Terry Gaasterland, Parke Godfrey, and Jack Minker. Relaxation as a platform for cooperative answering. Journal of Intelligent Information Systems, 1(3/4):293--321, 1992.  Home/Search   Document Details and Download   Summary   Related Articles   Check

....number of users. The Niagara Internet Query System [19] suggests an interface where sources in a query are automatically generated by looking up user specified keywords in a catalog of all sources. The cooperative database literature has long argued for looser modes of query specification (e.g. [28, 18, 7]) where the system aids the user in specifying queries. For example, the CoBase project [28] automatically adds other relevant sources to a query even if the user has not asked for them. A common denominator among these approaches is a broad, generalist approach to query specification. Such query ....

....and generate a parachute query to complete the original query when all sources are available. But they do not give out query results incrementally, and do not adapt to dynamic volatilities or user preference changes. There is also a rich literature on looser modes of query specification (e.g. [28, 18, 7]) where the system automatically relaxes query constraints, or adds additional sources, to provide expanded answers. As discussed in Section 1,the techniques of this paper are especially appropriate for such systems. Algorithmically, our work is related to the recent body of work on adaptive ....

T. Gaasterland et al. Relaxation as a platform for cooperative answering. J. Intel. Info. Sys., 1(3/4), 1992.

....formalisms and implemented systems. Most work concerning metaprogramming in logic takes its origin in the extremely simple self interpreter for Prolog known as Vanilla. prove(true) prove( A,B) prove(A) prove(B) prove(A) clause(A,B) prove(B) Gaasterland, Godfrey, and Minker [17] has used an extension of Vanilla for describing cooperative answering based on relaxation by taxonomy, i.e, generalizing the query in case of an insufficient answer. In our own work [2; 3] we have taken this approach further allowing (in principle) arbitrary nonstandard inference rules kept in ....

....travels in which one or more subdistances is replaced by another means of transportation, e.g. using another taxonomy rule in the normal modus ponens direction. klause(tax, subdist(X,Y) bus ride(X,Y) Our approach, here, is inspired by earlier work of Gaasterland, Godfrey, and Minker [17] who performed similar transformations on the initial query, but without going into derived subgoals as we do. We can illustrate the difference between the two by an example. Having submitted a query for a flight travel from Copenhagen to Budapest, the typical traveller may accept as an answer ....

Gaasterland T., Godfrey P., and Minker J., Relaxation as a Platform for Cooperative Answering. Journal of Intelligent Information Systems, 1, 3/4, pp. 293-321, 1992.

....in human discourse. For instance, if a customer asks a travel agent asking the flight from the city X to the city Y on the date of Z, then when no satisfactory flights are found, the agent is very likely to respond with alternatives by changing one or many of X, Y , and Z of the initial query [GGM90]. In general, query modification broadly describes the process of changing a query when the answer to the query does not meet the expectations of the user. In the relational database jargon, this change to a query can occur in two di#erent places con21 dition (i.e. FROM and WHERE clauses) ....

....size, location) and contents matching [CHC98] logistic planning application for searching objects with similar characteristics in the neighboring desired locations [CYC94] etc. Cooperativeness of the information system with the focus on the integrity constraints was studied in [Gal88] In [GGM90, Gaa97], Gaasterland et al. introduced query relaxation as a platform of cooperative answering from the perspective of deductive database and logic programming. Godfrey [God97] presents the theoretical complexity results of minimal failing sub query and maximal succeeding sub query problems, which are ....

T. Gaasterland, P. Godfrey, and J. Minker. "Relaxation as a Platform for Cooperative Answering". J. Intelligent Information Systems (JIIS), 1(3/4):293--392, 1990.

....gone by If so, where This is reasonable because in normal conversation the second part of Q2 will be asked in almost all cases. This is also a more generalized form of the original question. 13 [GGMN92] gives a more detailed survey of the above concepts. 4. 2 Relaxation Gaasterland et al.[GGM92b] provide various examples of relaxing the constraints in the query for a more cooperative answer. They provide a cooperative method of relaxation for expanding deductive database and logic programming queries. The relaxation method expands the scope of a query by relaxing the constraints implicit ....

....the taxonomy clause. b.Replace the query atom with the head atom of the taxonomy clause. c.Add the non key literals from the body of the taxonomy clause to the new query as constraints on the variables. Using this process the predicates can be relaxed. The details and examples can be seen in [GGM92b]. 4.2.2 Partitioning In gure 4.1 we see how intermediate types may partition a type along a number of dimensions. The types dalas co, kawasaki co etc. partition companies by their locations. The types chem co, electronics co, etc. partition companies by their industries. And there will be many ....

Terry Gaasterland, Parke Godfrey, and Jack Minker. Relaxation as a platform for cooperative answering. volume 1, pages 293-321, 1992.

....every rule making use of the predicate p k Gamma1 can be deleted. 5 Application to Intelligent Query Answering The framework of semantic query optimization is sufficiently broad to capture forms of intelligent query answering. Intelligent query answering has been studied by several researchers [4, 5, 11] In this section, we shall show via examples that some of the techniques studied by Motro and Yuan [11] can be captured within our framework for semantic query optimization. The basic idea in intelligent query answering is that it must be possible to elicit answers from the database that ....

T. Gaasterland, P. Godfrey and J. Minker, "Relaxation as a Platform for Cooperative Answering, " Journal of Intelligent Information Systems, 1, (1992), 293-321.

....into specific query conditions by specialization. Query modification may also be specified explicitly by the user through a set of cooperative operators such as similar to , approximate , near to , etc. This approach was also adopted in the cooperative deductive databases by Gaasterland, et al. [15] for providing query relaxations. The notion of multi level object representation is not captured by the conventional semantic network and object oriented database approaches for the following reasons: grouping objects into a class and grouping several classes into a super class only provide a ....

T. Gaasterland, P. Godfrey, and J. Minker. Relaxation as a platform of cooperative answering. Journal of Intelligent Information Systems, 1:293--321, 1992.

....) Q(x; Y ) as the variable symbols x and y denote the values in a database. We call these conjunctions model schemas. A model schema determines true interpretation of rules and constraints as they apply to a database. The idea is that rules are harmonized by removing inconsistent models (e.g. [4,5] ) To resolve inconsistencies, the following theorem is proposed. Method 1 (Knowledge Harmonization) Let 4 = fK; X; Qg be a closed theory of the possible world containing the existing rules K and the discovered rules X which correspond to a query Q. Let E(K) and E(X) be the models of K and X, ....

....predicates will be our future research. Moreover, the rules discovered in this paper can be represented in a concise and elegant form by generalization techniques. Discovered rules may be refined [8,10] as a database is evolved. The approach may also be useful for research in cooperative answers [14,5] in which the system responds with knowledge rather than simply providing the tuples satisfying the query. ....

Terry Gaasterland, Parke Godfrey, and Jack Minker. Relaxation as a platform for cooperative answering. Journal of Intelligent Information Systems, 1(3-4):293--321, 1992.

.... organization [25, 39] clustering and browsing [5, 34] case [11, 13] and analogy based retrieval [30] and hybrid organizations [4] Similarity based retrieval helps to access relevant artifacts when the query is not complete and precise by iteratively relaxing or restricting the initial query [15, 34]. This becomes useful when a repository is large, and created by many users over a longer time period. Generally, it is impossible to decide in advance what needs to be retrieved, when, and in what form. Flexible and iterative similarity based retrieval diminishes this problem [26] CAReT is an ....

T. Gaasterland, P. Godfrey and J. Minker. Relaxation as a platform of cooperative answering. In. Proc. International Workshop on Nonstandard Answers and Queries. Toulouse, France, 1991.

....in intelligent query answering. 7. 1 Introduction Cooperative (or intelligent) query answering refers to a mechanism which answers information system queries cooperatively and intelligently by analyzing the intent of a query and providing some generalized, neighborhood, or associated answers [20, 24, 39, 8]. Many interesting techniques [58, 25, 19, 38, 63] have been developed for cooperative query answering, by integration of the methods developed in several related fields, such as semantic data modeling, deductive databases, knowledge discovery in databases, etc. In this chapter, we propose a new ....

....25, 19, 38, 20] can be extended to cooperative query answering in MLDBs, easily, effectively and efficiently. The following reasoning may convince us that an MLDB can greatly facilitate cooperative query answering. 1. Many cooperative query answering techniques need certain kinds of generalization [20, 39]; whereas different kinds of frequently used generalizations are performed and stored in the higher layers of an MLDB. CHAPTER 7. MULTIPLE LAYERED DATABASES 165 2. Many cooperative query answering techniques need to compare the neighborhood information [25, 20] whereas the generalized ....

T. Gaasterland, P. Godgrey, and J. Minker. Relaxation as a platform for cooperative answering. Journal of Intelligent Information Systems, 1:293--321, 1992.

.... Approach Although knowledge about the application domain can be expressed as a set of logical rules, such a rule based approach lacks a systematic organization to guide the query transformation process [10, 13] To remedy this problem, we use the notion of a type abstraction hierarchy (TAH) [4, 6, 7, 12] for providing an efficient and organized framework for cooperative query processing. Type abstraction emphasizes the abstract representation of object instances (see Figure 1) For example, runway length range in the domain of range (e.g. 4,000 to 8,000 ft. is a more abstract representation ....

T. Gaasterland, P. Godfrey, and J. Minker. Relaxation as a platform of cooperative answering. In Proceedings First International Workshop on Nonstandard Queries and Answers, Toulouse, France, 1991.

....10am, 6pm, #076) and DELTA flight(LAX, National, 11am, 5pm, #024) match each other approximately via CC flight(Los Angeles, Washington, morning, afternoon) 3. 2 Query Pattern Relaxation The proposed cooperative query answering mechanism can also be viewed as a query pattern relaxation mechanism [12]. To show this we introduce the notions of pattern and pattern instance. A pattern is defined on a type by specifying a condition [3] For example, given the following schema definition, AA flight(departure airport, arrival airport, departure time, arrival time, flight#) conditions such as ....

T. Gaasterland, P. Godfrey, and J. Minker. Relaxation as a platform for cooperative answering. Journal of Intelligent Information Systems, 1(3/4):293--321, 1992.

....potential incorrect presuppositions (like CS400 was offered last term) and a synthesizer which provided a natural language response. Part of the response generator included a set of heuristics used to select pertinent information for inclusion in the cooperative answer. Gaasterland, et.al. [54, 55, 56] expand on Gal s work in a number of ways still with the goal of providing more cooperative explanations. They, as most in the cooperative answering area, take inspiration from Grice [60] and aim to follow his maxims of cooperative conversation. One area they tackle is correcting invalid user ....

T. Gaasterland, P. Godfrey, and J. Minker. Relaxation as a platform for cooperative answering. In Journal of Intelligent Information Systems, 1:293-321, 1992.

....(single layer) databases [17, 3, 2, 10, 16] can be extended effectively to cooperative query answering in MLDBs. The following reasoning may convince us that an MLDB may greatly facilitate cooperative query answering. Many cooperative query answering techniques need certain kinds of generalization [2, 11]; whereas different kinds of frequently used generalizations are performed and stored in the high layers of an MLDB. Also, they often need to compare the neighborhood information [3, 2] The generalized neighborhood tuples are usually stored in the same high layer relations, ready for comparison ....

T. Gaasterland, P. Godgrey, and J. Minker. Relaxation as a platform for cooperative answering. Journal of Intelligent Information Systems, 1:293--321, 1992.

....of attributes required to match: given m of n matching, the required number of attributes is reduced from m to p, where 0 p m n. Generalization is a context transformation that relaxes the context by enlarging the set of allowable values for an attribute. This technique has been used in [10, 15] for relaxation of database queries. Consider the example presented in Table 3. Applying reduction to the first category, attributes motor and screw, changes the requirement that both attributes must satisfy the constraint to the requirement that if either of the two attributes satisfies the ....

T. Gaasterland, P. Godfrey, and J. Minker. Relaxation as a platform of cooperative answering. In Proc. International Workshop on Nonstandard Answers and Queries, Toulouse, France, 1991.

....query answering in database systems. 1 Introduction Cooperative (or intelligent) query answering refers to a mechanism which answers queries cooperatively and intelligently by analyzing the intent of a query and providing some generalized, neighborhood, or associated answers [5, 11, 2]. Many interesting techniques [14, 6, 5, 10, 13, 15] have been developed for cooperative query answering, by integration of the methods developed in several related fields, such as semantic data modeling [3] Research partially supported by the Natural Sciences and Engineering Research Council ....

....[14, 6, 5, 10, 13] can be extended to cooperative query answering in MLDBs, easily, effectively and efficiently. The following reasoning may convince us that an MLDB may greatly facilitate cooperative query answering. Many cooperative query answering techniques need certain kinds of generalization [5, 11]; whereas different kinds of frequently used generalizations are performed and stored in the higher layers of an MLDB. Also, they often need to compare the neighborhood information [6, 5] The generalized neighborhood tuples are usually stored in the same higher layer relations, ready for ....

T. Gaasterland, P. Godgrey, and J. Minker. Relaxation as a platform for cooperative answering. Journal of Intelligent Information Systems, 1:293-- 321, 1992.

....or how much to relax a condition, must be determined too. This type of information is not available with databases, and somehow the information has to be manifested. Motro [37] introduces a supposition generalizer into the architecture of his system SEAVE, which decides how to relax atoms. In [16], we consider how to use taxonomic information represented inherently in the rules of a deductive database to relax atoms in queries. Chu et al. 5, 6] consider how to extract and employ such taxonomic information in relational databases, and have implemented such in their system CoBase. ....

T. Gaasterland, P. Godfrey, and J. Minker. Relaxation as a platform for cooperative answering. Journal of Intelligent Information Systems, 1:293--321, 1992.

....providing associated information which is relevant to a query. Generalizing a query in order to capture neighboring information is one means to obtain possibly relevant information. Gaasterland, Godfrey and Minker have defined a method to relax a query in order to find neighboring information. [9] A query can be relaxed in at least three ways: 1) rewriting a predicate with a more general predicate; 2) rewriting a constant (term) with a more general constant (term) and 3) breaking a join dependency across literals in the query. The first two relaxations are achieved in a general manner ....

....the address 34 Cherry Lane, upon failing, can be relaxed to produce alternative queries: Q: mail(Terry, 34 Cherry Lane , invitation) Relaxed Q: communicate(Terry, P, invitation) lives at( 34 Cherry Lane , P) After the relaxation step, SLD resolution can be used to find related answers. [9] discusses search strategies and the use of user constraints for controlling relaxation. The method has been incorporated into the cooperative answering system. 4 Presenting the Logic Formula in Natural Language The cooperative answering strategies discussed in the previous section produce ....

T. Gaasterland, P. Godfrey, and J. Minker. Relaxation as a platform for cooperative answering. In Proceedings of Workshop on Nonstandard Queries and Answers, Toulouse, France, July, 1991.

....following language generation tasks: ffl Selection of coordination sites in the target language. ffl Selection of subordinate clauses and pronominalization. ffl Selection of tense, aspect, and temporal connecting words. ffl Incorporation of synonyms into the target text via taxonomy clauses [GGM91] 7 Summary and Future Work Semantic information in a deductive database is a relatively untapped source of information that can be used in responses to users queries. Cooperative answers that use semantic information such as integrity constraints and user constraints can provide the user with ....

T. Gaasterland, P. Godfrey, and J. Minker. Relaxation as a platform for cooperative answering. In Proceedings of Workshop on Nonstandard Queries and Answers, Toulouse, France, July, 1991.

....terms from the query with corresponding relations and terms from higher in the hierarchy. The resulting query is considered more general than the original. Along similar lines, we have introduced a cooperative method called relaxation for expanding deductive database and logic programming queries [20]. The relaxation method expands the scope of a query by relaxing the logical constraints implicit in the query. Thus, the database may return answers related to the original query as well as the literal answers themselves. travel (From, To) serves area (A, From) serves area (B, To) flight (A, ....

....with a in the body) meaning it can be used in an abductive direction to relax a query to a more general query, for instance Q 8 to Q 9 . The marked atom, flight (A, B) in C 6 , unifies an atom in the query and is replaced with the head of the clause and the other atoms in the body. See [20]. Our relaxation method is a general approach to seek additional answers to a query that may or may not be of direct interest to the user. We plan to extend this work to consider user models and user preferences to help determine when a query relaxation might be relevant. In the system FLEX, ....

[Article contains additional citation context not shown here]

T. Gaasterland, P. Godfrey, and J. Minker. Relaxation as a Platform of Cooperative Answering. In T. Imielinski, editor, Proceedings of the First International Workshop on Nonstandard Queries and Answers, pages 101--120, volume 2, Toulouse, France, July 1991.

....on integrity constraints directly. Query relaxation helps the user formulate follow up queries by relaxing the current query into more general queries. A more general query s answer set subsumes the query s answer set. This helps the user to browse further for related answers and concepts. In [Gaasterland et al. 1992b] the rules of the database are employed in an abductive direction to accomplish relaxation. In the CoBase system ( Chu et al. 1993] and [Chu et al. 1994] type abstraction hierarchies are used in conjunction with the relational schema to relax constants used in the query, and a ....

Gaasterland, T., Godfrey, P., and Minker, J. (1992b). Relaxation as a platform for cooperative answering. Journal of Intelligent Information Systems, 1:293--321.

....data itself and how it is organized to ask accurate queries. Sometimes queries contain misconceptions and are bound to fail. Sometimes they are too general. Other times they are too specific. When users ask queries that do not obtain the information that they seek, a technique called relaxation [ 5, 7 ] enables a knowledge based system in the form of a deductive database to work interactively with a user to find alternative answers that are related to the answers of the original query. The relationships are defined through taxonomy clauses, which are first order expressions about semantic ....

....including [ 1, 3, 4, 12, 13, 15, 16, 18 ] one form of cooperative behavior involves providing associated information that is relevant to a query. Generalizing a query in order to capture neighboring information is a means to obtain possibly relevant information. Gaasterland, Godfrey, and Minker [ 7 ] have defined a method to relax a query in order to find neighboring information. A query can be relaxed in at least three ways: 1) rewriting a predicate with a more general predicate; 2) rewriting a constant (term) with a more general constant (term) and (3) breaking a join dependency across ....

[Article contains additional citation context not shown here]

T. Gaasterland, P. Godfrey, and J. Minker. Relaxation as a platform for cooperative answering. Journal of Intelligent Information Systems, 1993. To appear.

....providing associated information which is relevant to a query. Generalizing a query in order to capture neighboring information is one means to obtain possibly relevant information. Gaasterland, Godfrey and Minker have defined a method to relax a query in order to find neighboring information [12]. A query can be relaxed in at least three ways: 1) rewriting a predicate with a more general predicate; 2) rewriting a constant (term) with a more general constant (term) and 3) breaking a join dependency across literals in the query. The first two relaxations are achieved in a general manner ....

....the address 34 Cherry Lane, upon failing, can be relaxed to produce alternative queries: Q: mail(Terry, 34 Cherry Lane , invitation) Relaxed Q: communicate(Terry, P, invitation) lives at( 34 Cherry Lane , P) After the relaxation step, SLD resolution can be used to find related answers. [12] discusses search strategies for relaxation. The method has been incorporated into the cooperative answering system. 4 On Presenting the Cooperative Response in Natural Language The cooperative answering strategies discussed in the previous section produce potentially large and complex logic ....

T. Gaasterland, P. Godfrey, and J. Minker. Relaxation as a Platform for Cooperative Answering. Journal of Intelligent Information Systems, 1992. To appear.

....or how much to relax a condition, must be determined too. This type of information is not available with databases, and somehow the information has to be manifested. Motro [31] introduces a supposition generalizor into the architecture of his system SEAVE, which decides how to relax atoms. In [13], we consider how to use taxonomic information represented inherently in the rules of a deductive database to relax atoms in queries. Chu et al. 3, 4] consider how to extract and employ such taxonomic information in relational database, and have implemented such in their system CoBase. Heuristics ....

T. Gaasterland, P. Godfrey, and J. Minker. Relaxation as a platform for cooperative answering. Journal of Intelligent Information Systems, 1:293--321, 1992.

....a misconception in a query is the ultimate optimization; in such a case, the query does not have to be evaluated at all. June 1994 An Architecture for a CDBS GNM 94 p. 6 of 22 As work proceeded with cooperative answering, new cooperative techniques were found that should be supported. In [16], Gaasterland, Godfrey, and Minker present a mechanism to relax a given query into a set of more general, related queries. Evaluating one of these queries results in new answers with respect to the original query, and may be of interest to the user. In [14, 19, 20] Gaasterland, Minker, and ....

....Godfrey, and Minker in the survey [15] review a number of other cooperative answering techniques that warrant consideration. With the Carmin I prototype, Gaasterland, Godfrey, Minker, and Novik aimed to combine a number of fundamental cooperative techniques uniformly into one cooperative system [16, 18]. Carmin I is written in QUINTUS Prolog and exists entirely in the Prolog session. The system was meant as a testbed for cooperative techniques, to explore how they could be combined effectively and how they could be implemented. Work on the Carmin II prototype commenced in spring of 1993, and ....

[Article contains additional citation context not shown here]

T. Gaasterland, P. Godfrey, and J. Minker. Relaxation as a platform for cooperative answering. Journal of Intelligent Information Systems, 1:293--321, 1992.

No context found.

Terry Gaasterland, Parke Godfrey, and Jack Minker. Relaxation as a platform for cooperative answering. Journal of Intelligent Information Systems, 1(3/4):293--321, 1992.

No context found.

T. Gaasterland, P. Godgrey, and J. Minker. Relaxation as a platform for cooperative answering. Journal of Intelliilent Information Systems, 1:293-321, 1992.

No context found.

T. Gaasterland, P. Godfrey, and J. Minker. Relaxation as a platform for cooperative answering. Journal of Intelligenet Information Systems, 1(3,4):293--321, Dec 1992.

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