N. J. Nilsson. Logic and artificial intelligence. Artificial Intelligence 45:31--56, 1991.  Home/Search   Document Not in Database   Summary   Related Articles   Check

....We agree with Zadeh [13] that the imprecision that is intrinsic in natural languages is, for the most part, possibilistic rather than probabilistic in nature . the denotation of [an imprecise] word is generally a fuzzy . subset of a universe of discourse. We also agree with Nilsson [10] that For the most versatile machines, the language in which declarative knowledge is represented must be at least as expressive as first order predicate calculus (FOPC) So, we are interested in languages which support the expressive power of the full fuzzy FOPC. The inference engines for such ....

Nilsson N, 1991, "Logic and artificial intelligence," Artif. Intel., 47, 31-56.

....policies can also be expressed by rules. Logic has become the primary focus of knowledge representation in AI. This is because it provides a way to give meaning to symbols and a way to specify what you want to compute independently of how it s computed [42] It has often been argued (e.g. [33]) that any general representation scheme must be at least as rich as the first order predicate calculus. One of the problems with the first order predicate calculus is the way it handles uncertainty; all it has available is disjunction. This is a rather blunt instrument and doesn t do justice to ....

N. J. Nilsson. Logic and artificial intelligence. Artificial Intelligence, 47:31-- 56, 1991.

....Intelligence, 65:29 70, 1994. Also IRST Technical Report 9110 07, IRST, Trento, Italy. 18] F. Giunchiglia, L. Serafini, and A. Simpson. Hierarchical meta logics: intuitions, proof theory and semantics. In Proc. of META 92, Workshop on Metaprogramming in Logic, number 649 in LNCS, pages 235 249, Uppsala, Sweden, 1992. Springer Verlag. Also IRST Technical Report 9101 05, IRST, Trento, Italy. 19] F. Giunchiglia and L. Serafini. Hierarchical Meta Logics: some Proof Teoretical Results. Technical Report 9303 11, IRST, Trento, Italy, 1993. 26 ....

....of META 92, Workshop on Metaprogramming in Logic, number 649 in LNCS, pages 235 249, Uppsala, Sweden, 1992. Springer Verlag. Also IRST Technical Report 9101 05, IRST, Trento, Italy. 19] F. Giunchiglia and L. Serafini. Hierarchical Meta Logics: some Proof Teoretical Results. Technical Report 9303 11, IRST, Trento, Italy, 1993. 26 ....

[Article contains additional citation context not shown here]

N. Nilsson. Logic and artificial intelligence. Artificial Intelligence, 47:31--56, 1991.

....domains, the constraint language must be richly expressive. Knowledge representation research in Artificial Intelligence indicates that for . versatile [systems ] the language in which declarative knowledge is represented must be at least as expressive as first order predicate calculus [26]. We consider the discovery that constraint networks could provide the full expressive power of the predicate calculus to be an important conceptual breakthrough to arise from our work. Thus, in Galileo4, any sentence in first order predicate calculus is a well formed constraint. The language is ....

Nilsson, N. (1991) Logic and artificial intelligence. Artificial Intelligence, 47, 31-56.

....because any knowledge based system must work based on reasoning and logic is the systematic study of fundamental principles that underlie various valid reasoning forms. However, the hot controversy about the role of logic in AI has been repeated so far and probably will continue on as usual [15 18]. An important fact is that the logic as the center of the controversy is classical mathematical logic (CML for short) and or its various extensions, though there are some more logical logic systems. Until recently, what is debated by the researchers working on the fundamentals of AI is, among ....

N. J. Nilsson, "Logic and Artificial Intelligence," Artificial Intelligence, Vol.47, pp.31-56, 1991.

....possible world wW, the intended structure of w according to C is the structure S wC = D, R wC , where R wC = r(w) r is the set of extensions (relative to w) of the elements of . We shall denote with S C the set S wC wW all the intended world structures of C. 6 In a subsequent paper [37], Nils Nilsson stresses the importance of the conceptualization for a modeling task. 7 For instance, the two relations trilateral and triangle turn out to be the same, as they have the same extension in all possible worlds. 8 In the following, symbols denoting structures and sets of sets ....

Nilsson, N. 1991. Logic and Artificial Intelligence. Journal of Artificial Intelligence: 31-55.

....this allows to have links among links (and so on) as the one, represented by the dashed line in Figure 1(b) that links two links of the same kind, i.e. two is a links. The above assumptions are widely spread in many fields, for instance: artificial intelligence (under the label logicism [12, 22]) situation semantics [7] cognitive science [11] and human computer interaction [8] They can be criticized from many points of view (e.g. 4, 17, 18, 6] but they will be useful in the following of this paper. Thus, again, I do not take them as established truths, but as useful work ....

N. J. Nilsson. Logic and artificial intelligence. Artificial Intelligence, 47:31--56, 1991.

....the detail of arguments. An example of Tillers is given is used to show the how the framework could be used for legal reasoning. The code to run this example is available from the authors web site. 1 1 Introduction There are good normative arguments for using logic to represent knowledge (Nilsson, 1991; Poole, Mackworth Goebel, 1998) These arguments are usually based on reasoning with symbols with an explicit denotation, allowing relations amongst individuals, and permitting quantification over individuals. This is often translated as needing (at least) the first order predicate calculus. ....

Nilsson, N. J. (1991). Logic and artificial intelligence, Artificial Intelligence 47: 31-- 56.

....that goal may not be worthwhile when compared to another plan that gets to a less valuable state (e.g. it may not be worth trying to achieve world peace if that entails a risk of killing everyone on Earth) 1. 3 Logic and Uncertainty There are many normative arguments for the use of logic in AI (Nilsson 1991, Poole, Mackworth Goebel 1998) These arguments are usually based on reasoning with symbols with an explicit denotation, allowing relations amongst individuals, and permitting quantification over individuals. This is often translated as needing (at least) the first order predicate calculus. ....

Nilsson, N. J. (1991). Logic and artificial intelligence, Artificial Intelligence 47: 31--56.

....sentence would be meaningless to us. Hence understanding natural intelligence by necessity has always been among the goals of Intellectics (as of Cognitive Science) Different points of view for approaching the goal of creating artificial intelligence have been distinguished [Kus96] Logicism [Nil91] cognitivism [LNR87] and situated action [Agr95] are three out of several such points of view. In a nutshell, the logistic point of view argues that man can describe his creations (including an artificial intelligence) only by natural linguistic, hence logical means; thus any way towards ....

Nils J. Nilsson. Logic and artificial intelligence. Artificial Intelligence Journal, 47(1--3):31--56, 1991.

....that we, the designers of systems, can arbitrarily assign meanings to messages begs a number of important questions, one of which being how we as designers can construct intelligent systems composed of many autonomous agents 2 . 2 In an analogous fashion, Nilsson s remarks in a recent paper [10] begs a number of questions about the knowledge and role of designers in the construction of conventional intelligent systems. 2 The starting point for the discussion that follows is this: messages must serve some purpose. For a long time, I have believed that the very fact that there is a ....

Nilsson, N.J., Logic and Artificial Intelligence, Artificial Intelligence, Vol. 47, pp. 31-56, 1991.

....be objected that the fact that payroll suites often run in batch is an irrelevance this is correct. It can also be objected that one of the criteria that I have just presented explicit representation is rather operational. Explicit representation, according to the logicist view of AI (e.g. [32]) brings with it the benefit that represented items can be used more flexibly a hint at meaning asuse (which has turned out to be notoriously difficult for logic) One should be able to determine whether one has an AI program on the grounds of the behaviour it exhibits: one should not have to ....

....will concentrate on the case that he presents, as well as the stronger position that he is often supposed to take. The stronger position (i.e. the one that FOL must be used as the representation language) has been adopted by others, for example Moore [28] Genesereth and Nilsson [13] and Nilsson [32], and which was adopted by McDermott for a number of years (e.g. 24, 25] and [26] Of the various arguments in favour of FOL, I consider those of Hayes to be the clearest, the best articulated and, possibly, the most easily available: the critique that follows is, if anything, intended as a ....

Nilsson, N.J., Nilsson, N.J., Logic and Artificial Intelligence, Artificial Intelligence, Vol. 47, pp. 31-56, 1991.

....sentence would be meaningless to us. Hence understanding natural intelligence by necessity has always been among the goals of Intellectics (as of Cognitive Science) Different points of view for approaching the goal of creating artificial intelligence have been distinguished [Kus96] Logicism [Nil91] cognitivism [LNR87] and situated action [Agr95] are three out of several such points of view. In a nutshell, the logistic point of view argues that man can describe his creations (including an artificial intelligence) only by natural linguistic, hence logical means; thus any way towards ....

Nils J. Nilsson. Logic and artificial intelligence. Artificial Intelligence Journal, 47(1--3):31--56, 1991.

....policies can also be expressed by rules. Logic has become the primary focus of knowledge representation in AI. This is because it provides a way to give meaning to symbols and a way to specify what you want to compute independently of how it s computed [42] It has often been argued (e.g. [33]) that any general representation scheme must be at least as rich as the first order predicate calculus. One of the problems with the first order predicate calculus is the way it handles uncertainty; all it has available is disjunction. This is a rather blunt instrument and doesn t do justice to ....

N. J. Nilsson. Logic and artificial intelligence. Artificial Intelligence, 47:31-- 56, 1991.

....better, they possess different sets of consistent beliefs which may or may not be consistent among them. Moreover, it seems that, again, those sets of beliefs are dependent on the contexts of their use. It is thus acknowledged that no global consistency can model an intelligent agent (see also [25]) Models of inconsistent sets of knowledge about the same object are therefore common to both NLP and KR in general. 4 Conclusion I have presented two lines of argument of why NL should be a main concern for KR researchers, the first philosophical, the second essentially practical. I want now ....

Nilsson, Nils J. "Logic and artificial intelligence", Artificial Intelligence, 47, 1991, pp.31-56.

....because any knowledge based system must work based on reasoning and logic is the systematic study of fundamental principles that underlie various valid reasoning forms. However, the hot controversy about the role of logic in AI has been repeated so far and probably will continue on as usual [15 18]. An important fact is that the logic as the center of the controversy is classical mathematical logic (CML for short) and or its various extensions, though there are some more logical logic systems. Until recently, what is debated by the researchers working on the fundamentals of AI is, among ....

N. J. Nilsson, "Logic and Artificial Intelligence," Artificial Intelligence, Vol.47, pp.31-56, 1991.

....abductive logics, which are more context dependent and contradiction tolerant than deductive logic, are suitable for managing SC. Birnbaum s critique of logic based AI is really aimed at a more specific target: deductive logic. Birnbaum s arguments are an explicit reaction to Nilsson s advocacy [71] of logic based AI. Birnbaum claims that Nilsson s goal of contextfree characterisation of knowledge [7, p67] is doomed to failure. Given some theory containing a rule ff fi, then with deductive logic, in all interpretations where ff is true, then fi is also true. That is, in deduction, rules ....

N.J. Nilsson. Logic and Artificial Intelligence. Artificial Intelligence, 47:31--56, 1991.

....that goal may not be worthwhile when compared to another plan that gets to a less valuable state (e.g. it may not be worth trying to achieve world peace if that entails a risk of killing everyone on Earth) 1. 3 Logic and Uncertainty There are many normative arguments for the use of logic in AI (Nilsson 1991, Poole et al. 1998) These arguments are usually based on reasoning with symbols with an explicit denotation, allowing relations amongst individuals, and quantification over individuals. This is often translated as needing (at least) the first order predicate calculus. Unfortunately, the ....

Nilsson, N. J. (1991). Logic and artificial intelligence, Artificial Intelligence 47: 31--56.

....there have been two distinct approaches to achieving the general goal of intelligent computing systems. We will refer to these as the symbolic computation approach and the neural networks approach. In very general terms, those pursuing the symbolic computation approach, whether from the logicist [294] or the knowledge is all there is [235] schools, have been concerned with developing advanced computing systems which can perform complex problem solving tasks (e.g. chess playing, theorem proving, medical diagnosis, chemical structure analysis, constraint satisfaction, rule based deduction) ....

N J Nilsson. Logic and artificial intelligence. Artificial Intelligence, 47(1-3):31--56, January 1991.

....to get an objective and accurate artificial language. Model theoretic semantics was founded by Tarski s work. Although Tarski s primary target formal language, he also hoped that the ideas could be applied to reform everyday language [17] This approach is accepted by the logical approach to AI [11]. For a language L, defined by a finite formal grammar, a model M consists of a description of the relevant part of a domain, in another language ML, and an interpretation I, which maps the items in L into the items in ML. Given the above components, the meaning of a term in L is defined as its ....

N. Nilsson. Logic and artificial intelligence. Artificial Intelligence, 47:31-- 56, 1991.

....about AI, but rather a way of doing AI research. 1 The analysis in this section was suggested by [Agr93b] SA is not the only point view one might take towards AI. There is, of course, the classical point of view (which I will refer to as Cognitivism) Connectionism [Smo88] and Logicism [Nil91] are other perspectives on AI, although I will not discuss either in this paper. Cognitivism is, like SA, a perspective on AI rather than any particular theory. It has its own collection of ways of thinking about intelligence, strategies for deciding what aspects of AI are important and which are ....

N. Nilsson. Logic and artificial intelligence. Artificial Intelligence, 47, 1991.

....statements. Rule based programming systems restrict in various ways the well formed sentences (or rules) that may appear in a theory (or program) Because rule based systems admit only singleconsequent implication statements, they cannot, for example, handle negated or disjunctive knowledge [5]. However, for many applications, the most natural formation requires features of the FOPC that are not provided by Horn clauses. Thus the limited expressivenes of rule based languages may require considerable manipulation of the application domain knowledge in order to achieve a Horn clause ....

Nilsson N, 1991. "Logic and artificial intelligence," Artificial Intelligence, 47, 31-56.

....does not have to understand much how the receiver functions or how or whether the receiver will use it. The same fact can be used for many purposes, because the logical consequences of collections of facts can be available . Nillson made a similar point of view a central thesis of his recent paper [Nil87]: Thus my thesis: General intelligence depends on context free, declarative knowledge and on the means to manipulate it . 1 This part of the paper closely follows [Prz88a] The sharpest boundary exists between us and the proceduralists who claim that intelligence consists of having ....

N.J. Nilsson. Logic and artificial intelligence. In MIT Workshop on Foundations of AI, 1987.

No context found.

N. J. Nilsson. Logic and artificial intelligence. Artificial Intelligence 45:31--56, 1991.

No context found.

Nilsson, N. J. (1991) Logic and artificial intelligence. Artificial Intelligence 45, 31-56.

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