Evans C., Negation-as-failure as an an approach to the Hanks and McDermott problem, Proc. of the 2nd International Symposium on Arti cial Intelligence, 1989.  Home/Search   Document Not in Database   Summary   Related Articles   Check

....the frame axiom correctly. A nonmonotonic reasoning technique that was mistakenly not considered by these authors is negation as failure. It has been shown that the YTS representation in situation calculus or event calculus with negation as failure solves the problem correctly ( 1] 4] [5]) Negation as failure alone is not sufficient for representing many temporal reasoning problems. A major restriction is its incapacity of representing incomplete knowledge. The original event calculus only supports the prediction of a goal state, starting from a complete description of the ....

C. Evans. Negation as failure as an approach to the Hanks and McDermott problem. In Proc. of the second International Symposium on Artificial Intelligence, 1989.

....Noninertial(loaded; load; S) may be dropped from this program without effect on the semantics of Holds 2. In general, all Noninertial=3 rules for initiating effects of actions may be dropped, without effect on the semantics of Holds 2. DY TS strongly resembles the YTS solution in [1] and in [14]. For example, 1] proposes a Prolog program analogous to D Y TS : Holds(alive; s 0 ) Holds(F; Result(A; S) Holds(F; S) Noninertial(F; A; S) Holds(loaded; Result(load; S) Noninertial(loaded; shoot; S) Noninertial(alive; shoot; S) Holds(loaded; S) Note that this program ....

....S) iff A is an action of registering St in C or, St was enrolled at S and A is not an action of dropping St from the course C. As said above, this type of theory shows a strong similarity with the completion of the open programs produced by and with the completions of the logic programs found in [14] and [1] If we forget about P oss(A; S) and introduce Result and Holds, we find: 8St; C; A; S : Holds(enrolled(St; C) Result(A; S) A = register(St; C) Holds(enrolled(St; C) S) A 6= drop(St; C) Similar formulas are subsumed by D. The first disjunct corresponds to a rule initiating ....

C. Evans. Negation as failure as an approach to the Hanks and McDermott problem. In Proc. of the second International Symposium on Artificial Intelligence, 1989.

....are dealt with in Apt [Apt95] The Yale Shooting Problem was extensively discussed in the literature and its formalizations in various formalisms for non monotonic reasoning were given. The solution in Prolog presented here was found independently by Apt and Bezem [AB91] Elkan [Elk89] and Evans [Eva89] where its declarative and procedural interpretation were also studied. Summary 317 The formal aspects of the term inspection facilities have been considered in Apt et al. AMP94] where procedural and declarative interpretation of logic programs with arithmetic and the term inspection ....

C. Evans. Negation-as-failure as an approach to the Hanks and McDermott problem. In Proceedings of the Second International Symposium on Artificial Intelligence, Monterrey, Mexico, 1989.

....results. In particular, there is an extension of the default theory in which the action of waiting unloads the gun. Morris [1988] showed that formalizing the frame default with the non normal default : Ab(f; a; s) Holds(f; s) j Holds(f; Result(a; s) 2.8) solves the Yale Shooting Problem. Evans [1989] and Apt and Bezem [1990] represent properties of actions in logic programming languages with negation as failure. Gelfond and Lifschitz [1993] point out the 1 A default of the form ff : w=w is called a normal default. 14 advantages of representing actions in extended logic programs (logic ....

Chris Evans. Negation-as-failure as an approach to the Hanks and McDermott problem. In Proc. of the Second Int'l Symp. on Artificial Intelligence, 1989.

....defeasible and abnormal effects. Our goal is to handle these three types of defeasibilities in reasoning about action under a framework of logic programming. The issue of representing action properties in logic programming languages is not new. It was explored by some researchers previously, e.g. [5, 6]. However, probably Gelfond 1 Informally, A ) B represents a semantics like if A then B, from which we cannot derive :B ) A. See [17] for detail. 4 and Lifschitz s work [9] was the first time to make major progress in this issue. By introducing a simple action language A, Gelfond and ....

Evans, C., Negation-as-failure as an approach to the Hanks and McDermott problem. In Proceedings of the Second International Symposium on Artificial Intelligence, 1989.

....default theory which has no facts but contains the two non normal defaults : M q p : M r q yields only one extension, containing q, which is the correct result. In contrast, all natural representations of the problem in Theorist give incorrect results. As Eshghi and Kowalski [32] Evans [34] and Apt and Bezem [3] observe, the Yale shooting problem has the form of a logic program, and interpreting negation in the problem as negation as failure yields only the correct result. This is the case for both the semantics and the proof theory of logic programming. Moreover, 32] and [62] show ....

Evans, C.A., Negation as failure as an approach to the Hanks and McDermott problem. Proc. 2nd International Symposium on Artificial intelligence, Monterrey, Mexico (1989) 68

....of the programs from [Tur94] DDS93] HT93] Dun93] Since these formalizations use different languages with different semantics, establishing this fact directly, without the use of A and AC , will be probably more difficult. Before A was introduced there were several works (which includes [AB90, EK89, Eva89]) that reasoned about actions using logic programs. In [AB90] Apt and Bezem consider the 16 case when there is complete information about the initial situation. If we consider only such domain descriptions then our f is an extension of Apt and Bezem s formalization. This can be easily shown ....

....Future research will show if this is a right assessment. The second theme of the paper is centered around the question of applicability of logic programming languages to representing knowledge about actions. The early attempts on such representation include Eshghi and Kowalski [EK89] Evans [Eva89], Apt and Bezem [AB90] among others. These formalizations used the language of general logic programs and therefore assumed the closed world assumption about all predicates. Formalization in the language of extended logic programs which has no such limitation was suggested in [GL92] The ....

C. Evans. Negation-as-failure as an approach to the Hanks and McDermott problem. In Proc. of the Second Int'l Symp. on Artificial Intelligence, 1989.

.... provide formalizations and implementations for special forms of non monotonic reasoning (see e.g. 2, 7] For example, the Prolog negation as finite failure operator has been used to implement a formulation as logic program of the temporal persistence problem in Artificial Intelligence (see [16, 13, 1]) Termination of glp s is a relevant topic (see [12] for a general discussion on termination in Logic Programming) also because the implementation of the operators for the negation, like Clark s negation as failure [10] and Chan s constructive negation [9] are based on termination conditions. ....

C. Evans. Negation as Failure as an Approach to the Hanks and McDermott Problem. Proc. of the 2nd International Symposium om AI, pp. 23-27, 1990.

....theory which has no facts but contains the two non normal defaults : M q p : M r q 12 yields only one extension, containing q, which is the correct result. In contrast, all natural representations of the problem in Theorist give incorrect results. As Eshghi and Kowalski [32] Evans [34] and Apt and Bezem [3] observe, the Yale shooting problem has the form of a logic program, and interpreting negation in the problem as negation as failure yields only the correct result. This is the case for both the semantics and the proof theory of logic programming. Moreover, 32] and [62] show ....

Evans, C.A., Negation as failure as an approach to the Hanks and McDermott problem. Proc. 2nd International Symposium on Articial intelligence, Monterrey, Mexico (1989)

....y 4 is not stratified it is possible to show (see Theorem 2.5) that it has a unique stable model. From this and Lemma 2. 4 it is easy to see that Y entails holds(alive; res(load; s 0 ) and :holds(alive; res(shoot; res(wait; res(load; s 0 ) 2 As we can see, the logic programming solution [Eva89, AB91, EK89] to the original Yale Shooting Problem is rather natural and simple. This is not, of course, to say that it can be easily generalized to more complicated forms of reasoning about actions. It is worth recalling that the original formulations of this story in the formalisms of circumscription and ....

C. Evans. Negation-as-failure as an approach to the Hanks and McDermott problem. In Proc. of the Second Int'l Symp. on Artificial Intelligence, 1989.

....in [26] to express negative information, and it has been later recognized as a suitable form of non monotonic reasoning. For instance, negation as failure provides a simple and elegant solution to the well known frame problem in AI, as exemplified by the so called Yale Shooting Problem [13, 31, 43]. More recently, other extensions have been proposed to further enrich the expressive power of LP in order to turn it into a general formalism for knowledge representation (for a survey, see [15] Among others, we mention abductive LP [42, 35, 48] and LP with different forms of negation (e.g. ....

C. Evans. Negation-as-Failure as an Approach to the Hanks and McDermott Problem. In Proc. 2nd Int. Symp. on Artificial Intelligence, Monterrey, Mexico, 1989.

.... concentration of such work has been in circumscription [Hau87, Lif87a, LR89, Lif90, Bak91, GLR91, Lif91, LS91, CE92, LR94, KL95, Lin95, Gus96, Lin96] There were also early published solutions to the Yale Shooting problem in autoepistemic logic [Gel88] default logic [Mor88] and logic programming [EK89, Eva89, AB90]. A primary methodological weakness of much of the work cited above is the fact that it is motivated by, and validated for, only a small set of examples. In many cases, there is no clear claim about what kinds of action domains can be represented correctly by a given approach. Moreover, it is ....

Chris Evans. Negation-as-failure as an approach to the Hanks and McDermott problem. In Proc. of the Second Int'l Symp. on Artificial Intelligence, 1989.

.... actions can be automated using SLG, a logic programming interpreter based on SLG resolution, developed by Chen and Warren [1993a,1993b] Formalizations of action based on the situation calculus and logic programming with negation as failure were first introduced by Eshghi and Kowalski [1989] and Evans [1989]. These logic programs are closely related to some formalizations that use default logic [ Morris, 1988 ] and autoepistemic logic [ Gelfond, 1989 ] Apt and Bezem [1990] observed that such programs are acyclic, which implies that the SLDNF procedure will terminate for them on all nonfloundering ....

Chris Evans. Negation-as-failure as an approach to the Hanks and McDermott problem. In Proc. of the Second Int'l Symp. on Artificial Intelligence, 1989.

....substantially influenced the logic programming approach to formalizing actions and changes. ffl Logic programming. Instead of using various circumscription policies, a logic program uses an operator called negation as failure (not) The YSP can be expressed as a logic program (see, for example, [20, 4]) The resulting logic program can be proved to be locally stratified and thus has a unique perfect model [57] as intended 1 . Because there are so many different approaches to reasoning about actions and changes, it is very difficult, if not impossible, to compare and contrast them with each ....

Evans, C., Negation-as-failure as an approach to the Hanks and McDermott problem, Proc. of the 2nd Int't Symp. on Artificial Intelligence, 1989

.... to logic programming we note three independent references that of Apt and Bezem [4] who proved that the translation of the Yale Shooting Problem to a logic program results in an acyclic program, Elkan [56] who showed that this translation results in a locally stratified program, and Evans [58], who observed that SLDNF resolution can be used to compute desired consequences of the original formulation of the problem in first order logic. In contrast, the program EVEN of Section 6.2 is not locally stratified, so a fortiori not acyclic. However, it is possible to apply to it a result of ....

C. Evans. Negation-as-failure as an approach to the Hanks and McDermott problem. In Proceedings of the Second International Symposium on Artificial Intelligence, Monterrey, Mexico, 1989.

....is applicable to temporal projection problems with incomplete information, as well as to reasoning about the past. It is proved to be sound relative to a semantics of action based on states and transition functions. 1 Introduction This paper extends the work of Eshghi and Kowalski [6] Evans [7] and Apt and Bezem [1] on representing properties of actions in logic programming languages with negation as failure. Our goal is to overcome some of the limitations of the earlier work. The existing formalizations of action in logic programming are adequate for only the simplest kind of temporal ....

Chris Evans. Negation-as-failure as an approach to the Hanks and McDermott problem. In Proc. of the Second Int'l Symp. on Artificial Intelligence, 1989.

....1. A restricted monotonicity theorem for Bakerstyle formalizations can be proved using Theorem 3 from [ Kartha, 1993 ] The logic programming method of [ Gelfond and Lifschitz, 1992 ] and [ Baral and Gelfond, 1993 ] builds on the ideas of [ Morris, 1988 ] Eshghi and Kowalski, 1989 ] Evans, 1989 ] and [ Apt and Bezem, 1990 ] A restricted monotonicity theorem for this method can be derived from Proposition 2. Sandewall [1989] proposed to apply a nonmonotonic formalism to a set of axioms that does not include initial conditions ( observations ) and to get first the set of all possible ....

Evans, Chris 1989. Negation-as-failure as an approach to the Hanks and McDermott problem. In Proc. of the Second Int'l Symp. on Artificial Intelligence.

....y 4 is not stratified it is possible to show (see Theorem 2.5) that it has a unique stable model. From this and Lemma 2. 4 it is easy to see that Y entails holds(alive; res(load; s 0 ) and :holds(alive; res(shoot; res(wait; res(load; s 0 ) 2 As we can see, the logic programming solution [Eva89, AB91, EK89] to the original Yale Shooting Problem is rather natural and simple. This is not, of course, to say that it can be easily generalized to more complicated forms of reasoning about actions. It is worth recalling that the original formulations of this story in the formalisms of circumscription and ....

C. Evans. Negation-as-failure as an approach to the Hanks and McDermott problem. In Proc. of the Second Int'l Symp. on Artificial Intelligence, 1989.

....of the programs from [Tur94] DDS93] HT93] Dun93] Since these formalizations use different languages with different semantics, establishing this fact directly, without the use of A and AC , will be probably more difficult. Before A was introduced there were several works (which includes [AB90, EK89, Eva89]) that reasoned about actions using logic programs. In [AB90] Apt and Bezem consider the case when there is complete information about the initial situation. If we consider only such domain descriptions then our f is an extension of Apt and Bezem s formalization. This can be easily shown ....

....Future research will show if this is a right assessment. The second theme of the paper is centered around the question of applicability of logic programming languages to representing knowledge about actions. The early attempts on such representation include Eshghi and Kowalski [EK89] Evans [Eva89], Apt and Bezem [AB90] among others. These formalizations used the language of general logic programs and therefore assumed the closed world assumption about all predicates. Formalization in the language of extended logic programs which has no such limitation was suggested in [GL92] and also in ....

C. Evans. Negation-as-failure as an approach to the Hanks and McDermott problem. In Proc. of the Second Int'l Symp. on Artificial Intelligence, 1989.

....example can be found in research on the frame problem. It is known that expressing temporal persistence by the formula Holds(f; s) Ab(f; a; s) oe Holds(f; Result(a; s) leads to difficulties [ Hanks and McDermott, 1987 ] This fact prompted several authors ( Eshghi and Kowalski, 1989 ] Evans, 1989 ] Apt and Bezem, 1990 ] to experiment with the corresponding logic programming rule Holds(f; Result(a; s) Holds(f; s) not Ab(f; a; s) 2) Independently, Morris [1988] proposed to express the same principle of reasoning by the default Holds(f; s) Ab(f; a; s) Holds(f; Result(a; ....

Chris Evans. Negation-as-failure as an approach to the Hanks and McDermott problem. In Proc. of the Second Int'l Symp. on Artificial Intelligence, 1989.

.... voor Programmatie van Wetenschapsbeleid , under the contract RFO AI 03, and by ESPRIT BRA Compulog II under contract 6810 y supported by the Belgian National Fund for Scientific Research been shown that the YTS representation in these formalisms solves the problem correctly [1] 14] [15]. Negation as failure alone is not sufficient for representing many temporal reasoning problems. A major restriction is its inability of representing incomplete knowledge. The original event calculus only supports the prediction of a goal state, starting from a complete description of the initial ....

C. Evans. Negation as failure as an approach to the Hanks and McDermott problem. In Proc. of the second International Symposium on Artificial Intelligence, 1989.

....proposals of several kinds, differing by the choice of the underlying logic. Some of them combine the situation calculus with classical logic (for instance, Schubert, 1990 ] Reiter, 1991 ] Elkan, 1992 ] some with logic programming (for instance, Eshghi and Kowalski, 1989 ] Evans, 1989 ] Apt and Bezem, 1990 ] some with circumscription (for instance, Haugh, 1987 ] Lifschitz, 1987 ] Baker, 1989 ] or with other nonmonotonic formalisms (for instance, Morris, 1988 ] and [ Gelfond, 1989 ] Some of the more recent work on representing action makes use of ....

Chris Evans. Negation-as-failure as an approach to the Hanks and McDermott problem. In Proc. of the Second Int'l Symp. on Artificial Intelligence, 1989.

....has two guns, and at least one of them is initially loaded. The turkey is initially alive. The available actions are loading and shooting. We are interested in representing this action domain using the situation calculus and negation as failure. Early work of this kind is described in [EK89] [Eva89], AB90] and [Gel91] The subject is treated more systematically in several recent papers, including [GL93] DDS93] and [Dun93] We assume that the reader is familiar with the main ideas of this work. The following formulation of the two gun example is close to the one proposed in [Tur94] It ....

Chris Evans. Negation-as-failure as an approach to the Hanks and McDermott problem. In Proc. of the Second Int'l Symp. on Artificial Intelligence, 1989.

.... special forms of non monotonic reasoning, as illustrated by Apt and Bol (1994) and Baral and Gelfond (1994) For example, Prolog s negation as finite failure operator can be used to implement the temporal persistence problem in Artificial Intelligence as a logic program (Kowalski Sergot, 1986; Evans, 1990; Apt Bezem, 1991) The implementation of operators like Clark s negation as failure (Clark, 1978) and Chan s constructive negation (Chan, 1988) is based on termination conditions. Therefore the study of termination of glp s (e.g. De Schreye Decorte, 1994) is an important topic. Two classes ....

Evans, C. (1990). Negation as failure as an approach to the hanks and mcdermott problem.

No context found.

Evans C., Negation-as-failure as an an approach to the Hanks and McDermott problem, Proc. of the 2nd International Symposium on Arti cial Intelligence, 1989.

First 50 documents

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