98), Trento, June 2-5 1998. To appear. Also IRST-Technical Report 9701-07, IRST, Trento, Italy.

....of the federated database must be defined in terms of these partial descriptions. The semantics for federated database proposed in this paper, called Local Model Semantics for federated database (LMS hereafter) is an extension to first order language of the semantics of contexts proposed in [11]. It is based on the intuition that the databases of a federation can be though as partial views (thought as contexts) on a common world. LMS for a federated database is constituted by a set of local semantics each formalizing the view of a database, as it was not part of the federation, and by ....

....and when a federated database is of a given federated database schema. Being a federated database composed of a set of distributed and autonomous databases, its formal model must be specifiable in terms of the composition of the models of each single database. We take the perspective described in [11] for the semantics of contexts by formalizing each database as a context. The formal semantics associated to each database i represents the description of the real world from the i th partial point of view. Therefore the formal semantics of a federated database on the federated database schema hfS ....

F. Giunchiglia and C. Ghidini. Local Models Semantics, or Contextual Reasoning = Locality + Compatibility. In Proceedings of the Sixth International Conference on Principles of Knowledge Representation and Reasoning (KR'98), pages 282--289. Morgan Kaufmann, 1998. Also IRST-Technical Report 9701-07, IRST, Trento, Italy.

....point to understand, specify, and verify the behavior of a multi agent system for electronic commerce. The semantics for information integration proposed in this paper, called Local Model Semantics (LMS hereafter) is an extension to first order languages of the semantics of contexts proposed in [9]. It is based on the intuition that the database of an agent can be though as a partial view (thought as a context) on a common world. The paper is structured as follows. In Section 2 we introduce motivating examples. In Section 3 we review the basic concepts of semantics of databases. In Section ....

....notion of information integration state, and we define when an information integration state satisfies the constraints of an information integration schema. An information integration state is defined by formalizing the database of each agent as a context and by taking the perspective described in [9] for the semantics of contexts and of contextual reasoning. The formal semantics associated to each database i represents the description of the current state of the real world from i th partial point of view. Therefore the formal semantics of the information integration schema hfS i g; fIC ij gi ....

F. Giunchiglia and C. Ghidini. Local Models Semantics, or Contextual Reasoning = Locality + Compatibility. In Proceedings of the Sixth International Conference on Principles of Knowledge Representation and Reasoning (KR'98), pages 282--289. Morgan Kaufmann, 1998. Also IRST-Technical Report 9701-07, IRST, Trento, Italy.

....OM pairs, however, allow the improper inference rules (52) to be part of the formalism. Indeed OM pair are based on Multi Context Systems [14, 31] a formal framework which allows us to cope with rules across di erent theories. From Multi Context System we can borrow the semantics for such rules [16], and we can compare di erent re ection principles in a well founded formalism, as we have done in this paper. Metaprogram (theory) generation In the non amalgamated approach the metatheory is de ned only informally for any object program. The metaprogram generation is not part of the logic. ....

F. Giunchiglia and C. Ghidini. Local Models Semantics, or Contextual Reasoning = Locality + Compatibility. In Proceedings of the Sixth International Conference on Principles of Knowledge Representation and Reasoning (KR'98), pages 282-289. Morgan Kaufmann, 1998. Also IRST-Technical Report 9701-07, IRST, Trento, Italy.

.... shifting; however, it formalizes a quite weak form of partiality (via the use of partial functions for interpreting a global language) and only a special form of push pop (i.e. making explicit or implicit the context itself) MultiContext systems (MCS) 14] and their Local Model Semantics (LMS) [9], provide a logic for contextual reasoning based on the principles of locality and compatibility. These principles impose that: 1. each context c i is associated with a di erent formal language L i , used to describe what is true in that context. The semantics of L i is local to the context ....

C. Ghidini and F. Giunchiglia. Local Models Semantics, or Contextual Reasoning = Locality + Compatibility. Articial Intelligence, To Appear.

....an inference engine, expressing the reasoning capabilities of the subsystem. Following this idea DFOL syntax is composed of a family of rst order languages, each language describing a piece of the global knowledge contained in a subsystem. DFOL semantics is an extension of Local Models Semantics [11]. It is de ned in terms of sets of rst order interpretations for the languages of the di erent subsystems and relations between domains of interpretation of the di erent languages. DFOL calculus is based on ML systems [12] It is composed of a set of rst order natural deduction systems, ....

....L i . When no ambiguity arises, labeled formulae are called formulae. Given a set of labeled formulae , j denotes the set of formulae f jj : 2 g. From now on, a formula of the language L i is called i formula . The semantics for fL i g is an extension of Local Models Semantics de ned in [11]. The languages fL i g are interpreted in a structure which consists of two components: i) a local interpretation for each L i , and (ii) a binary relation between the interpretation domains of each pair of languages L i and L j . Let M i be the set of all the rst order models of L i on a given ....

[Article contains additional citation context not shown here]

F. Giunchiglia and C. Ghidini. Local Models Semantics, or Contextual Reasoning = Locality + Compatibility. In Proc. of the 6th Int. Conference on Principles of Knowledge Representation and Reasoning (KR'98), 1998. Morgan Kaufmann.

....in [11, 7] is based on the fact that different contexts may have different structural properties and may be related in different ways with other contexts. Inspired by the same intuitions as ML systems, a semantics for contextual reasoning, called Local Models Semantics (LMS) has been defined in [9]. LMS is based on the ideas that (i) each context has its own local semantics and (ii) relations between different contexts are compatibility relations. In [8] the authors apply LMS in order to model agent s beliefs and to provide the semantics for MBK, a ML system which has a standard formulation ....

....aa formalize the beliefs that the agent a ascribes to itself. Iterating the nesting, the view aaa formalizes the view of the agent a about beliefs about its own beliefs, and so on 2 . In this section we are principally interested in applying propositional Local Models Semantics, introduced in [9], in order to define the classes of models for the belief systems defined in [10] which have a standard formulation using normal modal logics. In order to achieve this, we focus our attention on single agent beliefs in which a is on top of an infinite chain of views. The infinite chain ....

[Article contains additional citation context not shown here]

F. Giunchiglia and C. Ghidini. Local Models Semantics, or Contextual Reasoning = Locality + Compatibility. In Proceedings of the Sixth International Conference on Principles of Knowledge Representation and Reasoning (KR'98), pages 282-- 289, Trento, 1998. Morgan Kaufmann.

....with some important modal systems. The goal of this analysis is to provide intuitions about the expressive power and conceptual importance of the multicontext systems defined. Till Section 5 we limit ourselves to the case of only two reasoners. This is a very strong hypothesis which, among 3 [13,14] give a context based semantics to some of the systems for non logical omniscience presented in this paper. This work allows us to give a semantic interpretation to some (but not all ) the possible forms of reality introduced below. 3 other things, forces us to deal only with the case of no ....

F. Giunchiglia and C. Ghidini. Local Models Semantics, or Contextual Reasoning = Locality + Compatibility. Technical Report 9701-07, IRST, Trento, Italy, February 1997. To appear in the proceedings of the AAAI Fall 1997 symposium on context in KR and NL. 30

....In most of the cases, indeed, each DB of a FDB has its own semantics which corresponds to a partial description of the real world. Therefore the semantics of the FDB must be defined in terms of these partial descriptions. This is analogous to the intuition of the semantics of contexts described in [8], in which each DB of the federation is considered to be a context. Following this intuition, in [18; 19] we have defined a context based semantics for federated databases, called Local Model Semantics for FDB (LMS hereafter) which is an extension to first order language of the semantics of ....

....DB of the federation is considered to be a context. Following this intuition, in [18; 19] we have defined a context based semantics for federated databases, called Local Model Semantics for FDB (LMS hereafter) which is an extension to first order language of the semantics of contexts proposed in [8]. LMS for a FDB is constituted by a set of local semantics each formalizing the view of a DB, as it was not part of the federation, and by a compatibility relation between local semantics, which represents the fact that only certain combinations of views are allowed, as views are on the same ....

[Article contains additional citation context not shown here]

F. Giunchiglia and C. Ghidini. Local Models Semantics, or Contextual Reasoning = Locality + Compatibility. Technical Report 9701-07, IRST, Trento, Italy, February 1997. Submitted to AAAI Fall 1997 symposium on context in KR and NL.

No context found.

F. Giunchiglia and C. Ghidini. Local Models Semantics, or Contextual Reasoning = Locality + Compatibility. In Proceedings of the Sixth International Conference on Principles of Knowledge Representation and Reasoning (KR'98), pages 282-289. Morgan Kaufmann, 1998. Also IRST-Technical Report 9701-07, IRST, Trento, Italy.

No context found.

F. Giunchiglia and C. Ghidini. Local Models Semantics, or Contextual Reasoning = Locality + Compatibility. In Proceedings of the Sixth International Conference on Principles of Knowledge Representation and Reasoning (KR'98), pages 282--289. Morgan Kaufmann, 1998.

No context found.

F. Giunchiglia and C. Ghidini. Local Models Semantics, or Contextual Reasoning = Locality + Compatibility. In Proceedings of the Sixth International Conference on Principles of Knowledge Representation and Reasoning (KR'98), pages 282--289. Morgan Kaufmann, 1998.

First 50 documents

Online articles have much greater impact   More about CiteSeer.IST   Add search form to your site   Submit documents   Feedback  

CiteSeer.IST - Copyright Penn State and NEC