R.J. Brachman, V.P. Gilbert, H.J. Levesque ; An Essential Hybrid Reasoning System: Knowledge and Symbol Level Accounts of KRYPTON ; Proceedings of IJCAI 9, pp. 532-539, 1985.  Home/Search   Document Not in Database   Summary   Related Articles   Check

....formal semantics for specification of descriptive constants and axioms relating them. Remarkably, 99.9 of 7,000 vocabulary items in our natural language applications could be adequately axiomatizad in the taxonomic language. 1. Introduction Hybrid representation systems have been expietad before [9, 24, 31], but until now only one has been used in an extensive natural language processing system. KL TWO [31] based on a prepositional logic, was at the core of the mapping from formulae to lexical items in the Penman generation system [28] In this paper we report some of the design decisions made in ....

....inference in propositional logic is the goal of KL TWO. Limited aspects of universal quantification are achieved via allowing demons in the inference process. KL TWO and its classification algorithm [27] are at the heart of the lexicalization process of the text generator Penman [28] KRYPTON [9], which marries a frame system with first order logic. The frame system is designed to be less expressive than NIKL to allow rapid checking for disjointheSS of two class concepts in order to support efficient resolution theorem proving. KRYPTON has not as yet been used in any natural language ....

Brachman, R.J., Gilbert, V.P., and Levesque, H.J. An Essential Hybrid Reasoning System: Knowledge and Symbol Level Accounts of Krypton. Proceedings of IJCAI85, International Joint Conferences on Artifi- cial Intelligence, Inc., Los Angeles, CA, August, 1985, pp. 532-539.

....notions of KL ONE was rather unclear. TSLs are formal knowledge representation languages derived from KL ONE providing well defined semantics which enables the decision whether the inferences are sound and complete. A number of KR systems based on TSLs have been developed, for instance, Krypton [Brachman et al. 85] KL Two [Vilain 85] Back [Peltason et al. 89] Loom [MacGregorBates 87] Besides a component for defining concepts and reasoning about the relationships between concepts (terminological component, TBox) these systems include an assertional component (ABox) that allows the definition of ....

R. Brachman, V. Gilbert and H. Levesque. An Essential Hybrid Reasoning System Knowledge and Symbol Level Accounts of KRYPTON. In Proc. of [JCA[-85, pp. 532-539, Los Angeles, CA, August 1985. IJCAI, Inc.

....their values. To our knowledge, no existing data model supports retrieval of inherited categorical properties in a context, or the concept of inherited multi categorical property. Support for this kind of retrieval is the main goal of our work. Certain models, such as terminological languages [7, 5, 24, 21, 4] or F logic [6] do support instance inheritance. However, although terminological languages support instance inheritance with value specialization, they do not support reference to a context or inherited multi categorical properties (see also subsection 8.1) As for F logic, it supports default ....

....Work The form of instance inheritance supported by our data model is monotonic, in the sense that any new declaration can only specialize the value of an instance inherited property, but cannot override that property. Monotonic instance inheritance is supported by terminological languages [7, 5, 24, 21, 4] and by several deductive object oriented models, such as DOT [27, 26] and QUIXOTE [28, 29] In contrast, several systems, such as [6, 14, 20, 1, 13, 17] support non monotonic instance inheritance, where the value of an instance inherited property can be overridden over subclasses, or over ....

[Article contains additional citation context not shown here]

R.J. Brachman, V.P. Gilbert, H.J. Levesque, An essential hybrid reasoning system: Knowledge and symbol level accounts of KRYPTON, Proceedings of the 9th Intern. Joint Conference on Artificial Intelligence, 532-539 (1985).

....in figure 1. In most knowledge representation formalisms those taxonomies are represented using network formalisms which allow the definition of networks of classes like animals or birds together with a definition of attributes that hold for all elements of such a class. Krypton [2] is such a hybrid system which uses the network language Kl one for terminological and predicate logic for assertional knowledge. The connection to order sorted logic becomes obvious when the taxonomy is understood as the definition of sort information including a partial order for the subsumption ....

R. J. Brachman, V. P. Gilbert, and H. J. Levesque. An essential hybrid reasoning system: Knowledge and symbol level accounts of Krypton. In Proceedings of the International Joint Conference on Artificial Intelligence, pages 532--539, 1985.

.... Amongst the earliest formal results was that any interesting such system was, in the worst case, computationally intractable i.e. that the time required to compute classification could grow exponentially and that there were cases in which the no algorithm could be guaranteed to terminate at all [6] 6 . Since much of the point of the exercise had been to find a part of inference which was relatively easy, this was a disheartening result. It gave rise to the strategy of dividing knowledge representation systems into two: a terminological part or T Box with a very restricted language which ....

R. Brachman, R. Fikes, H. Levesque, An essential hybrid reasoning system; knowledge and symbol level accounts of KRYPTON, International Joint Conference on Artificial Intelligence (IJCAI-85) (Morgan Kaufmann, 1985) 532-539.

....checking, subsumption checking, retrieval, and querying. 1 Introduction and Motivation In the last decade many knowledge representation systems in the tradition of kl one [BS85] have been built, for example back [NvL88, Neb90] classic [BBMR89] kandor [Pat84] kl two [Vil85] krypton [BPGL85] loom [MB87] nikl [KBR86] sb one [Kob89] A common feature of these systems is the separation of the knowledge into a terminological part and an assertional part. Knowledge about classes of individuals and relationships between these classes is stored in the TBox, and knowledge concerning ....

R. J. Brachman, V. Pigman Gilbert, H. J. Levesque. \An essential hybrid reasoning system: knowledge and symbol level accounts in KRYPTON." In Proceedings of the 9th IJCAI, pp. 532-539, Los Angeles, Cal., 1985.

....The analysis of how efficiency is affected, when the expressivity increases, is an active research field. Two solutions are available if one needs a large choice of knowledge representation formalisms: i) to dispose of a large number of system building tools, or (ii) to use a hybrid system (e.g. [4]) We report on Mantra 1 , a system that supports hybrid knowledge representation which has been designed at the University of Karlsruhe during the last years and is still under development. The system was implemented according to the following design principles: i) several cooperating ....

....a particular domain. It is adequate to domains where the knowledge is largely unstructured and consists of a collection of independent facts. The main drawback of the logic formalism is the inherent inefficiency of the inference method: automatic deduction. Most of the existing hybrid systems [4] contain a first order logic method, because of its advantages: i) naturality of the representation, ii) modularity and flexibility of the represented knowledge which can easily be modified and extended, and additionaly, iii) classical logic has a formal semantics due to Tarski. Adopting the ....

R.J. Brachman, V.P. Gilbert, H.J. Levesque, An Essential Hybrid Reasoning System: Knowledge and Symbol Level Accounts of KRYPTON , Proceeding of IJCAI 9, 1985.

....for the wrong attribute are part of the answer. In addition it is possible to include any other membership conditions for instances of a query class using the logical representation. The integration of structural and logical representation formalisms leads to a hybrid query language. KRYPTON [BGL85], a derivative of KLONE and ancestor of Telos, uses both formalisms for its query language too but as in its data description language they are not combined to one closed format. In order to avoid the reformulation of similar more specialized queries, attributes of query classes can be ....

Brachman,R.J., Gilbert,V.P., Levesque,H.J. (1985). An essential hybrid reasoning system: knowledge and symbol level accounts of KRYPTON. Proc. Int. Joint. Conf. on Artificial Intelligence

.... 1980s the formal foundations of such systems were developed in terms of intensional logics (Maida and Shapiro, 1982) and complexity theory (Brachman and Levesque, 1984) and during the 1980s increasingly principled system designs were developed such as KL ONE (Brachman and Schmolze, 1985) KRYPTON (Brachman, Gilbert and Levesque, 1985), BACK (Nebel, 1990) LOOM (MacGregor, 1991) CLASSIC (Borgida, Brachman, McGuiness and Resnick, 1989) and KRIS (Baader and Hollunder, 1991) Deep theoretical foundations have been developed for such technologies in recent years (At Kaci, 1986; Nebel, 1990) The availability of these foundations ....

Brachman, R.J., Gilbert, V.P. and Levesque, H.J. (1985). An essential hybrid reasoning system: knowledge and symbol level accounts of KRYPTON. Proceedings of IJCAI85. pp.547-551. San Mateo, California, Morgan Kauffman.

....representation system, kl one [7] was an implementation of Brachman s [5] work on structured inheritance networks. In the last decade many knowledge representation systems based on these ideas have been built, for example back [37] classic [36] kandor [34] kl two [42] k rep [30] krypton [6], kris [3] loom [29] meson [13] nikl [41] sb one [23] and yak [10] Moreover, formal aspects of ter1 minological representation languages have been thoroughly investigated, with the highest emphasis having been placed on the decidability and complexity of the subsumption problem [24, 31, 39, ....

R. J. Brachman, V. Pigman Gilbert, and H. J. Levesque. An essential hybrid reasoning system: Knowledge and symbol level accounts in KRYPTON. In IJCAI-85 [20], pages 532-539.

....of C is always a subset of the extension of D. A whole family of knowledge representation systems have been built using these languages and for most of them complexity results for the subsumption algorithm are known (e.g. KL ONE[BrS85] BACK [Neb90] CLASSIC [Bo 89] KANDOR [PaS84] KRYPTON [Br 85], LOOM [MGB87] LITE ( Ron90] Ron92] on the other hand, is a variation to first order predicate logic where in particular the notion of object is revised from being an indivisible entity into being a temporal structure of versions. Each object version is then indivisible and unchanged in ....

Brachman, R.J., Pigman Gilbert, V., Levesque, H.J., 'An Essential Hybrid Reasoning System : Knowledge and Symbol Level Accounts in KRYPTON', in Proceedings of the International Joint Conference on Artificial Intelligence -IJCAI85, pp 532-539, 1985.

....CLP is a natural implementation vehicle for hybrid knowledge representation languages. Terminological concepts are often expressed as constraints [7] The implementation of Krypton extended the basic inference rule of the assertional reasoning mechanism by altering the meaning of unification [1]. With logic programming supporting the assertional component of a hybrid system, and the constraints the terminological component (as in [19] CLP offers precisely the right architecture since it provides logic programming with unification replaced by constraint solving. Given an underlying ....

R. Brachman, V. Gilbert and H. Levesque, "An Essential Hybrid Reasoning System: Knowledge and Symbol Level Accounts of Krypton", Proceedings of the 9 th IJCAI, 1985.

....However, the distinction is sometimes elusive because links can be used for both purposes simultaneously. In KL ONE type languages, there is a clear distinction between terminological information, which is used to define concepts, and assertional information [12] For 6 example, in KRYPTON [13] a TBox component represents definitional and terminological information and creates a hierarchy of class concepts; an ABox component makes assertions using first order predicate calculus. Unary predicates used in the ABox must be previously defined in the TBox. The distinctions own member ....

R. Brachman, V. Gilbert, and H. Levesque, "An essential hybrid reasoning system: Knowledge and symbol level accounts of KRYPTON," Proceedings of the Ninth International Joint Conference on Artificial Intelligence, IJCAI-85, pp. 532-539, 1995.

....modeled as (e.g. Lisp procedures attached to frame slots. The debate continues to this day [Nilsson 1991, Birnbaum 1991] but the complexity of reasoning about procedures (e.g. Etherington Reiter 1983] drove most researchers to declarative characterizations of their frame based knowledge (e.g. [Brachman, Gilbert Levesque 1989]) Pure PROC OO KEng researchers are rare these days, but some still keep the faith (e.g. Birnbaum 1991, Brooks 1991] Few PROC OO researchers are known amongst the HPKB and Sisyphus communities. IFA: InFerence Axioms. No explicit representation meta knowledge when modeling KBS and a single ....

Brachman, R., Gilbert, V. & Levesque, H. [1989], An Essential Hybrid Reasoning System: Knowledge and Symbol Level Accounts of Krypton, in J. Mylopoulos & M. Brodie, eds, `Readings in Artificial Intelligence and Databases', Morgan Kaufmann, pp. 293--300.

....they can also be used a constraint language within other formalisms, such as Horn rule languages. In fact, many applications would significantly benefit from having the expressive power of both formalisms. This has been the driving force behind previous work on designing hybrid languages (e.g. [2, 5, 8]) Two prime examples of such applications are information gathering from multiple sources [14, 12, 13] where we require the expressive power of both formalisms in order to model meta data about information sources, and the problem of modeling of complex physical devices [9, 10] in which we need ....

....works (e.g. AL log [8] TaxLog [1] had the goal of using a description logic as a rich typing language on the variables already appearing in the rules (which could also be recursive) In those works, only unary predicates from the description logics are allowed in the rules. Other works (e.g. [5, 11]) considered a more tight integration of the two for 3 One main reason for the limited coreference constraints in description logics is that subsumption becomes undecidable when coreference constraints are applied to roles with multiple fillers. 13 malisms, which is more in the spirit of ....

[Article contains additional citation context not shown here]

Ronald J. Brachman, Victoria P. Gilbert, and Hector J. Levesque. An essential hybrid reasoning system: Knowledge and symbol level accounts of krypton. In 14 Proceedings of the Ninth International Joint Conference on Artificial Intelligence, 1985.

....their values. To our knowledge, no existing data model supports retrieval of inherited categorical properties in a context, or the concept of inherited multi categorical property. Support for this kind of retrieval is the main goal of our work. Certain models, such as terminological languages [7, 5, 24, 21, 4] or F logic [6] do support instance inheritance. However, although terminological languages support instance inheritance with value specialization, they do not support reference to a context or inherited multi categorical properties (see also subsection 8.1) As for F logic, it supports default ....

....Work The form of instance inheritance supported by our data model is monotonic, in the sense that any new declaration can only specialize the value of an instance inherited property, but cannot override that property. Monotonic instance inheritance is supported by terminological languages [7, 5, 24, 21, 4] and by several deductive object oriented models, such as DOT [27, 26] and QUIXOTE [28, 29] In contrast, several systems, such as [6, 14, 20, 1, 13, 17] support non monotonic instance inheritance, where the value of an instance inherited property can be overridden over subclasses, or over ....

[Article contains additional citation context not shown here]

R.J. Brachman, V.P. Gilbert, H.J. Levesque, An essential hybrid reasoning system: Knowledge and symbol level accounts of KRYPTON, Proceedings of the 9th Intern. Joint Conference on Artificial Intelligence, 532-539 (1985).

....that can be used for implementing ontologies are description logics for representing declarative knowledge. Most of such languages support subsumption checking, automatic classification and consistency maintenance. Examples of such description logics include KL ONE [26] LOOM [97] KRYPTON [25], CYCL [94] A dedicated language for specifying ontologies is Ontolingua [74] which has become quite popular in the last years. Ontolingua is based on KIF the Knowledge Interchange Format [67] which is basically first order predicate logic extended with meta capabilities to reason about ....

R.J. Brachman, V.P. Gilbert, and H.J. Levesque, An Essential Hybrid Reasoning System: Knowledge and Symbol Level Accounts of KRYPTON, in: Proceedings IJCAI-85, 1985.

....24 5.1 Reasoning Techniques : 25 5. 2 Inclusions versus Concept Definitions : 26 6 Discussion 29 1 Introduction A general characteristic of many proposed Terminological Knowledge Representation Systems (TKRSs) such as kripton [BPL85] nikl [KBR86] back [QK90] loom [MB87] classic [BBMAR89] KRIS [BH91] and others (see [Ric91, WS92] is that they are made up of two different components. Informally speaking, the first is a general schema concerning the classes of individuals to be represented, their general properties and ....

Ronald J. Brachman, Victoria Pigman Gilbert, and Hector J. Levesque. An essential hybrid reasoning system: Knowledge and symbol level accounts in KRYPTON. In Proc. of the Int. Joint Conf. on Artificial Intelligence, pages 532--539, Los Angeles, Cal., 1985.

....to URmod each new formula inferred, and how it has been inferred (i.e. its depending from other formulae) Better that any explanation, a simple example will illustrate this simple idea. Example 1. We imagine using the M KRYPTON KR system 2 , an extension of the KRYPTON hybrid KR system (Brachman et al. 1985), as our KR mod, and Dempster Shafer theory of evidence (Shafer, 1976; Smets, 1988) as our UR mod. We want to model the following interesting problem: Robert is not very learned in Palaeontology: though he is almost sure that brontosauri are animals, he is not quite sure about which are the ....

Brachman, R.J., Pigman Gilbert, V. and Levesque, H.J. (1985) "An Essential Hybrid Reasoning System: Knowledge and Symbol Level Accounts of Krypton", Proc. of IJCAI-85: 532-539.

No context found.

R.J. Brachman, V.P. Gilbert, H.J. Levesque ; An Essential Hybrid Reasoning System: Knowledge and Symbol Level Accounts of KRYPTON ; Proceedings of IJCAI 9, pp. 532-539, 1985.

No context found.

R.J. Brachman, V.P. Gilbert, H.J. Levesque, An Essential Hybrid Reasoning System: Knowledge and Symbol Level Accounts of KRYPTON , Proceeding of IJCAI 9, 1985.

No context found.

Brachman, R.J., Gilbert, V.P., Levesque,H.J., An Essential Hybrid Reasoning System: Knowledge and Symbol Level Accounts of KRYPTON, Proceedings of IJCAI 9, pp. 532--539, 1985.

No context found.

, pages 532{ 539, Los Angeles, 1985.

No context found.

Brachman R.J., Gilbert V.P., Levesque H.J.: An essential hybrid reasoning system: knowledge and symbol level accounts of Krypton. Proceedings of IJCAI-85, Los Angeles CA, 1985, pages 532-539.

No context found.

Ronald J. Brachman, Victoria Pigman Gilbert, and Hector J. Levesque. An essential hybrid reasoning system: Knowledge and symbol level accounts in KRYPTON. In Proc. of the 9th Int. Joint Conf. on Arti#cial Intelligence #IJCAI-85#, pages 532#539, Los Angeles, 1985.

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