S. Hanks and D. McDermott. Default reasoning, nonmonotonic logics, and the frame problem. In Proc. 5th Nat. Conf. on Artificial Intelligence (AAAI'86), pages 328--333, Philadelphia, 1986. Morgan Kaufmann Publishers.  Home/Search   Document Not in Database   Summary   Related Articles   Check

....actions, ii) non determinism is not treated adequately, iii) no partial knowledge and no temporal post diction. However there are some successful attempts in order to overcome these problems proving the great exibility of this formalism [KM97, Sha97] 1. 2 Features Fluents The frame problem [HM86] is a typical problem in AI in the eld of commonsense reasoning. Given the description of a world and of an agent in it, a scenario description , we try to infer intuitive conclusions about all the properties that remain unchanged when actions are performed. Features Fluent is a logical ....

Steve Hanks and Drew McDermott. Default Reasoning, Nonmonotonic logics, and the Frame Problem. In Proceedings of the Fifth National (U.S.) Conference on Articial Intelligence, pages 328333, Los Altos, CA, 1986. Morgan Kauman.

.... ] so) Notice that bts[ is instantiated at the final state do(T, So) and this is the only state mentioned in the formula. So, we can see that we have transformed our problem of answering a temporal query wrt to a virtually updated database into the temporal projection problem of AI [4], that is, the problem of querying a future state obtained by the execution of an actions sequence. To solve this problem we may apply some existing techniques for Reiter like specifications, e.g. Reiter s query regression [9] minimal rolling forward of the database based on information that is ....

S. Hanks and D. McDermott. Default Reasoning, Nonmonotonic Logics, and the Frame Problem. In Proc. National Conference on Artificail Intelligence, pages 328-333, 1986.

....logic ofplans miniraises over several dimensions, including, e.g. the number of steps in a plan. 96 The whole area of non monotonic reasoning is in a state of great flux, in part because of the diversity of systems and the technical difficulty of arguments within and between the various systems [HM]. Nevertheless, the simple fact that no. other approach comes close to dealing with so many of the relevant problems suggests to me that three questions should be investigated on the way to a semantic theory of discourse: whether other pragmatic problems can be couched in the same terms (e.g. ....

Hanks, S., McDermott, D., Default reasoning, nonmonotonic logics, and the frame problem, AAAI-86, 1986.

....would not be initiated after all) The e#ect on spill thus depends on the absence of an e#ect on up r . Applying circumscription to rules incorporating such negative dependencies runs into the same sort of problems as the (naive) application of circumscription to the Yale Turkey Shooting problem ([9]) Due to the absence of syntactical constraints on e#ect rules, theories of e#ect rules can be constructed which cannot be given a consistent semantics. A simple example is the e#ect rule initiating a causes if true. In other cases, contradictions may appear in more subtle ways. A ....

S. Hanks and D. McDermott. Default reasoning, nonmonotonic logic, and the frame problem. In Proceedings of the National Conference on Artificial Intelligence, Philadelphia, pages 328--333, 1986.

....with respect to a given action: ab(action 1 ; object 1 ) The frame problem: the system cannot infer automatically that all the properties not modi ed by a given action should hold in the resulting situation. The di culty of this problem can be illustrated by the Yale Shooting Problem [18]. However, Sandewall proposed with Feature Fluents [41] a framework to deal with the frame problem. The rami cation problem: the system cannot infer automatically all the indirect changes after an action is performed. In order to deal with some rami cation problems, Gelfond and Lifschitz ....

.... put[path[speed] plus[get[speed] 1] This view is provided with a count process, responsible for increasing the value when Coco is running (this de nes the proactive behaviour) and four reactions (stop, start, speed up and slow down) The second example is the Yale Shooting Problem [18]: view[dead[false] loaded[false] load[ event[load] put[path[loaded] true] shoot[ event[shoot] guard[get[loaded] put[path[loaded] true] wait[ event[wait] 2.3.9 Properties In my thesis [36] I prove that: VDL is a Turing complete language (any Turing machine can be modelled ....

S. HANKS and D.V. McDERMOTT. Default reasoning, nonmonotonic logics, and the frame problem. In Proc. AAAI'86, pages 328333, 1986.

....derive successor state axioms for these views, since their contents at each state depends on their contents in the previous state and the transaction executed to reach the current state. It will turn out that Chomicki s task, in our setting, corresponds to solving the temporal projection problem [9] for a query stated in terms of the R ff s at the final state in the transaction list, that is, the problem of querying a virtual state obtained after the (hypothetical) execution of the whole transaction log. We can say that while Chomicki answers the query by positioning himself at the final ....

....we have transformed our problem of answering a temporal (dynamic) query wrt to a virtually updated database into the temporal projection problem. In artificial intelligence planning, this is the problem of querying a state of the world that is reached by execution of a sequence of legal actions [9]. Example 12. Let us consider the specification Sigma given in section 4. Suppose that in this database we want to know if is there anyone that has always been working in the company, in all states generated by the execution of the sequences of actions T = hire(Sue) fire(John) at S 0 , that ....

S. Hanks and D. McDermott. Default Reasoning, Nonmonotonic Logics, and the Frame Problem. In Proc. National Conference on Artificail Intelligence, pages 328--333, 1986.

....n ] S 0 ) Notice that bts [ is instantiated at the final state do(T ; S 0 ) and this is the only state mentioned in the formula. So, we can see that we have transformed our problem of answering a temporal query wrt to a virtually updated database into the temporal projection problem of AI [4], that is, the problem of querying a future state obtained by the execution of an actions sequence. To solve this problem we may apply some existing techniques for Reiter like specifications, e.g. Reiter s query regression [9] minimal rolling forward of the database based on information that is ....

S. Hanks and D. McDermott. Default Reasoning, Nonmonotonic Logics, and the Frame Problem. In Proc. National Conference on Artificail Intelligence, pages 328--333, 1986.

....that they eventually believed in the need to state frame axioms if they want to use deductive logic. For example, Patrick J. Hayes stated in [17] that And yet such laws : are necessary in first order axiomatizations. They are called frame axioms. Or, Steve Hanks and Drew McDermott expressed in [16] that Deductive logic forces us into the position of assuming an event [action] occurrence may potentially change the truth value of all facts [fluents] thus if it does not change the value of a particular fact in a particular situation we must explicitly say so. Interestingly, the idea for a ....

S. Hanks and D. McDermott. Default reasoning, nonmonotonic logics, and the frame problem. In Proceedings of the AAAI National Conference on Artificial Intelligence, pages 328--333, 1986.

....task. Following McCarthy [23] again, we intend to provide a system which meets the subsequent specifications. 1. General properties of causality and facts about the possibility and results of actions are given as axioms. This problem is similar but not equivalent to the Yale Shooting Problem [14]. The Yale Shooting Problem addresses not only the frame problem but also the qualification problem and we want to isolate the frame problem in this paper. Nevertheless, if we formalize our problem and apply circumscription in analogy to [14] then we also obtain two models as in the Yale Shooting ....

....similar but not equivalent to the Yale Shooting Problem [14] The Yale Shooting Problem addresses not only the frame problem but also the qualification problem and we want to isolate the frame problem in this paper. Nevertheless, if we formalize our problem and apply circumscription in analogy to [14], then we also obtain two models as in the Yale Shooting Problem. One could rephrase the example in a non monotonic reasoning framework but this would be a complication. I would like to keep things simple and just consider the frame problem in this paper. I will not discuss philosophical ....

[Article contains additional citation context not shown here]

S. Hanks and D. McDermott. Default reasoning, nonmonotonic logics, and the frame problem. In Proceedings of the AAAI National Conference on Artificial Intelligence, pages 328--333, 1986.

....features, this program illustrates how, in the GOLOG framework, one can represent concurrent processes with explicit time. 1 Introduction The situation calculus [15] has long been the formalism of choice in artificial intelligence for theoretical investigations of properties of actions (e.g. [12, 1, 4]) but until recently, it has not been taken seriously as a specification or implementation language for practical problems in dynamic world modeling. Exceptions to this are the situation calculus based programming languages GOLOG and CONGOLOG [11, 2] and some of their applications to planning ....

S. Hanks and D. McDermott. Default reasoning, nonmonotonic logics, and the frame problem. In Proceedings of the National Conference on Artificial Intelligence (AAAI'86), pages 328--333, 1986.

....zeitgeist almost from the very beginning of the field. It is included in the standard material of every introductory course on AI, and it is the language of choice for investigations of various technical problems that arise in theorizing about actions and their effects (e.g. 19] 8] [13]) But only recently has it been taken seriously as a foundation for practical work in planning, control, simulation, database updates, agent programming and robotics (e.g. 18, 9, 17, 34, 33, 2, 14, 16] In parallel with these developments of its applications, there have emerged axiomatizations ....

S. Hanks and D. McDermott. Default reasoning, nonmonotonic logics, and the frame problem. In Proceedings of the National Conference on Artificial Intelligence (AAAI'86), pages 328--333, 1986.

....addresses, telephone numbers, fees, etc. etc. has been executed , we must determine whether Database j= 9c) enrolled(John; c; do(register(Mary; C100) do(drop(John; C100) S 0 ) Querying an evolving database is precisely what is called the temporal projection problem in AI planning [18]. 4. AN AXIOMATIZATION OF UPDATE TRANSACTIONS The example education domain illustrates the general principles behind our approach to the specification of database update transactions. In this section we precisely characterize a class of databases and updates of which the above example will be an ....

.... and if so, what is the answer to Q in that state of the database that would result from performing these transactions in the indicated sequence, beginning with the initial state S 0 of the database This is exactly what is called the temporal projection problem in the AI planning literature ([18]) For the class of databases of this paper, Reiter [43] has provided a closed form solution to this problem, which we now describe. 5.1. Legal Transaction Sequences In this section we provide necessary and sufficient conditions that a sequence 1 ; n of update transactions be legal. ....

S. Hanks and D. McDermott. Default reasoning, nonmonotonic logics, and the frame problem. In Proceedings of the National Conference on Artificial Intelligence, pages 328--333, 1986.

....partial plans; the order of events is left unspecified when they do not interfere. 1 Introduction Temporal reasoning is an excellent domain for testing nonmonotonic reasoning techniques because of the frame problem : the problem of representing the tendency of facts to endure over time ([6]) Hanks and McDermott used the famous Yale turkey shooting problem (YTS) to show that well known nonmonotonic reasoning systems failed to represent the frame axiom correctly. A nonmonotonic reasoning technique that was mistakenly not considered by these authors is negation as failure. It has ....

S. Hanks and D. McDermott. Default reasoning, nonmonotonic logic, and the frame problem. In Proc. of the National Conference on Artificial Intelligence, Philadelphia, pages 328--333, 1986.

....in logic programming ever since the first experiments in Kowalski s Logic for problem solving [21] first edition in 76) That Prolog with negation as finite failure can be interpreted as a sound theorem prover wrt to the completion of a Prolog program is known since Clark s work [2] in 78. When [18] introduced the YTS problem to show the failure of several nonmonotonic solutions to the frame problem, 4 Here the version of Noninertial=3 is needed which contains only rules for terminating effects of actions. 24 it was soon realised by some in the logic programming community that the Prolog ....

S. Hanks and D. McDermott. Default reasoning, nonmonotonic logic, and the frame problem. In Proc. of the National Conference on Artificial Intelligence, Philadelphia, pages 328--333, 1986.

.... transaction sequence drop(John; C100) register(Mary; C100) has been executed , we must determine whether Database j= 9c) enrolled(John; c; do(register(Mary; C100) do(drop(John; C100) S 0 ) Thus, querying an evolving database is precisely the temporal projection problem in AI planning [4]; this motivates the theoretical focus of this paper. 1 In the sequel, lower case roman letters will denote variables. All formulas are understood to be implicitly universally quantified with respect to their free variables whenever explicit quantifiers are not indicated. 3 An Axiomatization ....

S. Hanks and D. McDermott. Default reasoning, nonmonotonic logics, and the frame problem. In Proceedings of the National Conference on Artificial Intelligence, pages 328--333, 1986.

....To design tableau method for LAP is straightforward. It is just to combine the tableau rules for the logics K and S4. We omit a presentation here, because it will be a particular case of that presented in section 4. For details, see [CH96] Example 1 (The Yale Shooting Scenario (YSS) part I) [HM86] Consider a scenario where w, l and s represent the actions of waiting, loading and shooting, respectively. Let the atoms Ld and Al represent the gun is loaded and the man is alive , respectively. The effect of the load action is that the gun is loaded and the effects of the shoot action are ....

S. Hanks and D. McDermott. Default reasoning, nonmonotonic logics, and the frame problem. In Proc. Nat. (US) Conf. on Artificial Intelligence (AAAI'86), pages 328--333, Philadelphia, PA, 1986.

....future behaviors) Standard techniques for model checking finite state systems may then be used to analyze both component as well as system models constructed in this way. ffl It should be clear how some standard problems about reasoning with action and change, such as the Yale shooting problem [HM86] have simple solutions in tcc. In essence, tcc makes it possible to say that the effects of certain actions should persist indefinitely unless interrupted by the environment. By computing the least solution above the input supplied by the environment, tcc ensures that there is no deus ex machina ....

S. Hanks and D. McDermott. Default reasoning, nonmonotonic logics, and the frame problem. In Proceedings of Fifth National Conference on Artificial Intelligence, pages 328--333, 1986.

.... definitions that have been proposed are very sophisticated ones, difficult to understand (see e.g. the different variants of circumscription) Third, and in consequence of the previous, there is a strong risk of getting unintuitive results, as exemplified by the classical Yale shooting scenario [27], and Sandewall s review of the domain [54] If nonmonotonic mechanisms are required anyway, then this might be used as an argument against modal approaches: it should be simpler to add nonmonotonicity to classical logic than to a nonclassical logic. But in the last years there was a trend to ....

....of formulas with the conjunction of its elements. 3 What is reasoning about actions In this section we state the terminology of reasoning about actions in terms of the above modal language. We illustrate our language and our logic throughout the paper by the classical Yale shooting scenario [27]. 4 3.1 Belief bases, queries, and domain laws We can describe in our modal language three main concepts: knowledge bases, queries, and domain laws. Knowledge bases (KB) are finite sets of formulas without the 2 operator. Usually they represent the current state of affairs, such as fWalking ; ....

[Article contains additional citation context not shown here]

S. Hanks and D. McDermott. Default reasoning, nonmonotonic logics, and the frame problem. In Proc. Nat. (US) Conf. on Artificial Intelligence (AAAI'86), pages 328--333, Philadelphia, PA, 1986.

....axioms that are relevant for A: IC frame = f2(L [ff]L) jLj occurs in A; and ff6;jLjg Clearly, A is LAP; satisfiable iff A IC frame is LAP satisfiable, and the size of A IC frame is linear in that of A. Example 5 (The Yale Shooting Scenario (YSS) We consider the classical scenario of [HM86] adapted by [Bak91] Let wait, load and shoot be the actions of waiting, loading and shooting, respectively. Let the atoms Loaded , Alive and Walking represent the gun is loaded , the man is alive , and the man is walking respectively. The effect of the load action is that the gun is loaded ....

S. Hanks and D. McDermott. Default reasoning, nonmonotonic logics, and the frame problem. In Proc. Nat. (US) Conf. on Artificial Intelligence (AAAI'86), pages 328--333, Philadelphia, PA, 1986.

....actions, while most of the current methods are inappropriate for indefinite actions in the general case. 1 Introduction Since Hanks and McDermott revealed the problem of temporal projection (e.g. the Yale shooting problem) and showed that classical nonmonotonic logics failed to solve it [ 3 ] many solutions have been proposed to the problem [ 6; 12; 7; 8; 5; 1; 2 ] Although there are differences among these approaches, most of them employ some minimality principle in their formalisms. Generally, there are two ways to deal with the problem. One is either to extend the expressive ....

....in T . Instead of that, we will use the PSA [ 14 ] to construct the persistent extension of T when a situation is carried to another by an action or a sequence of actions, so as to from the persistent extension of T we can infer the facts that are true in the resulting situation. Definition 3 Let T be a primitive theory of ACT in situation S 0 , and fA 1 ; Delta Delta Delta ; An g ACT . We denote S i = do(A i ; S i Gamma1 ) for 1 i n. P ersist(T ; A 1 ; Delta Delta Delta ; An ) is a persistent extension of T with respect to the sequence of actions A 1 ; Delta Delta Delta ....

[Article contains additional citation context not shown here]

S. Hanks and D. McDermott. Default reasoning, nonmonotonic logics, and the frame problem. In Proceedings of the Fifth National Conference on Artificial Intelligence, pages 328-- 333, 1986.

....into autoepistemic logic. Even assuming rather strongly restricted forms of knowledge, the frame problem turned out to be unexpectedly difficult to solve in some of the above mentioned 15 systems of nonmonotonic reasoning. This was shown in the cases of circumscription and default logic by Hanks and McDermott [1986,1987] using the Yale Shooting problem as an example (cf. Section 1.3) In particular, they showed that an apparently natural formalization in circcumscription turned out to have consequences that were unexpectedly weak. Various solutions to the Yale Shooting problem were proposed in the ....

Steve Hanks and Drew McDermott. Default reasoning, nonmonotonic logics, and the frame problem. In Proc. AAAI-86, pages 328--333, 1986.

....if and only if exists a small positive 2 T such that H( Gamma ffl) c( for all in [ The intended and classical model set are defined in the usual fashion. 30 3.7 Examples The following is the most famous example of chronicle in Ksp IbAd . Scenario 3.7. 1 (Yale Shooting Scenario [HM86] The turkey Fred is initially alive and the gun is not loaded. Successively, a load, a wait and a shoot event occur. The common sense conclusion is the turkey dies after the shooting. Using the F F language as in [San94a] its representation is as follows: OBS: t 0 ]a:l t1 s2 SCD: s1 ; ....

Steve Hanks and Drew McDermott. Default Reasoning, Nonmonotonic logics, and the Frame Problem. In Proceedings of the Fifth National (U.S.) Conference on Artificial Intelligence, pages 328--333, Los Altos, CA, 1986. Morgan Kauffman. 85

....2 L ae in the unique stable expansion of a stratied theory Sigma can be tested using the query stratification( Sigma,S) reduct(S,R) l o consequence( R,true) 5. 6 An Example Application: The Yale Shooting Problem As an example application we present a solution to the Yale Shooting Problem of Hanks and McDermott [ 1986 ] which was constructed to show that the solution of the frame problem using nonmonotonic logics is not as trivial as had been thought earlier. The Yale Shooting problem is formalized as a simple situation calculus which is embedded in ordinary rst order logic. The binary predicate holds(F,S) is ....

S. Hanks and D. McDermott. Default reasoning, nonmonotonic logic, and frame problem. In Proceedings of the 5th National Conference on Articial Intelligence, pages 328333, Philadelphia, August 1986.

No context found.

S. Hanks and D. McDermott. Default reasoning, nonmonotonic logics, and the frame problem. In Proc. 5th Nat. Conf. on Artificial Intelligence (AAAI'86), pages 328--333, Philadelphia, 1986. Morgan Kaufmann Publishers.

No context found.

Hanks S., and D. McDermott, Default Reasoning, Nonmonotonic Logics, and the Frame Problem. Proc. of the 5th National Conference on Arti cial Intelligence (AAAI), Philadelphia, PA, AAAI Press, 328-333, 1986.

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