R. MacGregor, A deductive pattern matcher, Proc. AAAI-88, Nat. Conf. Artif. Intell., St. Paul, MN, 1988.  Home/Search   Document Not in Database   Summary   Related Articles   Check

....information from semi structured Web sources. We then provide a description of our framework for materializing data in mediators. Finally we outline the factors in selecting data to materialize and the portion focused on in this paper. 2. 1 SIMS Architecture In the SIMS system we use the LOOM [16] knowledge representation language (we can also view this as a data modeling language) for modeling data. The user is presented with an integrated view of the information in several different sources, which is known as the domain model. We describe the contents of the individual information ....

R. MacGregor. A deductive pattern matcher. In Proceedings of AAAI-88, The National Conference on Artificial Intelligence, St.Paul, MN, 1988.

....In Ariadne the mediator designer defines a domain model, which is an ontology of the application domain that integrates the information in the sources and provides a single terminology over which the user poses queries. The domain model is represented using the Loom knowledge representation system [MacGregor 1988]. Each information source is defined described using the terms in the domain model [Arens et al. 1996; Knoblock et al. 2001] The source descriptions are utilized to reformulate user queries into source queries. We have also developed Theseus, an efficient execution system for information agents ....

MacGregor, R. 1988. A Deductive Pattern Matcher. In the Proceedings of the Seventh Nation Conference on Artificial Intelligence, Saint Paul, Minnesota

....of the concepts which underly that domain. In current practice there is often a spectrum rather than a clear cut division between the ontological definition of a knowledge source and its rule based or schema specification. Some ontological description languages such as CLASSIC [6] or LOOM [28] permit the inclusion of a significant amount of operational detail in the definition of an ontology. Conversely, an object oriented approach encourages the production of schemata which take a concept based view of a domain rather than the traditional task based enterprise model. In order to ....

R.M. MacGregor. A deductive pattern matcher. In Proceedings of the Seventh National Conference on Artificial Intelligence, pages 403--408. AAAI Press, 1988.

....the reader to appreciate the complexity of Description Logics, and further understand the scope of this research, the following sub sections introduce Description Logics and GRAIL in an informal manner. 1.2. 1 Fundamentals of Description Logics Description Logics (DL) such as CLASSIC [19] LOOM [101, 102] and GRAIL [125] are a class of powerful declarative languages in the KL ONE family [174] The family is based on knowledge representation such as semantic networks frames with subsumption and multiple inheritance. Description Logic models have been used in a wide variety of data oriented ....

....1990s) and 2) it has been applied to manually creating terminological specifications. For a comprehensive review of design environments for earlier knowledge bases, see [110] The list of systems can be categorised into: # Description Logics based (CLASSIC [19] and NeoClassic [120] LOOM [101, 102, 103] and its front end Ontosaurus [156] and # frame based (Cyc [93] Conceptually Oriented Description Environment (CODE4) 147] Co4 [51] and its front end Hytropes [52] Ontolingua [54] Ontology Design Environment (ODE) 17] APECKS [158] WebOnto and Tadzebo [49] Generic Knowledge Base Editor ....

[Article contains additional citation context not shown here]

MACGREGOR, R. M. A Deductive Pattern Matcher. In Proceedings of AAAI '88, the Seventh National Conference on Artificial Intelligence (St. Paul, MN, 1988), pp. 403--408.

....of Quillian s original semantic network [42] Quite a number of KL ONE descendants exist, which makes this probably the largest and most successful family of implemented knowledge representation systems. Two excellent overviews are [43, 44] while some of the family members are described in [45, 40, 46, 47, 48] [49, 50] 51, 52, 53, 54] 55] Several other semantic networks that do not belong to the family of Description Logics exist, such as those described in [56, 57] Our own choice of using Object Oriented Databases instead of Description Logics is based on two kinds of reasons, which may be ....

MacGregor RM. A deductive pattern matcher. In Seventh National Conference on Artificial Intelligence 1988;403--408.

....to determine which rules unify with the posted goal. Our work stems from the EXPECT project [ Swartout and Gil, 1995; Gil, 1994; Gil and Melz, 1996 ] and its predecessor system EES [ Swartout , 1991 ] an architecture for developing knowledge based systems that is tightly coupled with LOOM [ MacGregor, 1988; 1991 ] a description logic system. EXPECT represents domain objects and classes in LOOM, as well as the goals of the methods to manipulate those objects. We have also extended the EXPECT matcher to work in a relaxed mode and retrieve rules that a posted goal based on the subsumption ....

R. MacGregor. A deductive pattern matcher. In , St Paul, MN, August 1988.

....there is no number restriction. A tractable and complete subsumption procedure is provided [2] An object oriented ABox with an abstraction classi cation propagation realizing procedure [6, 8] A Constraint Box which includes: disjointness, implies rules, transitive roles, test operator [1, 7]. Constraints inference ( a la Allen) added to ABox, to exploit set and time reasoning [5] A compositional uni cation based query language with typed variables and simple connectives. A simple mechanism for multiple KBs handling. Viewpoints according to the relevant beliefs theory ....

R. MacGregor. A deductive pattern matcher. In Proc. of AAAI-88, pages 403-408, St. Paul, MN, 1988.

....knowledge representation systems for its high expressivity and its provably sound and complete reasoning procedures. With respect to the comparable systems available in the research community, i.e. KRIS [ Baader and Hollunder,1991; Baader et al. 1994 ] Classic [ Brachman et al. 1991 ] and Loom [ MacGregor,1988 ] CRACK is more expressive, it is expandable to new constructs, it treats the conceptual and individual levels in a homogeneous way, it is modular, it is comparably fast. Special features handled currently by CRACK are: feature selection, agreement and disagreement; inverse roles; collections, ....

R. MacGregor. A deductive pattern matcher. In Proc. of AAAI-88, pages 403-408, St. Paul, MN, 1988.

....these optimizations have on the classi cation process are evaluated on three di erent sets of test data, which are described below. It should be noted that we do not claim that all the presented optimizations are novel. Similar optimizations can probably be found in many of the existing systems [27, 28, 38, 44]. Further, the optimizations on the rst level described below are very similar to methods that can be found in the conceptual graphs literature [25, 14, 26] and which have been used in the implementation of the peirce system [15] However, until now it was not possible to nd a coherent ....

....or on positive information (i.e. a subsumption test was successful) To use negative information during the top search phase one has to check whether for some predecessor z of y the test c z has failed. In this case, we can conclude that c 6 y without performing the expensive subsumption test [28]. In order to gain maximum advantage, all direct predecessors of y should have been tested before the test is performed on y [25] This can be achieved by using a modi ed breadth rst search where the already computed hierarchy is traversed in topological order, as described by Ellis [14] and ....

[Article contains additional citation context not shown here]

R. MacGregor. A deductive pattern matcher. In Proceedings of the 7th National Conference of the American Association for Articial Intelligence, pages 403-408, Saint Paul, MI, Aug. 1988. 44

....of access for all the information in such a domain. The user interacts directly with the SIMS mediator expressing queries against the domain model, without knowledge about the schemas or locations of the sources. The global model and the user queries are speci#ed in the Loom description logic #MacGregor 1988#, whichis the knowledge representation subsystem of the SIMS mediator. Selecting the sources, translating between global domain terms and source terms, and ordering the operations is the task of the query planner. 1 The speci#cation of the operators for distributed query processing and the ....

MacGregor, R. 1988. A deductive pattern matcher. In Proceedings of the Seventh National Conferenceon Arti#cial Intelligence.

....between them. This knowledge is derived from a study of a multilingual corpus of software manuals and is treated as language independent, an important requirement for multilingual generation. The knowledge base is hierarchically organized and represented using a terminological language, i.e. loom (MacGregor, 1988). As most knowledge bases represented in a terminological language such as loom, our knowledge base is separated into a terminology box, Multilingual Document Production 13 the T box, and an assertional box, the A box. The T box contains the definitions of all the concepts and relations relevant ....

MacGregor, R.: 1988, `A Deductive Pattern Matcher'. In: Proceedings of the 1988 Conference on Artificial Intelligence. St Paul, MN, pp. 403--408.

....changed. If so, it will request the additional information needed from the user. In this way, EXPECT helps users modify and adapt a knowledge based system while freeing them from the need to understand the details of the implementation. Figure 1 shows the architecture of EXPECT. EXPECT uses Loom [MacGregor 1988] to represent domain facts and domain ontologies. Loom is a description logic based representation. Like other description logics, Loom is based on a semantic network approach to knowledge representation. Concepts in Loom are descriptions of objects (which may or may not actually exist) while Loom ....

MacGregor, R. A Deductive Pattern Matcher. In Proceedings of the Seventh National Conference on Artificial Intelligence (AAAI-88). St. Paul, MN, August 1988.

....does not imply that the object it d escribes actually exists concepts are just descriptions that may or may not apply in any given situation. Domain facts, on the other hand, use domain terminology to represent what exists. To represent both domain facts and terminology, EXPECT uses Loom [MacGregor 1988]. Loom provides a descriptive logic representation language and a classifier for inference. Facts are represented as Loom instances, while terminology is represented using Loom concepts. Both instances and concepts are structured, frame based representations with slots that indicate relations in ....

MacGregor, R. A Deductive Pattern Matcher. In Proceedings of the Seventh National Conference on Artificial Intelligence. St. Paul, MN, August 1988.

....of these ontologies have not been included due to space limitations. 2.3. 1 CLASSIC: A system based on Description Logics Systems based on DLs, also known as terminological systems, are descendants of KL ONE [6] Some systems based on DLs are CLASSIC [5] used in our prototype) BACK [32] LOOM [17] and KRIS [1] The main features of the DL systems are described below: ffl The language contains unary relations called concepts which represent classes of objects in the domain and binary relations called roles which describe relationships between objects. Concepts and roles are created via ....

R. MacGregor. A deductive pattern matcher. In Proceedings AAAI-87, 1987.

.... Description of Ontologies: Description Logics In our proposal, ontologies are described using a system based on Description Logics (DL) These systems, also known as Terminological Systems, are descendants of KLONE [6] Some systems based on Description Logics are CLASSIC [5] BACK [37] LOOM [23] and KRIS [1] In our prototype we use CLASSIC but any other could be used because the features we use are common to all DL systems. The natural semantics of CLASSIC have been described in [4] In this section we discuss some of the common features of the DL systems that are relevant to the ....

R. MacGregor. A deductive pattern matcher. In Proceedings of AAAI-87, 1987.

....information from semi structured Web sources. We then provide a description of our framework for materializing data in mediators. Finally we outline the factors in selecting data to materialize and the portion focused on in this paper. 2.1. SIMS Architecture In the sims system we use the loom [19] knowledge representation language (we can also view this as a data modeling language) for modeling data. The user is presented with an integrated view of the information in several different sources, which is known as the domain model. We describe the contents of the individual information ....

R. MacGregor. A deductive pattern matcher. In Proceedings of AAAI-88, The National Conference on Artificial Intelligence, St.Paul, MN, 1988.

....an aircraft is handed over from one region to the next. The underlying representation for this example consisted of a semantic network of 18 instances, defined in terms of 27 air traffic domain concepts and 8 domain relations, implemented as frames in the Loom knowledge representation system [MacGregor 88] The structure planner built the paragraph tree shown in Figure 10. Though the form of the text closely mirrors that of the actual Air Traffic Control Manual [ASA 89] the differences in formatting are significant; and these differences make the manual much more readable. The manual contains ....

MacGregor, R. 1988. A Deductive Pattern Matcher. Proceedings of the 6th National Conference on Artificial Intelligence AAAI-88, St. Paul (696--701).

....in S, I satisfies A I C I (A I = C I ) Now, C is subsumed by D w.r.t. a schema S (S j= C v D for short) iff C I D I for all models I of S. 7 The various DLs investigated so far have different kinds of implementations: the most widely used, including classic[ABM 89] and Loom[Mac87] are implemented in a normalize then compare approach, which is quite different than standard theorem proving, and relies on finding a normal form for descriptions that detects nested incoherences, explicates implicit concepts, and removes redundancies. A second family of DLs, typified by ....

R.M. MacGregor. A deductive pattern matcher. In Proc. AAAI'87, pages 403--408, 1987.

....a sound and complete inference procedure for non recursive ALCNR carin programs; it forms the core of a backward chaining algorithm for carin programs, and forms the core of several algorithms for optimizing carin programs. 1 Introduction Description logics (e.g. classic [4] loom [15], back [17] kris [3] ALCNR [6] are declarative object oriented languages that have been designed especially for the purpose of representing domains with rich hierarchical structure, and have been used in several applications of Artificial Intelligence. A description logic is a subset of first ....

....approaches are either incomplete or guarantee only refutation completeness, relying on a full first order logic theorem prover. A different approach to integrating rules and description logics is to add rules as an additional constructor in description logics (e.g. classic [4] back [17] loom [15]) These works allowed only rules of a restricted form: C(x) GammaD(x) where C and D are concepts. Furthermore, the rules are generally not integrated in subsumption inferences but they are just used to derive additional knowledge about concept instances. MacGregor [16] and Yen [20] describe ....

R. M. MacGregor. A deductive pattern matcher. In Proceedings of the Seventh National Conference on Artificial Intelligence, pages 403--408, 1988.

....in the relational database literature, and for which, incidentally, subsumption is definitely decidable. The obvious route to follow in this direction is to add formulas with variables, or their equivalent (e.g. relational algebra expressions) into description languages. For example, in Loom [15] FOPC formulas may be given as arguments to the :satisfies concept constructor. Another approach is to integrate descriptions and Horn formulas, as in [12] Alternatively, one can consider ways of presenting queries that can create new objects or relationships (like relational algebra in ....

R.M. MacGregor, "A deductive pattern matcher", in Proceedings AAAI-87, St. Paul, Minnesota (1987), pp.403--408.

.... it corresponds to the predicate False) And, though not relevant to this paper, there is more logic dealing with individuals, facts, and their connection to concepts. The various DLs investigated so far have di#erent kinds of implementations: the most widely used, including classic[7] and loom[18], are implemented in a normalize then compare approach, which is quite di#erent than standard theorem proving, and relies on finding a normal form for descriptions that detects nested incoherences, explicates implicit concepts, and removes redundancies. A second family of DLs, typified by ....

R.M. MacGregor, "A deductive pattern matcher", in Proc. AAAI'87, pp.403--408.

....I satisfies A I # C I (A I = C I ) Now, C is subsumed by D w.r.t. a schema S (S = C # D for short) i# C I # D I for all models I of S. The various DLs investigated so far have di#erent kinds of implementations: the most widely used, including classic[ABM 89] and Loom[Mac87] are implemented in a normalize then compare approach, which is quite di#erent than standard theorem proving, and relies on finding a normal form for descriptions that detects nested incoherences, explicates implicit concepts, and removes redundancies. A second family of DLs, typified by ....

R.M. MacGregor. A deductive pattern matcher. In Proc. AAAI'87, pages 403--408, 1987.

....the Global Information System. Description Logics Systems based on DLs, also known as Terminological Systems, are descendants of KLONE [7] and allow us to define ontologies by using terminological descriptions. Some systems based on DLs are CLASSIC [6] used in our prototype) BACK [24] LOOM [18] and KRIS [1] The main features of the DL systems are described below: ffl The language contains unary relations called concepts which represent classes of objects in the domain and binary relations called roles which describe relationships between objects. Concepts and roles are created via ....

R. MacGregor. A deductive pattern matcher. In Proceedings AAAI-87, 1987.

.... of the resulting algorithms are usually not polynomial time ( 12] is a notable exception) and there is evidence that with large realistic KBs performance is slow enough to require additional heuristics [25] Most implemented systems that have a wider distribution and use, such as Classic[9] loom[23] and back[22] follow a so called normalize compare paradigm, where most of the reasoning work is performed in an initial normalization phase, whose goal is to find a normal form for concepts which explicates implicit facts, eliminates redundancies and detects inconsistencies. Once this is ....

....MARRIED, and now that holdout, e, is also asserted or inferred to be MARRIED, then b itself can be reclassified as an instance of all(friends,MARRIED) Without a priori knowledge, after an update of e one might have to reconsider every other individual in the KB. One alternative, used in Loom [23], is to keep track of all questions asked about an individual as part of the previous processing ( hits and misses ) and if answers to these do not change as a result of the update then no re processing is needed. Another alternative is to use an elaborate truth maintenance system (as available ....

R.M. MacGregor, "A deductive pattern matcher", in Proceedings AAAI-87, St. Paul, Minnesota (1987) 403--408.

.... 1992; 1994; 1995 ] Such an extension, the autoepistemic description logic ALCK, both constitutes a promising framework for reasoning about actions [ De Giacomo et al. 1996 ] and allows for the formalization of several non first order aspects of KR systems based on DLs [ Borgida et al. 1989; MacGregor, 1988 ] like procedural rules, defaults, a weak form of concept definition and some forms of closed world reasoning [ Donini et al. 1994; 1995 ] However, other nonmonotonic features of DL based KR systems, in particular role and concept closure inside the knowledge base, lack an intuitive ....

R. MacGregor. A deductive pattern matcher. In Proc. of AAAI-88, pages 403--408, 1988.

....information from semi structured Web sources. We then provide a description of our framework for materializing data in mediators. Finally we outline the factors in selecting data to materialize and the portion focused in this paper. 2. 1 SIMS Architecture In the sims system we use the loom [ MacGregor, 1988 ] knowledge representation language (we can also view this as a data modeling language) for modeling data. The user is presented with an integrated view of the information in several different sources, which is known as the domain model. We describe the contents of the individual information ....

MacGregor, R. 1988. A deductive pattern matcher. In Proceedings of AAAI-88, The National Conference on Artificial Intelligence.

....also borrows heavily from that of sims, sims is used to integrate data from mainly database systems whereas Ariadne is more focussed on integrating semi structured Web sources. In fact the sims architecture is typical of several other mediator systems. In the sims system we use the loom [ MacGregor, 1988 ] knowledge representation language (we can also perceive this as a data model) for modeling data. The user is presented with an integrated view of the information in several different sources which is known as the domain model. We describe the contents of the individual information sources in ....

MacGregor, R. 1988. A deductive pattern matcher. In Proceedings of AAAI-88, The National Conference on Artificial Intelligence.

....by using a clever order in which to make the subsumption calls, and by appropriately propagating the results of each call in the already computed part of the hierarchy, most of the n 2 calls can be avoided. 2 Such techniques are employed in many of the implemented terminological KR systems [15, 16, 19, 28], and similar methods have independently been developed in the conceptual graphs community [12, 8, 13, 9] The subsumption hierarchy provides only a limited amount of information about the interaction between defined concepts. For example, assume that we have appropriately defined concepts ....

R. MacGregor. A deductive pattern matcher. In Proceedings of the 7th National Conference of the American Association for Artificial Intelligence, pages 403--408, Saint Paul, MI, Aug. 1988.

....of these features can be achieved in our framework, thus clarifying both their semantics, and their interaction with the other parts of the knowledge base. Moreover, we show that epistemic sentences provide an account for weak forms of concept definitions similar to those found in LOOM [16] and other systems. This formalization makes it clear that weak definitions provide a form of incomplete reasoning that is both computationally advantageous, and semantically well founded. The paper is organized as follows. Section 2 recalls the basic notions about the concept language ALC, which ....

....as Epistemic Statements Recent studies on the formal properties of concept languages [4, 20, 21] show that one of the critical aspects of the implementation of knowledge representation systems based on concept languages is the treatment of inclusions. This problem is addressed for example in LOOM [16] by adopting a weak form of inclusion, which applies only to known individuals and disregards many inferences based on the use of contrapositives. In this section we argue that the class of epistemic sentences used in the formalization of trigger rules can be regarded as a form of weak inclusion ....

MacGregor, R. A deductive pattern matcher. In Proc. of the 6th Nat. Conf. on Artificial Intelligence AAAI-88, pages 403--408, 1988.

....aspects of object oriented database technology to address heterogeneity problems arising in the creation of integrated views. Pegasus also supports limited materialization; see below. Another recent system is SIMS [ACHK93] where integrated views are represented in the frame based model of LOOM [Mac88] a descendant of KL ONE [BS85] and query plans are generated dynamically using the LOOM inference engine. The commercial UniSQL system also provides support for the construction of virtual integrated views. One of the early projects advocating materialization of integrated views is WorldBase ....

R. MacGregor. A deductive pattern matcher. In Proc. AAAI-88, The Natl. Conf. on Artif. Intell., St. Paul, MN, 1988.

....axis) against number of concepts (horizontal axis) test was successful) To use negative information during the top search phase one has to check whether for some predecessor z of y the test c z has failed. In this case, we can conclude that c 6 y without performing the expensive subsumption test [28]. In order to gain maximum advantage, all direct predecessors of y should have been tested before the test is performed on y [25] This can be achieved by using a modified breadth first search where the already computed hierarchy is traversed in topological order, as described by Ellis [14] and ....

....we get the top search part of the enhanced traversal method. The enhanced top search procedure just described makes maximum use of failed tests. Alternatively, it is possible to use positive information. Before checking c y, one can look for successors z of y that have passed the test c z [28]. If there exists such a successor, one can conclude that c y without performing an actual subsumption test. Although we are only interested in minimizing the Optimization Techniques for Terminological Representation Systems 18 enhanced top subs (y,c) if marked (y, positive ) then return ....

R. MacGregor. A deductive pattern matcher. In Proceedings of the 7th National Conference of the American Association for Artificial Intelligence, pages 403--408, Saint Paul, MI, Aug. 1988. Optimization Techniques for Terminological Representation Systems 35

....is due to a relation property, extraneous to the entities themselves. As relations does not exist in the object model, we will prefer bind the opposition relation to the objects. The classification languages possess a notion more similar to what we are looking for. As an example, in LOOM [Mac88], there are particular concepts (called partitions) which specializations are mutually exclusive. This is more similar to what we want, as the opposition is expressed on the concepts. The object model has a similar notion with the abstract classes. Abstract classes define a general frame that ....

R.F. MacGregor. A deductive pattern-matcher. In Proc. of AAAI'88, volume 2, pages 403--408, 2929 Campus Drive, Suite 260, San Mateo, CA 94403, 1988. Morgan Kaufmann Publishers, Inc.

....as follows: Context = C 1 , V 1 ) C 2 , V 2 ) C k , V k ) We shall explain with examples the meaning of the symbols C i and V i and how they can be used. We shall also explain each example by using a description logic like language. This language has been used in and is exemplified by [BS85, BBMR89, Mac87, PS84, vLNPS87, KBR86]. Using this language, it is possible to define primitive classes and in addition specify classes using intensional descriptions phrased in terms of necessary and sufficient properties that must be satisfied by their instances. This can be used to express the collection of constraints that make up ....

....objects in the query context or the definition contexts of the object classes in the databases. We view ontology as the symbolic layer closest to the concepts in the real world. We are looking into the possibility of the Knowledge Interchange Format [GF92] and description logic based languages [BS85, BBMR89, Mac87, PS84, vLNPS87, KBR86] for context representation. 4.3.2 The Ontology Problem An ontology may be defined as the specification of a representational vocabulary for a shared domain of discourse which may include definitions of classes, relations, functions and other objects [Gru93] In constructing the contexts as ....

R. MacGregor. A deductive pattern matcher. In Proceedings AAAI-87, 1987.

.... the expressiveness (as exercised, e.g. in krypton [11] and classic [8] by using weaker semantics that permit tractable subsumption determination [38] or by restricting the inferential capabilities in a pragmatic way leading to sound and fast but incomplete systems, such as back [31] and loom [29]) Only recently, complete algorithms for a number of powerful term description languages for which the subsumption problem is NP hard have been developed [47, 20] It is yet unclear, however, whether these algorithms are feasible, i.e. whether the worst case does not show up for naturally ....

R. MacGregor. A deductive pattern matcher. In Proceedings of the 7th National Conference of the American Association for Artificial Intelligence, pages 403--408, Saint Paul, Minn., Aug. 1988. Terminological Reasoning and Information Management 28

....domain model that must be expressed in language are required to be subordinated to the upper model. It provides, among other things, the basic distinction between objects, processes, and qualities, and the definition of grammatical case roles. 3.3 The Knowledge Representation Language LOOM 2. 0 (MacGregor, 1988) has been chosen as the project s primary knowledge representation tool. There are several reasons for this choice: ffl LOOM is a well established knowledge representation tool that has been widely used in the Artificial Intelligence community. It has, for example, been used in the Penman Text ....

MacGregor, R. (1988). A Deductive Pattern Matcher. In Proceedings of the 1988 Conference on Artificial Intelligence, St Paul, Mn. American Association of Artificial Intelligence.

....contexts as a meet semi lattice. Inferences on a new context with respect to the knowledge present in the context set can now be supported by determining its position in the semi lattice We have expressed our context descriptions using DL expressions. Well known DL systems are KL ONE [BS85] LOOM [Mac87], BACK [vLNPS87] and CLASSIC [BBMR89] We are investigating the use of CLASSIC as the DL system for representing context. The advantage of using CLASSIC is that it is sufficiently expressive and has a polynomial time classification algorithm. Classification or taxonomies of schematic differences ....

....the semantic content of the information present in the various databases. In any attempt to represent the context of objects in a database, issues of language and vocabulary become important. We are looking into the possibility of the knowledge interchange format [GF92] and DL based languages [BS85, BBMR89, Mac87, PS84, vLNPS87, KBR86] for context representation. In designing the definition context of an object, it is necessary to choose the contextual coordinates and their values in a controlled manner. We are experimenting on using domain specific ontologies to construct these contexts in a methodical manner. In cases where a ....

MacGregor R. (1987) A deductive pattern matcher. In:<F3.17e+05> Proceedings<F3.733e+05> AAAI-87.

....emphasize typicality and defaults. Another key issue in the kl one community has been the tension between the need for expressiveness in the language and the desire to keep implementations computationally reasonable. Two somewhat different approaches can be seen: nikl [17] and subsequently loom [19], added expressive power to the original kl one language, and admitted the possibility of incomplete classification. krypton [12] and subsequently kandor [26] on the other hand, emphasized computational tractability and completeness. While neither of these approaches is right for every ....

....Subsumption is defined formally in [18] and [4] Concept a subsumes concept b iff instances of b are instances of a in all possible interpretations. 5 Note that classic and its cousins all do normal inheritance of properties. Most of these systems are strictly monotonic for simplicity, but loom [19] has a default component. 4.2 From LISP Functions to Languages The realization that the structure of a concept is the only source of its meaning, and that any is a hierarchy is induced by such structures, leads to another significant point of departure for the classic system. classic has a true ....

MacGregor, R. M., "A Deductive Pattern Matcher," Proc. AAAI-87, St. Paul, MN, pp. 403--408.

....and they are based on a clear semantics while associated with algorithms that provide classification services. The computational and expressive properties of description logics have been extensively studied. Several systemshavebeen built basedondescription logics (e.g. CLASSIC [3] LOOM [8], BACK [12] KRIS [1] they have been used in several applications (e.g. 16, 14] A description logic knowledge base includes descriptions of concepts representing classes of objects in the domain. A concept is defined by the conditions that must be satisfied by elements in the class. A ....

....In particular, theoretical results ( 10] show that the range restriction for roles combined with the number restriction constructors cause subsumption checking to be intractable in the worst cases. In parallel, several systemshavebeen built basedon description logics (e.g. CLASSIC [3] LOOM [8], BACK [12] KRIS [1] Specific description systems do not necessarily offer all the constructors that are considered in this paper. For instance, CLASSIC does not include the range restriction constructor for roles. The counterpart of its restricted language is that the subsumption algorithm ....

R. M. MacGregor, `A deductive pattern matcher', in Proceedings of the Seventh National Conference on Artificial Intelligence, pp. 403--408, (1988).

.... Knoblock 1998) SIMS and Ariadne are mediator systems that provide integrated access to heterogeneous sources in an application domain by building a model for the domain and mapping the contents of the sources to this domain model. The domain model is expressed in the Loom description logic (MacGregor 1988). The user poses queries in terms of the domain model and the system generates a plan that answers the query by combining information from the available relevant sources. SIMS has focussed more on the integration of databases and structured sources, while Ariadne addresses the issues arising in ....

MacGregor, R. 1988. A deductive pattern matcher.

....person , John has two children , Philip is John s son , Angela is John s daughter , both Philip and Angela are college graduates . The integration of terminological capabilities with rules is intended to address three problems with 1 We use the syntax of LOOM knowledge representation system[6] to define concepts and relations in this paper. rule based systems that critics have identified as hindering system maintenance and limiting the ability to generate high quality explanations and justifications [8, 9] First, rules fail to explicitly separate different kinds of knowledge; ....

....can be implemented by integrating the realizer with an efficient pattern matching algorithm (e.g. RETE match algorithm) C. 1 The CONCRETE Matching Network The semantic pattern matcher in CLASP is implemented by combining Forgy s Rete matching algorithm[20] with the deductive matcher of LOOM [6], which is a counterpart of Vilain s KL TWO s realizer[14] The rule compiler builds a CONcept Classification RETE (CONCRETE) net as rules are loaded into the rule base. As external changes are made to the facts database, the LOOM matcher computes assertional changes that can be deduced from the ....

R. M. MacGregor, "A deductive pattern matcher," In Proceedings of AAAI-88, 1988.

No context found.

R. MacGregor, A deductive pattern matcher, Proc. AAAI-88, Nat. Conf. Artif. Intell., St. Paul, MN, 1988.

No context found.

R. M. MacGregor. A deductive pattern matcher. In Proceedings of AAAI-88, pages 403--8, 1988.

No context found.

R. MacGregor. A deductive pattern matcher. In Proc. of AAAI-88, pages 403-408, St. Paul, MN, 1988.

No context found.

R. MacGregor, "A deductive pattern matcher," in Proceedings of AAAI-88, The National Conference on Artificial Intelligence, St. Paul, MN, USA, 1988.

No context found.

R. MacGregor. A Deductive Pattern Matcher. In Proceedings of AAAI-88, The National Conference on Artificial Intelligence, pages 403--408, St. Paul, Minnesota, USA, August 1988.

No context found.

R. MacGregor, A deductive pattern matcher, Proc. AAAI-88, Nat. Conf. Artif. Intell., St. Paul, MN, 1988.

No context found.

MacGregor R.: A Deductive Pattern Matcher. Proceedings of AAAI-88, St.Paul MINN, 1988, pages 403-408.

No context found.

R.M. MacGregor, "A deductive pattern matcher", in Proc. AAAI'87, pp.403--408.

No context found.

MacGregor R, "A Deductive Pattern Matcher", Proceeding of AAAI-88, the national conference of artificial intelligence, ST Paul, 1988.

No context found.

R. M. MacGregor. A Deductive Pattern Matcher. In Proceedings of the Seventh National Conference on Artificial Intelligence, St. Paul, Minn, August, 1988, pp 403--408. 225

First 50 documents  Next 50

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