Eshghi K., Abductive Planning with Event Calculus, Proc. of the 5th International Conference and Symposium on Logic Programming (ICSLP), Seattle, WA, MIT Press, 562-579, 1988.  Home/Search   Document Not in Database   Summary   Related Articles   Check

....the limitations of deductive reasoning in classical logic. The role of abduction has been demonstrated in a variety of applications. It has been proposed as a reasoning paradigm in AI for diagnosis [8, 90] natural language understanding [8, 39, 4, 93] default reasoning [81, 29, 25, 50] planning [28, 110, 71, 59], knowledge assimilation and belief revision [54, 76] multi agent systems [7, 64, 102] and other problems. In the context of logic programming, the study of abductive inference started at the end of the eighties as an outcome of di#erent attempts to use logic programming for solving AI problems. ....

....inference started at the end of the eighties as an outcome of di#erent attempts to use logic programming for solving AI problems. Facing the limitations of standard logic programming for solving these problems, di#erent researchers proposed to extend logic programming with abduction. Eshghi [28] introduced abduction in logic programming in order to solve planning problems in the Event Calculus [65] In this approach, abduction solves a planning goal by explaining it by an ordered sets of events a plan that entails the planning goal. This approach was further explored by Shanahan [110] ....

K. Eshghi. Abductive planning with Event Calculus. In R.A. Kowalski and K.A. Bowen, editors, Proc. of the International Conference on Logic Programming. The MIT press, 1988.

....change applications and resource planning. 1 Introduction Abduction has been proposed as a reasoning paradigm in AI for fault diagnosis [4] natural language understanding [4] default reasoning [15] 39] In the context of logic programming, abductive procedures have been used for planning [14], 48] 35, 34] knowledge assimilation and belief revision [23] 21] database updating [22] 11] showed the role of an abductive system for forms of reasoning, di erent from planning, in the context of temporal domains with uncertainty. The term abduction was introduced by the logician and ....

....the existence of an abductive answer proves the satis ability of the observation. But abduction returns more informative answers, in the sense that it describes one, or in general a class of possible states of a airs in which the observation is valid. In the context of temporal reasoning, Eshghi [14] was the rst to use abduction. He used abduction to solve planning problems in the Event Calculus [28] This approach was further explored by Shanahan [48] Missiaen et al. 37, 34] 11] and [20] Planning in the event calculus can be seen as a variant of reasoning from observations to causes. ....

[Article contains additional citation context not shown here]

K. Eshghi. Abductive planning with Event Calculus. In R.A. Kowalski and K.A. Bowen, editors, Proc. of the International Conference on Logic Programming. MIT Press, 1988.

....only supports the prediction of a goal state, starting from a complete description of the initial state and the set of events. In planning, however, the set of events is the subject of the search, and thus, a priori unknown. A solution to this problem is to extend event calculus with abduction ([8], 20] In planning problems for example, the predicates which describe the events, i.e. happens=1; act=2 and are abducible. An abductive solution for a goal, describing the goal state, gives a description of a set of events and their order. A crucial property of the underlying abductive ....

....atoms cannot become ground. A procedure for planning should be able to cope with such situations. This condition is not satis ed by the abductive procedure de ned in [11] In the past, special abductive procedures for temporal reasoning with abductive event calculus have been presented ([8], 20] 16, 17] However, as the three authors argue, the treatment of constraints in [8] and the treatment of non ground abductive failure goals in [20] 16, 17] can be very inecient. The abductive procedure SLDNFA presented in this paper, provides an improved treatment of non ground abductive ....

[Article contains additional citation context not shown here]

K. Eshghi. Abductive planning with event calculus. In R.A. Kowalski and K.A. Bowen, editors, proc.of the 5th ICLP, 1988.

....between implementation and specification. Although this is a very important feature in the design of a planner the main issue of the planning task still remains: how to reduce the search space size There are several works aiming the construction of sound and complete logic based planning systems [7] [17] 12] 8] Moreover, some recent research results [6] demonstrate that, a good theoretical solution can co exist with a good practical solution despite of contrary widespread belief [13] In this paper we are going to analyze a particular logic based planner, named Abductive Event Calculus ....

....effects of actions are the only effects of actions, and by circumscribing Happens we assume that there are no unexpected event occurrences. A more thorough discussion of it can be found in [18] Planning in Event Calculus Planning in Event Calculus is naturally handled as an abduction process [7]. In this setting, planning a sequence of actions in order to satisfy a given goal G, given a domain description S is the abduction explanation D such that S, D G; which becomes with the circumscription policy stated above: G In other words, D is a plan for G . In [6] a planning system ....

K. Eshghi. Abductive planning with event calculus. In Proceedings of the Fifth International Conference on Logic Programming, 562-579, 1988.

....considered, in this version of the EC, to overcome the frame problem. The Event Calculus is a many sorted first order language comprising, basically: ffl A sort of time points t 1 ; t 2 ; t n ; t n 1 ; In this case, planning is considered to be an abduction task as defined in [14][43] From now on, when we use the term em Event Calculus to designate the simplified version of the Circumscriptive Event Calculus. ffl A sort of fluents f 1 ; f 2 ; f n ; f n 1 ; ffl A sort of event types a 1 ; a 2 ; a n ; a n 1 ; The following are the ....

....B and Sigma E comprises circumscriptions as we are going to discuss below. Acting Logically Together with the assimilation of the sensor data, we should have a machine to create sequences of actions based on the sensor information. In the present work this resembles planning in Event Calculus [14] [43] Roughly, the task is to find a sequence of robot actions Delta N such that given a goal Gamma Sigma B Sigma E Delta N Delta M j= Gamma; 37 where Sigma B ; Sigma E and Delta M are defined above. Thus, the symbol j= refers to second order inference as already discussed. In ....

K. Eshghi. Abductive planning with event calculus. In Proceedings of the Fifth International Conference on Logic Programming, pages 562--579, 1988.

....of backward persistence of properties and preprocessing away all reference to time periods. We do the latter by replacing all conditions referring to properties holding for time periods by their definition in terms of events happening at time points. Such a special case was introduced in ( 10] [5], 18] In [17] we showed that this special case, in completion form augmented with integrity constraints, has much of the power of the original event calculus, and even overcomes some of its deficiencies. The special case of the event calculus has at its core a single axiom: holds(P, T2) ....

Eshghi K., Abductive planning with event calculus, Proceedings of ICLP, MIT Press, Kowalski R.A. and Bowen K.A. (eds.), 562-579, 1988.

....the limitations of deductive reasoning in classical logic. The role of abduction has been demonstrated in a variety of applications. It has been proposed as a reasoning paradigm in AI for diagnosis [6, 79] natural language understanding [6, 33, 3, 81] default reasoning [70, 24, 20, 43] planning [23, 97, 61, 51], knowledge assimilation and belief revision [47, 65] multi agent coordination [5, 55] and other problems. In the context of logic programming, the study of abductive inference started at the end of the eighties as an outcome of di erent attempts to use logic programming for solving AI problems. ....

....inference started at the end of the eighties as an outcome of di erent attempts to use logic programming for solving AI problems. Facing the limitations of standard logic programming for solving these problems, di erent researchers proposed to extend logic programming with abduction. Eshghi [23] introduced abduction in 1 logic programming in order to solve planning problems in the Event Calculus [56] In this approach, abduction solves a planning goal by explaining it by an ordered sets of events a plan that entails the planning goal. This approach was further explored by Shanahan ....

K. Eshghi. Abductive planning with Event Calculus. In R.A. Kowalski and K.A. Bowen, editors, Proc. of the International Conference on Logic Programming. The MIT press, 1988.

.... to the frame problem have been worked out, and deductive planning based on the situational calculus has been considered extensively, in particular by the Toronto group, leading to the GOLOG planning language [40] In parallel, planning in the event calculus [38] has been pursued, starting from [15, 63]. Formulating planning problems as logical satisfiability problems, proposed by Kautz and Selman [36] enabled to solve large planning problems which could not be solved by specialized planning systems, and led to the efficient Blackbox planning system [37] In the same spirit, other approaches ....

K. Eshghi. Abductive Planning with Event Calculus. In Proc. 5th International Conference and Symposium on Logic Programming, pages 562--579. MIT Press, 1988.

....description of the initial state and the set of events. In many problems, the initial state or the events are not known. In planning, for example, the set of events is the subject of the search, and thus, a priori unknown. A solution to this problem is to extend event calculus with abduction ([3], 12] In planning problems for example, the predicates which describe the events, i.e. happens=1; act=2 and , are abductive. An abductive solution for a goal, describing the goal state, gives a description of a set of events and their order. A crucial property of the underlying abductive ....

....atoms will never become ground. A procedure for planning should be able to cope with such situations. This condition is not satisfied by the abductive procedure defined in [7] In the past, special abductive procedures for temporal reasoning with abductive event calculus have been presented ([3], 12] Recently, 9, 10] described an implementation of such a planner based on a special purpose abductive procedure. However, for the procedures in [3] and [12] no formalisation of the procedure is given and soundness and completeness results are lacking. In [9, 10] the abductive procedure ....

[Article contains additional citation context not shown here]

K. Eshghi. Abductive planning with Event Calculus. In R.A. Kowalski and K.A. Bowen, editors, Proc. of the International Conference on Logic Programming, 1988.

....1 Introduction The role of abduction [48] as a reasoning paradigm in AI is widely accepted. Abduction has been used for fault diagnosis [6] natural language understanding [6] default reasoning [22] 50] In the context of logic programming, abductive procedures have been used for planning [21], 52] 46, 44] knowledge assimilation and belief revision [32] 30] database updating [31] 18] showed the role of an abductive system for forms of reasoning, different from planning, in the context of temporal domains with uncertainty. In [15, 17] the role of abductive logic programming ....

....and satisfiability checking on incomplete knowledge has been shown in the context of a translation from a temporal language A [27] to abductive logic programming. In the past, a number of abductive extensions of SLDNF resolution have been proposed for abductive logic programs with negation [21], 52] 46, 45, 44] 31] 8] 51] 29] 13] 54] Anticipating the discussion of these procedures in section 11, we can say that either these procedures have not been formalised and proven correct senior research associate of the Belgian National Fund for Scientific Research 1 [21] 52] ....

[Article contains additional citation context not shown here]

K. Eshghi. Abductive planning with Event Calculus. In R.A. Kowalski and K.A. Bowen, editors, Proc. of the International Conference on Logic Programming, 1988.

....from A to extended logic programming [16] and proving its soundness, they show the expressivity of this formalism for representing temporal knowledge and, more in general, incomplete knowledge. In the past, another approach has been explored for temporal reasoning, based on event calculus [13], 31] 25] 10] 10] proposes solutions for the same benchmarks as in [17] This approach makes use of the formalism of abductive logic programming. One may interpret an abductive program as an open logic program in the sense that it contains only definitions for the non abducible predicates. ....

....to the frame problem. Yet, it seems that for many years, nobody, neither in the logic programming community nor in the A.I. community has come to this obvious conclusion or was interested in it. In the past, another approach has been explored for temporal reasoning, based on event calculus [22] [13] and [31] have simplified event calculus and have extended it with abduction for the purpose of planning. 31] extended event calculus to deal with necessary preconditions of actions. 25] implemented a planning system based on this formalism. Other work has been done to extend event calculus with ....

K. Eshghi. Abductive planning with Event Calculus. In R.A. Kowalski and K.A. Bowen, editors, Proc. of the International Conference on Logic Programming, 1988.

.... benchmark problems involving incomplete temporal knowledge and they proposed a sound transformation to extended programs, programs with both negation as failure and classical or explicit negation [12] In the past, another approach has been explored for temporal reasoning, based on event calculus [10], 21] 17] 8] 8] proposes solutions for the same benchmarks as in [13] This approach makes use of the formalism of abductive logic programming. One may interpret an abductive program as an incomplete logic program in the sense that it contains only definitions for the non abducible ....

....axiom in order to deal with formulas, universally quantified over all time points. In the future, we will investigate how such features can be integrated in the logic programming formalism. In the past, another approach has been explored for temporal reasoning, based on event calculus [15] [10] and [21] have simplified event calculus and have extended it with abduction for the purpose of planning. 21] extended event calculus to deal with necessary preconditions of actions. 17] implemented a planning system based on this formalism. Other work has been done to extend event calculus with ....

K. Eshghi. Abductive planning with Event Calculus. In R.A. Kowalski and K.A. Bowen, editors, Proc. of the International Conference on Logic Programming, 1988.

....P; T ) A planner uses this clause to introduce new events which initialise some desired property. Technically this is done by first skolemising and then abducing the happens goal. However, skolemisation requires explicit treatment of the equality predicate as an abducible satisfying FEQ [11]. The techniques proposed in this paper allow efficient treatment of the abduced equality atoms, and provide a declarative semantics for it. The paper is structured as follows. In section 2, we present the class of theories for which the model generation is designed. Section 3 recalls and extends ....

K. Eshghi. Abductive planning with Event Calculus. In R.A. Kowalski and K.A. Bowen, editors, Proc. of the International Conference on Logic Programming, 1988.

....explanations of the observations with respect to the knowledge base. Abduction was introduced by Peirce [Pei58,Pei92] in the beginning of the century and has been used in various AI applications [PMG98,Poo89] including: temporal explanations [Sha89,DMB92,Sha93] diagnosis [Reg83,Rei87] planning [Esh88,AKPT91,MD95] natural language understanding [HSAM90] Preprint submitted to Elsevier Science 9 March 2000 default reasoning [PGA87,Poo88a,EK89,KKT93] belief revision and updates [BB95,Bou96] and formulation of negation as failure [EK89,KKT93] Although, abduction is often used as a backward ....

K. Eshghi. Abductive planning with event calculus. In Proc. of the Fith ICLP., pages 562--579, 1988. 34

....free will action occurrences, closing the axioms with respect to these hypothetical occurrences, and testing whether the resulting axioms are satisfiable and entail the goal condition. This phenomenon of planning by abduction is quite widespread; it is used, for example, in the event calculus [5] and in Allen s temporal logic [1] In fact, it is the only way to do planning in logics which do not provide for branching futures. Unfortunately, abductive planning suffers from a number of drawbacks, when compared with the deductive approach: 1. It is a metalevel task; the planner must leave ....

K. Eshghi. Abductive planning with event calculus. In Proceedings of the Fifth International Conference on Logic Programming, pages 562--579. MIT Press, 1988.

.... each meta variable i is a time point constant such that i , and P is such that (EC1) EC7) T P j= G (2) or equivalently (EC1) EC7) T j= P G (3) Hence, under this formulation, planning can either be viewed as abducing sentences of the form P to add to the domain theory [4] [5] [14] 15] 16] or can be viewed as deducing theorems (of the form P G) in the spirit of Green, Reiter and others [7] 20] 21] In either case, it is necessary to ensure that the plan is consistent with the domain theory, i.e. that (EC1) EC7) T 6j= P (4) However, this need not ....

Kave Eshghi, Abductive Planning with Event Calculus, Proceedings of the 5th International Conference and Symposium on Logic Programming, ed.s Robert Kowalski and Kenneth Bowen, MIT Press, pages 562--579, 1988.

....agent, either a living creature or a machine. So, planning presents an important direction of research in AI. In general, given an initial state oe 0 of the agent s world W , and a goal G, a plan is a set of actions fffg ordered either fully (linear plans) or partially (nonlinear plans [2, 6, 27, 28]) such that their execution according to this order achieves the goal. The world W of the agent is described formally by a theory S (usually in a first order language) including a set of general constraints on the agent s activity, and a set AA of his her it admissible actions. Every action ff 2 ....

Eshghi, K. Abductive planning with event calculus. Proc. 5th International Conference and Symposium on Logic Programming, R. Kowalski and K. Bowen, Eds., Seattle, Wash., August 1988, MIT Press, 1988, 563-579.

....of various events. Introduced by Peirce [5] abduction underlies numerous processes of a great practical importance, such as medical diagnosing [2, 20, 57, 62, 68, 70, 83] testing and repairing technical devices and structures [4, 11, 13, 21, 24, 40, 55, 58, 72] planning a course of action [7, 18, 59], natural language understanding [32, 33, 43, 44, 39, 74, 75] learning [14, 61, 74, 75] Let S denote a Knowledge System (presented in the language of First Order Logic) that describes a part of a real world W . Suppose that a state Obs has been observed in W , and we wish to figure out what ....

....with regard to the notion of explanation. What accounts for an explanation What determines the quality of an explanation What is a best explanation What is a process of inference to the best explanation [48] Numerous works on abduction adopt the following general definition of explanation [18, 19, 46, 58, 59, 66, 76]. 1 Definition 1.1 Given a system S, and an observation Obs, an explanation of Obs in S is a formula expl(S; Obs) that satisfies the following conditions: i) S [ fexpl(S; Obs)g j= Obs; ii) expl(S; Obs) is consistent with S. Usually several refinements of this definition are considered. 1) ....

[Article contains additional citation context not shown here]

Eshghi, K., 1988, Abductive planning with event calculus. In Proceedings, 5th International Conference and Symposium on Logic Programming, 562--579.

....blocks world with one robot arm. De ning clear 2 (whether a block is clear at time t) in terms of on 3 (which expresses that a block is on a location (another block or the table) at a time t) we need only one basic property (on 3) and we can avoid the meta level of a holds 2 predicate as e.g. in [11]. constant maxstep = final state reached at time maxstep domain step = 0. maxstep 1. domain block = b1 b2 . domain table = table. domain location = block table. union domain action = move(block,location,location) skip. open function plan(step) action. PROGRAM ....

K. Eshghi. Abductive planning with event calculus. In R. Kowalski and K. Bowen, editors, Proc. of the International Conference on Logic Programming, pages 562-579. The MIT press, 1988.

.... THE RELATIONSHIP BETWEEN ABDUCTION AND DEDUCTION L uca CONSOLE , Daniele THESEIDER DUPRE, Pietro TORASSO Dipartimento di Informatica Universita di Torino 1 Corso Svizzera 185 0149 Torino (Italy) 2 Phone number : 39) 11 771200 Fax number : 39) 11 751603 E Mail: lconsole di.unito.it dtd di.unito.it t A torasso di.unito.i BSTRACT n a The aim of this paper is at analyzing from various points of view the relationships betwee bduction and deduction. In ....

.... THE RELATIONSHIP BETWEEN ABDUCTION AND DEDUCTION L uca CONSOLE , Daniele THESEIDER DUPRE, Pietro TORASSO Dipartimento di Informatica Universita di Torino 1 Corso Svizzera 185 0149 Torino (Italy) 2 Phone number : 39) 11 771200 Fax number : 39) 11 751603 E Mail: lconsole di.unito.it dtd di.unito.it t A torasso di.unito.i BSTRACT n a The aim of this paper is at analyzing from various points of view the relationships betwee bduction and deduction. In particular, we consider a meta level definition of ....

[Article contains additional citation context not shown here]

Eshghi, K., "Abductive Planning with Event Calculus," pp. 562-579 in Proc. 5th Int. Conf. and

....to temporal projection, and temporal projection in the event calculus is naturally cast as a deductive task. Given S, W and D as above, we re interested in HoldsAt formulae G such that, CIRC[S ; Initiates, Terminates, Releases] CIRC[D ; Happens] EC W G. 4 Conversely, as first pointed out by Eshghi [1988], planning in the event calculus can be considered as an abductive task. In terms of the circumscriptive event calculus, this task can be precisely characterised as follows. Definition 2.1. A domain description is a finite conjunction of formulae of the form, Initiates(a,b,t) P or, ....

....implementation. The basis of this implementation will be a resolution based abductive theorem prover, coded as a Prolog meta interpreter. This theorem prover is tailored for the event calculus by compiling the event calculus axioms into the meta level, resulting in an efficient implementation. Eshghi [1988] was the first author to show how abduction could be used to solve event calculus planning problems. The original event calculus of Kowalski and Sergot [1986] which was the inspiration for the present formalism, was also expressed a logic program. However, it cannot be executed directly as a ....

[Article contains additional citation context not shown here]

K.Eshghi, Abductive Planning with Event Calculus, Proceedings of the Fifth International Conference on Logic Programming (1988), pp. 562--579.

.... Plan ] j= Goal so that planning in the context of the Event Calculus can also be understood in terms of abduction (i.e. finding plans to add to the theory so that the goal is entailed) Indeed, it is this abductive view which is taken in the majority of work on Event Calculus planning, e.g. in [Eshg 88] Chle 96] JuFi 96] Shan 97b] Shan 98] and [Jung 98] 2.4 Non determinism In contrast to many versions of the Event Calculus, the axiomatisation described in (EC1) EC6) is non deterministic, in the sense that simultaneously initiating and terminating a fluent simply gives rise to two ....

.... has also been formulated in modal logic in [CeCh 95] CeCh 96] CeFr 97a] CeFr 97b] CeFr 98] and [ChMo 94] as an action description language in [KaMi 97] and [KaMi 98] and in an argumentation framework in [KaMi 99] The Event Calculus has been applied to planning using abduction in [Eshg 88] Chle 96] JuFi 96] Shan 97b] Shan 98] and [Jung 98] and in particular to cognitive robotics in [Shan 96a] Shan 96b] Shan 98] and [Shan 99b] Abduction in the context of the Event Calculus is also discussed in [DeMi 92] VaDe 94a] and [VaDe 94b] Other applications of the Event ....

K. Eshghi, Abductive Planning with Event Calculus, Proceedings of the 5th International Conference and Symposium on Logic Programming, ed.s Robert Kowalski and Kenneth Bowen, MIT Press, pp. 562-579, 1988.

....primary aim is to extract from the given theory T an explanation for the observation O and then reason with T augmented with this explanation. Completing the theory T with the information in the explanation is often a secondary issue. Consider for example the abductive framework of event calculus [4], 9] The general theory T contains the axiom holds(P; T 2 ) happens(E; T 1 ) T 1 T 2 ; initiates(E; P ) together with other (domain specific) rules that define the initiates predicate. This general axiom expresses the persistence of a property P from the time T 1 that it is initiated ....

Eshghi K. Abductive Planning with Event Calculus. Proceedings 5th International Conference on Logic Programming, 1988

....can be used in natural language understanding to interpret ambiguous sentences [9, 40, 53, 127] The abductive explanations correspond to the various possible interpretations of such sentences. In planning problems, plans can be viewed as explanations of the given goal state to be reached [30, 125]. These applications of abduction can all be understood as generating hypotheses which are causes for observations which are effects. An application that does not necessarily have a direct causal interpretation is knowledge assimilation [67, 72, 80, 89] The assimilation of a new datum can be ....

.... T 2 ) happens(E; T 1 ) T 1 T 2 ; initiates(E; P ) persists(T 1 ; P; T 2 ) 10 New information about the predicate holds at can be assimilated by adding an explanation in terms of some event that generates this property together with an appropriate assumption that the property persists [30, 62, 125]. This has the additional effect that the new KB will imply that the property holds until it is terminated in the future [125] This way of assimilating new information can also be used to resolve conflicts between the current KB and the new information [62, 125] Suppose for example that the ....

[Article contains additional citation context not shown here]

Eshghi, K., Abductive planning with event calculus. Proc. 5th International Conference and Symposium on Logic Programming, Washington, Seattle (1988) 562--579

No context found.

Eshghi K., Abductive Planning with Event Calculus, Proc. of the 5th International Conference and Symposium on Logic Programming (ICSLP), Seattle, WA, MIT Press, 562-579, 1988.

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