J. Crawford, "Access-Limited Logic-A Language for Knowledge Representation", Ph.D. Thesis, The University of Texas as Austin, 1990.  Home/Search   Document Details and Download   Summary   Related Articles   Check

....and thus the ability to add to the target text information that was omitted from the sohrce. We will assume that all the KB based repreentations can be treated as sets of logical assertions (although they can of course be implemented in a variety of ways, including the frame based system [Crawford, 90] that we are using) SOURCE LA.GUAGE understandinn SOURCE KBLF, mapping to. interlingua TARGET LANGUAGE STRING tactical generatio 7 TARGET KBLF strategic generation INTERLINGUA KNOWLEDGE BASE EXPRESSION Figure l: An In erlingua Based Architecture for MT To translate a ....

J. Crawford, "Access-Limited Logic-A Language for Knowledge Representation", Ph.D. Thesis, The University of Texas as Austin, 1990.

.... frames ; Each frame encodes a CG along with addition information, in the form of access paths, about how to infer values of particular nodes (see Section 3. 2) This working note describes now conceptual graphs can be mapped onto the representation language Algernon [Crawford and Kuipers, 1991, Crawford, 1990] Section 3 describes one way in which CGs can be encoded in Algernon, where a set of rules encodes the various implications which a CG contains. Section 4 discusses this and alternative ways that CGs could be encoded with Algernon, and their relative merits and weaknesses. 2 Introduction to ....

Crawford, J. (1990). Access-limited logic: A language for knowledge representation. Technical Report AI90-141, Dept CS, Univ Texas at Austin, Austin, TX. (Phd thesis).

....( likes x chocolate) sister John x) meaning Is there someone who likes chocolate who is John s sister is not legal because the first clause does not have a bound frame argument and thus would require a global search of a 3 2 Sections 2.1 through 2. 4 are derived from Chapters 1 and 7 of [Crawford, 1990]. single indexed, frame based) knowledge base. In practice, this limitation does not severely restrict expressiveness because a knowledge base can always be constructed with one root frame containing relations to every other frame in the KB, thus allowing the equivalent of a global search ....

....this restriction does not seriously limit the expressiveness of the language. Many logic programs use access limitation to reduce the execution time of their systems (by reducing the branching factor) even though they are not required to do so. 2.4. 2 Time complexity As shown in Chapter 7 of [Crawford, 1990], the upper bound on runtime of a primitive query or assertion in an ALL system with the restrictions shown below is proportional to: c 5 x o 2 x r 5 x f 5v m where c is the the maximum number of clauses in the antecedent of any rule in the KB, o is the number of unique queries and ....

Crawford, J. (1990). Access-Limited Logic -- A Language for KnowledgeRepresentation. PhD Thesis, University of Texas at Austin, TR AI90-141.

....idraw) and one or more sentences of descriptive text. Its output is a fragment of a knowledge base, which can be equivalently described as either a set of ground logical assertions or a set of frames, slots, and values. In the current implementation, the Algernon knowledge representation language [Cra90, CK92] is used to maintain the knowledge base. We are using the Figure Understander as the input processing module in a magnetic fields problem solving system [RK94] Our approach is motivated by the observation that a figure is a communication act involving at least two agents [Gri75] ffl The author ....

Crawford, J. Access-Limited Logic - A language for knowledge representation. Ph.D. dissertation, TR AI-90-141, A.I. Lab, Dept. of Computer Sciences, Univ. of Texas at Austin, 1990.

....in object centered hierarchies with inheritance. Ontolingua provides an architecture for translating from this specification language into forms that can be efficiently stored and reasoned about by target representation systems. Currently, there are translators into LOOM, Epikit, and Algernon [Crawford and Kuipers 90], as well as into a canonical form of KIF. We plan to continue development of Ontolingua and to demonstrate its viability by working with engineers who will use the system to develop multi use ontologies. We will focus the continuing Ontolingua work on: Development of general purpose seed ....

J.M. Crawford and B. Kuipers; "Access-Limited Logic -- A Language for Knowledge representation"; University of Texas at Austin dissertation; Available as Technical Report number AI90-141; Artificial Intelligence Laboratory, The University of Texas at Austin.

....through the terminological hierarchy might be brought into the context by treating any deduction from terminological knowledge as a single rule application. One could also retrieve obvious related items for the context using Crawford s notion of the accessible portion of the knowledge base [ Crawford, 1990 ] Levesque s notion of limited inference [ Levesque, 1984 ] or other mechanisms that guarantee small computational costs for re trieval. In a longer version of this paper [ Etherington and Crawford, 1992 ] we discuss the context selection process in more detail. 4.3 The Mitigating Nature ....

Crawford, J.M. 1990. Access-Limited Logic--A Language for Knowledge Representation. Ph.D. Dissertation, University of Texas at Austin, Austin, Texas.

....selected in advance by a human system designer. Automatic specification of awareness is necessary, if a system is for general purpose or the system is very large scale. Both of two are limited within finite (depth awareness) on any existential or imaginable knowledge processing system. 3] and [4] try to cope with the first problem by restricting reasoning power. It is useful to deal with huge knowledge base. This problem also relates to the problem of logical omniscience. Approaches from Modal logic are applied to the problem[5, 11, 13] While the second one is much more difficult to ....

J. M. Crawford: "Access-Limited Logic --- A Language for Knowledge Representation", Univ.of Austin technical report "AI90-141", Austin, Texas 18712, (1990).

....awareness and finite resources of reasoning. There are several researches into finite awareness and limited knowledge. Some of them look at modal theoretic properties of awareness to cope with logical omniscience[Fagin, Halpern, 85, Konolidge, 86, Levesque, 84] Others[Crawford, Kuipers, 89, Crawford, 90] treat limited access to knowledge for efficiency. These are not necessarily suitable to treat comprehensibility. Machine)Learning is a high complexity task and so selecting an appropriate awareness or selecting an appropriate conceptual bias (words to describe hypotheses to be learned) is ....

Crawford J.M.: "Access-Limited Logic --- A Language for Knowledge Representation ", University of Austin technical report "AI90-141", Austin, Texas 18712, 1990.

....has significantly improved the efficiency of basic inference operations (e.g. performing joins) there is still a need for methods that guide their application so that questions can be efficiently answered from large CG knowledge bases. In this paper, we apply ideas from Access Limited Logic [1] to Conceptual Graphs, as a means of guiding CG inference. Access Limited Logic specifies access paths that (1) relate together the concepts in a knowledge base, and (2) constrain inference to follow only those paths when answering queries that require navigating the knowledge base. Although this ....

.... polynomial time inference for pure Algernon, which did not include statements involving existential quantification or restrictions on the cardinality of relations (and thus is similar to Datalog [14] and achieved tractibility by limiting the number of bindings a variable in a rule could take [1]. In contrast, our interest has been in full Algernon (due to our requirements for these features) and we exploit different sources of incompleteness, and hence efficiency, in its reasoning. There are three sources of incompleteness in full Algernon which we exploit. Consider the following ....

James Crawford. Access-limited logic: A language for knowledge representation. Technical Report AI90-141, Dept CS, Univ Texas at Austin, Austin, TX, Oct 1990.

....mid 1970s. Its members include FRL [71, 70] HPRL [44, 72] and GOLDWORKS. Several other FRSs do not exist within a larger family. They include PROTEUS [73, 68] FROBS [58] OZONE [45] KRL [7, 46, 8] BB [32] LOOPS [94, 17] KB [23] SNePS [84, 85] RHET[3, 56] TELOS [59] PARKA [24] ALGERNON [18, 19], FRAPPE [25] Conceptual Graphs [92] MOPS, ART, and NEXPERT. HPRL Goldworks FRL PARMENIDES FrameKit CRL KnowledgeCraft SRL Strobe Class JOSIE OPUS THEO ARLO CycL RLL Kappa KEE Unit Package LOOM KRYPTON KANDOR SPHINX KREP NIKL KL ONE KL TWO SB ONE KRS King Kong KRIS BACK CLASSIC MUNIN Figure 1: ....

J.M. Crawford. Access-Limited Logic --- A language for knowledge-representation. PhD thesis, University of Texas at Austin, 1990. Technical report AI90-141.

....Corpus Humans Large Scale KB No No No No No Yes System vs. Human No No No No No Yes Table 8: Evaluation Methodologies the Krl [5] and Kodiak [66] systems. More recent work in this area has demonstrated how views can be reified, i.e. encoded as first class objects in their representations [57, 11, 28]. Related research in intelligent tutoring systems explored how different views could be used to model different aspects of the domain and, hence, to answer different kinds of questions [58] Work in automated knowledge acquisition systems has shown how to use views to constrain search [50] ....

J. Crawford. Access-limited logic---a language for knowledge representation. Technical Report AI Laboratory AI90-141, University of Texas at Austin, Austin, Texas, 1990.

....for instances of the subclass. R rules work directly on all instances of a class so there is no need for a separate working memory nor even a 3 Such path based rules are not new, having appeared explicitly as access limited rules in the Algernon implementation of accesslimited logic [10, 7]. rule declaration : rule identifier ; rule definition : rule class name : identifier f condition = action g action : statement condition : binding boolean expression binding condition boolean expression condition binding : ....

J. M. Crawford. Access-Limited Logic---A language for knowledge representation. PhD thesis, Department of Computer Sciences, The University of Texas at Austin, 1990. Also published as Technical Report AI 90-141, Artificial Intelligence Laboratory, The University of Texas at Austin.

....match part of a RETE network. We require that trigger conditions bind variables when they are first referenced. This simple restriction eliminates the need for the join part of the RETE network and makes trigger conditions equivalent to a knowledge representation mechanism known as access paths [28]. Access paths provide a well defined semantics for ordering conjunctive queries so that knowledge base access is contained and controlled, and therefore is more efficient. In our mechanism, the discrimination network uses the event type, event level, attribute and or link as the match keys at its ....

J. Crawford, "Access-Limited Logic -- A Language for Knowledge Representation," PhD Thesis, University of Texas at Austin, Technical Report AI90-141 (1990).

....starting from a root object (the this pointer variable) 7 Such rules have the same access capabilities as member functions. While this property may seem unsurprising and even expected by object oriented programmers, it is not present (to our knowledge) in any other rule language save Algernon [14, 15], the forebear of R . Other rule languages support pattern matching rules that can perform arbitrary joins between unrelated objects. In our view this violates the locality of reference designed into a domain model, making program behavior harder to predict and understand. 8 7 Note that global ....

....practice (e.g. 18] Below we present empirical results (using both synthesized rules and real world application programs) that further support the efficiency of path based rules. Although reduction in complexity was a principle objective of the research upon which pathbased rules was founded [14, 15], an equally important aspect of path based rules for R is that it allows rules to be viewed as class members, and thus integrated cleanly into the object oriented framework. 16 Rule Processing There are three phases to rule processing: triggering, evaluation, and execution. A rule is ....

J. M. Crawford. Access-Limited Logic---A language for knowledge representation. PhD thesis, Department of Computer Sciences, The University of Texas at Austin, 1990. Also published as Technical Report AI 90-141, Artificial Intelligence Laboratory, The University of Texas at Austin.

....starting from a root object (the this pointer variable) 7 Such rules have the same access capabilities as member functions. While this property may seem unsurprising and even expected by object oriented programmers, it is not present (to our knowledge) in any other rule language save Algernon [15, 16], the forebear of R . Other rule languages support pattern matching rules that can perform arbitrary joins between unrelated objects. In our view this violates the locality of reference designed into a domain model, making program behavior harder to predict and understand. 8 Compared to the ....

....practice (e.g. 19] Below we present empirical results (using both synthesized rules and real world application programs) that further support the efficiency of path based rules. Although reduction in complexity was a principle objective of the research upon which path based rules was founded [15, 16], an equally important aspect of path based rules for R is that it allows rules to be viewed as class members, and thus integrated cleanly into the object oriented framework. 7 Note that global variables and static data members can be used in rule conditions, but changes to those variables ....

J. M. Crawford, Access-Limited Logic---A language for knowledge representation, Ph.D. thesis, Department of Computer Sciences, The University of Texas at Austin, 1990, Also published as Technical Report AI 90-141, Artificial Intelligence Laboratory, The University of Texas at Austin.

No context found.

Crawford, J. M., and Kuipers, B. (1990). Access-Limited Logic --- A language for knowledge-representation. University of Texas at Austin dissertation. Available as Technical Report number AI90-141, Artificial Intelligence Laboratory, The University of Texas at Austin.

....polynomial in the size of the locally accessible portion of the knowledge base. Furthermore, ALL is Socratically Complete: for any fact which is a semantic consequence of the knowledge base, there is a sequence of preliminary queries and assumptions, after which a query of the fact will succeed [10, 9]. While these formal properties are important, they do not necessarily guarantee that ALL can actually be used to represent knowledge. There are two further claims which can only be shown empirically: 1) that the syntactic restrictions on ALL still allow one to express common sense knowledge ....

.... and second, that these inference mechanisms are Socratically Complete (thus guaranteeing that any logical consequence of a knowledge base can be inferred after some series of leading questions) Time complexity and Socratic Completeness are formal properties which have been shown elsewhere [9]. The claim that the representation and inference mechanisms in ALL are sufficient for a large class of common sense reasoning problems can necessarily only be proven or disproven empirically. In this paper we overview results from the application of the lisp implementation of ALL to three fields: ....

[Article contains additional citation context not shown here]

Crawford, J. M. (1990). Access-Limited Logic --- A language for knowledge-representation. University of Texas at Austin dissertation. Available as Technical Report number AI90141, Artificial Intelligence Laboratory, The University of Texas at Austin.

.... rules must start with some root object and can reach other objects only by following paths of pointers (access paths) Such path based rules are not new, having appeared explicitly as access limited rules in the Algernon implementation of access limited logic (Crawford Kuipers 1991; Crawford 1990)) Aside from this limitation on their conditions, pathbased rules are similar in concept to OPS style rules, having a condition and an action. The conditions of path based rules are evaluated in response to certain activities in the rest of the system. If the condition is satisfied, then the ....

....such as ART IM (Inf 1987) CERS (Miranker et al. 1993) ILOG rules (Albert 1994) RAL C (Forgy 1994) and Rete (Hal 1993) Further, none of these systems has as close an integration with C as R has. Of course, the most closely related system to R is Algernon (Crawford Kuipers 1991; Crawford 1990). R moves the path based rules of Algernon into the object oriented language C and integrates the rules into the C object system. The work in active object oriented databases has some similarities to R . In particular, integrity constraints and triggers in object oriented databases are ....

Crawford, J. M. 1990. Access-Limited Logic---A language for knowledge representation. Ph.D. Dissertation, Department of Computer Sciences, The University of Texas at Austin. Also published as Technical Report AI 90-141, Artificial Intelligence Laboratory, The University of Texas at Austin.

....tightly bounded. For example, if the KB has a terminological component (c.f. Brachman Schmolze 1985) chains through the type hierarchy might be brought in by treating deductions from terminological knowledge as single rule applications. Also, obvious related items can be retrieved using Crawford s (1990) notion of the accessible portion of the KB, Levesque s (1984) notion of limited inference, or other mechanisms that guarantee cheap retrieval. The significant feature of our approach is the synergy between the two components: context focuses the consistency check on the part of the KB most likely ....

Crawford, J. 1990. Access-Limited Logic--A Language for Knowledge Representation. Ph.D. Dissertation, University of Texas at Austin, Austin, Texas.

No context found.

J. M. Crawford, 1990. "Access-Limited Logic---A language for knowledge representation", doctoral dissertation, Department of Computer Sciences, The University of Texas at Austin. Also published as Technical Report AI 90-141, Artificial Intelligence Laboratory, The University of Texas at Austin.

No context found.

Crawford, J. M., and Kuipers, B. (1990). Access-Limited Logic --- A language for knowledge-representation. University of Texas at Austin dissertation. Available as Technical Report number AI90-141, Artificial Intelligence Laboratory, The University of Texas at Austin.

No context found.

J.M. Crawford. Access-Limited Logic --- A language for knowledge-representation. PhD thesis, University of Texas at Austin, 1990. Technical report AI90-141.

No context found.

James Crawford. Access-Limited Logic-A language for knowledge representation. Ph.D. thesis, department of Computer Sciences, The University of Texas at Austin, 1990. Also published as technical report AI 90-141, Artificial Intelligence Laboratory, The University of Texas at Austin.

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