Levinson, R. A. 1994. UDS: A universal data structure. Technical Report UCSC-CRL-94-15, University of California, Santa Cruz.  Home/Search   Document Details and Download   Summary   Related Articles   Check

.... Conceptual graphs are significant as the basis for an operational knowledge science and technology encompassing natural language, formal language, visual language, and the wide range of reasoning processes that may be based on them , says Gaines [Gai93] Works by Ellis, Levinson [Ell93] [Lev94] and others, have strengthen a fourth pillar besides the three ones emphasized by Gaines: natural language, formal logic, and visual language of CG theory : graph theory and graph algorithms. We share the Peirce group point of view that graph theory and algorithms are at the core of the CG ....

R. Levinson. UDS: A Universal Data Structure. In W.M. Tepfenhart, J.P. Dick, and J.F. Sowa, editors, Lecture Notes in Artificial Intelligence, 835, Proceedings of the International Conference on Conceptual Structures'94, pages 230--250. Springer Verlag, 1994.

.... supplies the GS with a pruned lattice structure which may be represented by an inheritance network [11] Numerous works have been carried on knowledge retrieval from structures similar to the GS [6] 12] 24] Levinson proposed a representation of such structures supporting efficient retrieval [16]. However, the problem we are interested in is not to find an efficient representation of the Generalization Space but to construct it. In order to construct the GS, we have adopted an ascending method. The method builds upon the objects descriptions to construct generalizations of these ....

R. Levinson. Uds : A universal data structure. In Lectures Notes in AI n-835, Proc. Second International Conference on Conceptual Structures, ICCS'94, pages 230--250. 1994.

....be encountered (it is the case when labels on nodes are reasonably different one from another) MLL91] 3. restrict the problem to conceptual graphs classes for which the problem turns to be polynomial [MC92] CM92] LB94] 4. use efficient pre processing on conceptual graphs [LE91] Lev92] Lev94] EL94] 5. organise the database according to the specialization hierarchy [Ell91] Lev92] Ell93] Lev94] and (a) traverse the hierarchy so that predictable comparisons are avoided [LE91] Ell91] Lev92] Ell93] Lev94] see [Woo91] BFH 92] for a general discussion) b) ....

....the problem to conceptual graphs classes for which the problem turns to be polynomial [MC92] CM92] LB94] 4. use efficient pre processing on conceptual graphs [LE91] Lev92] Lev94] EL94] 5. organise the database according to the specialization hierarchy [Ell91] Lev92] Ell93] Lev94] and (a) traverse the hierarchy so that predictable comparisons are avoided [LE91] Ell91] Lev92] Ell93] Lev94] see [Woo91] BFH 92] for a general discussion) b) restricting the problem to some finite set of graphs, for each graph define a unique code significant to the hierarchy ....

[Article contains additional citation context not shown here]

R. Levinson. Uds: A universal data structure. In Proceedings of the 2nd Int. Conf. on Conceptual Structures, Maryland, USA, August 1994. Springer Verlag. Lecture Note in Artificial intelligence #835.

....ffl x y and y x implies x = y, ffl x y and y z implies x z. The implication order over logical formula is a preorder which is a partial order over equivalence classes. There has been significant research into the problem of classification of a KB into a hierarchy [Lev84, Lev85, Lev92, LE92, Lev94, Ell92, Ell95a, Ell93, Ell95b, ELR94, EL94, BHN 92, BFH 94, Mac88, Mac91] 4 Classification CS433 Lecture 12 Classification is name for the process of finding the closest related objects in a set of object for a given object. Stonebraker [Sto89] gave the following example as an ....

Robert A. Levinson. UDS: A universal data structure. In 2nd International Conference on Conceptual Structures, pages 230--250. Spring-Verlag, 1994.

....does not incorporate a linguistic knowledge base (or lexicon) into its approach, it has proven to be a very useful graphical modeling technique with good Nl paraphrasing possibilities. Although Cgs are designed for knowledge representation, they have been proposed to be used as Lcmts as well [2] [19], on which we will focus now. Cgs are a system of logic designed to map to and from Nl in a simple and direct manner. They combine extended quantifiers and type labels in a readable graphic notation. The Cg approach is mainly used in the Artificial Intelligence field. Cgs are very powerful, which ....

R. Levinson. UDS: A Universal Data Structure. In W.M. Tepfenhart, J.P. Dick, and J.F. Sowa, editors, Conceptual Structures: Current Practices, Lecture Notes in Artificial Intelligence. Springer-Verlag, 1994.

....three ways: removal of redundant data, use of simple instructions which ignore redundant checks when performing matching, and by sharing common processing between graphs. In future work, we will examine methods for handling complex conceptual graphs for use in such domains as chemistry. Levinson [7] has recently developed a new tuple and skeleton based compression technique called UDS. UDS is based on a new compact representation of conceptual graphs which make storage and retrieval more efficient. UDS can be extended so that processing in a hierarchical search can shared. Early work ....

R. A. Levinson. "UDS: A Universal Data Structure," 2nd International Conference on Conceptual Structures, Springer-Verlag, August, 1994.

....in the poset or the lengths of the inheritance paths. 3 A Relation Based Representation for CS Levinson s Universal Data Structure (UDS) combines the features and benefits of neural networks, semantic networks, relational databases, and conceptual structures in a single representational framework [7]. Its relation based representation of graphs is of particular relevance to the new work presented in this paper. In UDS, conceptual structures are transformed such that the relations and their adjacent concept types in the CS become nodes in the UDS graph. Edges in the UDS graph represent the ....

Robert Levinson. UDS: A universal data structure. In W. M. Tepfenhart, I. P. Dide, and J. F. Sowa, editors, Conceptual Structures: Theory and Practice, pages 230--250. Springer-Verlag, New York, 1994. Lecture Notes in AI 835.

....is equivalent to structured concepts with unary and binary predicates. Only one relation is allowed between any two parts. Negation can be expressed using special negated edges. 10 Multilevel objects, multinary relations, or background knowledge cannot be expressed, although recent research by Levinson (1994) corrects these problems. The central tenet of Levinson s work is that a taxonomy provides an efficient indexing mechanism for structured objects. This indexing mechanism can be used to answer three 9 In an attribute value hypothesis space such as that used by UNIMEM, there is a unique least ....

....The authors state that the remaining problem of this approach is that of setting up the modules [concepts] and network [concept taxonomy] automatically in a learning phase . Clearly, this remaining problem can now be addressed by the new generation of structured concept discovery systems. Levinson s (1994) Method V retrieval algorithm uses a very similar arrow caching technique. All of these methods are fundamentally different from Ellis (1991) technique, as the category theorem states that all injective mappings between parent and child concept must be cached, whereas Ellis approach only ....

Levinson, R. A. 1994. UDS: A universal data structure. Technical Report UCSC-CRL-94-15, University of California, Santa Cruz.

....logical contradictions) in a CG database. Some of the major contributing notions include: 1. The storing of the CG generalization hierarchy as a CNF lattice that exploits the duality between Boolean AND and OR for storage and retrieval efficiency. 2. The extension of previous implementation models [3, 18] that exploit containment links to avoid repeated storage of structures, to store nested contexts. 3. The association of labels (codes) with nodes and links that indicate whether the positive or negated version is being asserted or inferred. 4. The representation of a CG hierarchy as a hierarchy ....

....to the first. Section 3, Lattice Terminology, explains the basic terminology needed to understand and use lattices to store the CG database. Section 4, Adding Contexts, adds complex referents which include contexts. Previous work defining a Universal Data Structure (UDS) to store the CG database [18] did not include complex referents with icons, indexes, and symbols. This section shows how various forms of them, e.g. names, variables, literals, and graphs for contexts, can be added to the generalization lattice. Section 5 is on Adding Negation. It describes how to extend the generalization ....

[Article contains additional citation context not shown here]

R. Levinson. UDS: A Universal Data Structure. In Proceedings of Second International Conference on Conceptual Structures, ICCS'94, College Park, Maryland, USA, August 1994, pages 230-250.

No context found.

Levinson, R. A. 1994. UDS: A universal data structure. Technical Report UCSC-CRL-94-15, University of California, Santa Cruz.

No context found.

Robert A. Levinson. UDS: A universal data structure. In 2nd International Conference on Conceptual Structures, pages 230--250. Spring-Verlag, 1994.

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