D. Poole. Normality and faults in logic-based diagnosis. In IJCAI '89, pages 1304--1310, 1989.  Home/Search   Document Details and Download   Summary   Related Articles   Check

....vantages. First, any set of rules of this form will be unable to specify all the conditions (such as insincerity) that might also influence the agent s interpreta tion; a reasoner will need also to assume that there are no abnormalities relevant to the participants or the speech event [Poole, 1989]. This approach also ignores the many other possible interpretations that participants might achieve through negotiation, independent of their actual beliefs. For example, an agent s response to a yes no question might treat it as a question, a request, a warning, a test, an insult, a ....

....(Figure 1 ineludes some examples in Theorist. In cases of appar ent misunderstanding, the same explanation process suggests a misunderstanding, rather than a planned act, as the reason for the utterance. To handle these cases, the model needs a theory of the symptoms. of a failure to understand [Poole, 1989]. For example, a speaker S2 might explain an otherwise unexpected response by a speaker S1 by hypothesizing that S2 has mistaken some speech act by S1 for another with a similar decomposition or S2 might hypothesize that SI has misunderstood (see Figure 2) We shall now consider some ....

David Poole. Normality and faults in logic-based diagnosis. In Proceedings of the 1lib International Joint Conference on Artificial Intelligence, pages 1304-1310, 1989.

....conventional consistency based diagnosis assumes that the structure of the system and the connections between components are xed. In the case of a faulty program, a diagnosis may require identifying missing or super uous components, or restructuring existing components. The abductive approach [POOL89] does not distinguish correct and incorrect states: it simply de nes multiple modes of behavior. An abductive diagnosis is a minimal set of assumptions that produce the observed output from the system description. Heuristic diagnoses map speci c symptoms to probable errors based on expert ....

Poole, D., Normality and faults in logic-based diagnosis. Proc. 11th Intl. Joint Conf. On Arti cial Intelligence, IJCAI '89, pp. 1304-1310, Detroit, Morgan Kaufman, 1989

....of an ADP P = hH; T; Oi is also a diagnosis of the corresponding consistency based diagnosis problem (provided that hypotheses do not occur in the heads of the rules) while the vice versa does not hold in general. For comparisons between the abductive and consistency based approaches, see [32, 24, 6]. Example 6 Consider the circuit description of a full adder (depicted in Fig. 3) For the system description T add , we rst give de nitions for the gates involved. The and gates: out(X; 1) and(X) in1(X; 1) in2(X; 1) ab(X) out(X; 0) and(X) in1(X; 0) ab(X) out(X; 0) and(X) ....

D. Poole. Normality and Faults in Logic-Based Diagnosis. In Proceedings of the Eleventh International Joint Conference on Arti cial Intelligence (IJCAI'89), pages 1304-1310, Detroit, Michigan, USA, 1989.

....of an ADP P = hH; T ; Oi is also a diagnosis of the corresponding consistency based diagnosis problem (provided that hypotheses do not occur in the heads of the rules) while the vice versa does not hold in general. For comparisons between the abductive and consistencybased approaches, see [32, 24, 6]. x1 x2 a1 a2 o1 i1 i3 u1 u2 Fig. 3. Circuit diagram of a full adder Example 6 Consider the circuit description of a full adder (depicted in Fig. 3) For the system description T add , we first give definitions for the gates involved. The and gates: out(X; 1) and(X) in1(X; 1) ....

D. Poole. Normality and Faults in Logic-Based Diagnosis. In Proceedings of the Eleventh International Joint Conference on Artificial Intelligence (IJCAI'89), pages 1304--1310, Detroit, Michigan, USA, 1989.

....of MAB diagnosis will be discussed. The techniques proposed by Reiter are not the only possible ways to formalise DNSB and MAB diagnosis; D. Poole has proposed other logical techniques for the same purpose in terms of his Theorist framework of hypothetical reasoning [Poole et al. 1987; Poole, 1990a; Poole, 1990b; Poole, 1994] This work, however, bears great resemblance to the work by Reiter with respect to DNSB diagnosis, and to the work by Console and Torasso with respect to MAB diagnosis, which will be discussed in the following section. The Theorist framework is discussed in Section 5. ....

D. Poole (1990). Normality and faults in logic-based diagnosis. In Proceedings of the 11th International Joint Conference on Artificial Intelligence, 1304-1310.

....the last years, there has been increasing interest in abduction in di#erent areas of computer science. It has been recognized that abduction is an important principle of common sense reasoning, and that abduction has fruitful applications in a number of areas such diverse as model based diagnosis [19], speech recognition [11] maintenance of database views [14] and vision [3] In the past, most research on abduction concerned abduction from classical logic theories. However, we argue that the use of logic programming to perform abductive reasoning can be more appropriate in several ....

D. Poole. Normality and Faults in Logic Based Diagnosis. In Proceedings IJCAI-89, pp. 1304--1310, 1989.

....(see [3] These applications have in common that the models being searched for have to be nite. In Diagnosis a description of a system is given and symptoms of the system are observed. The generated models correspond to explanations of the observed symptoms and therefore must be nite (see e.g. [11, 1]) Solving Planning and Design problems can as well be seen as model generation. The speci cations can describe an environment, e.g. the possible movements of a robot, a starting position, and a goal to reach. Each nite model describes a solution while in nite models are meaningless. Several ....

D. Poole. Normality and faults in logic-based diagnosis. In 11th Int. Joint Conf. on Articial Intelligence, pages 1304-1310, 1985.

.... or plausibility, rather than re#ecting the changing needs for information that motivate explanation #Charniak Goldman, 1991; Charniak Shomony, 1994; Hobbs, Stickel, Appelt, Martin, 1993; Josephson Josephson, 1994; Kautz Allen, 1986; Konolige, 1990; Levesque, 1989; O Rorke, 1994; Poole, 1989; Zadrozny, 1994#. Although suchapproaches have proven useful in a number of contexts, problems arise when trying to apply these methods to the rich domain of everyday abductive explanation. Modeling everyday abductive reasoning requires confronting the rami#cations of imperfect knowledge, limited ....

.... as reasoning chains deriving or supporting belief in a state or event to be explained #Charniak, 1986; Charniak Goldman, 1991; Charniak Shomony, 1994; Hobbs et al. 1993; Josephson Josephson, 1994; Kautz Allen, 1986; Konolige, 1990; Leake, 1992; Levesque, 1989; O Rorke, 1994; Mooney, 1990; Poole, 1989; Wilensky, 1983#. 3 The #rst issue we will address is the nature of explanatory reasoning chains. In many models, these chains are viewed as deductive proofs that may depend on additional abductive assumptions; the case based model instead maintains that models of everyday abductive reasoning ....

[Article contains additional citation context not shown here]

Poole, D. #1989#. Normality and faults in logic-based diagnosis. In Proceedings of the Eleventh International Joint Conference on Arti#cial Intelligence, pp. 1304#1310 Detroit, MI. IJCAI.

....Reasoning activities As discussed above, we defined different reasoning activities to exploit the potentiality of the types of models we consider. As we will see in the rest of this section, they range in the broad spectrum of the possible diagnostic reasoning activities, as reported in [16] [40][COTO91] POOL89] and others. In particular, we show below how abductive and consistency based diagnosis can interact opportunistically, exploiting the characteristics of the models we have defined. We also discuss when fault models may provide useful information to reach a satisfactory ....

[POOL89]D. L. Poole, "Normality and Faults in Logic-Based Diagnosis" Proceedings of IJCAI-89, Morgan Kaufman, San Mateo, CA, pp. 1304-1310, 1989. Draft! Please, do not circulate. 37

....from the area of model based diagnosis. Diagnosis is the problem of detecting failures in a system, based on our knowledge of the system structure, its possible malfunctions, and observations of symptoms made on the system [ Reiter, 1987 ] There are basically two approaches to diagnosis ( Poole, 1989; Console and Torasso, 1991; Mozeti c, 1992; Hamscher et al. 1992 ] The consistency based approach. In this approach, the system is de ned by a set SD of rst order formulas, the system description, and a set COMPS of constants, the system components. The system description de nes how the ....

....SD contains di erent modes of behavior and does not distinguish between the normal and the abnormal behavior of the system. In other words, the normal behavior of the system is not sucient to obtain a diagnosis (we must also express the abnormal situations in the system description, as stated in [ Poole, 1989 ] An abductive diagnosis is a minimal set of assumptions which, together with SD entail OBS: SD [ fAB(c)jc 2 g[ f:AB(c)jc 2 COMPS g j= OBS SD [ fAB(c)jc 2 g[ f:AB(c)jc 2 COMPS g 6j= An abductive diagnosis problem can be restated in terms of an abductive expansion: given the set of ....

David Poole. Normality and faults in logicbased diagnosis. In Proceedings of the 11th International Joint Conference on Articial Intelligence, pages 1304-1310, Detroit, August 1989.

....the last years, there has been increasing interest in abduction in different areas of computer science. It has been recognized that abduction is an important principle of common sense reasoning, and that abduction has fruitful applications in a number of areas such diverse as model based diagnosis [60], speech recognition [28] maintenance of database views [36] and vision [12] In the past, most research on abduction concerned abduction from classical logic theories. However, we argue that the use of logic programming to perform abductive reasoning can be more appropriate in several ....

D. Poole. Normality and Faults in Logic Based Diagnosis. In Proc. IJCAI '89, pp. 1304--1310, 1989. 46

....approaches to extend or combine them have been presented [11] 28] 40] Abductive approaches have been introduced to cope with a system description based on fault models. In its original definition, abductive diagnosis implies the availability of a complete representation of any possible fault [32]. We consider only the available fault models and use them in some of our diagnostic activities. Since other activities operate on different knowledge, we do not need a model for any possible fault. In the consistency based approach [12] 15] 27] 33] the diagnostic space (i.e. the space of ....

....activities As mentioned above, we have defined different reasoning activities to exploit the potentiality of the types of models we consider. As we will see in the rest of this section, they range on the broad spectrum of the possible diagnostic reasoning activities, as reported in [11] [32] and others. In particular, we show below how abductive and consistency based diagnosis can interact opportunistically, exploiting the characteristics of the models we have defined. One of the main features of our approach is that all these reasoning activities are integrated in a unique ....

D. L. Poole, "Normality and Faults in Logic-Based Diagnosis" Proceedings of IJCAI-89, Morgan Kaufman, San Mateo, CA, pp. 1304-1310, 1989.

....of MAB diagnosis will be discussed. The techniques proposed by Reiter are not the only possible ways to formalise DNSB and MAB diagnosis; D. Poole has proposed other logical techniques for the same purpose in terms of his Theorist framework of hypothetical reasoning [Poole et al. 1987; Poole, 1990a; Poole, 1990b; Poole, 1994] This work, however, bears great resemblance to the work by Reiter with respect to DNSB diagnosis, and to the work by Console and Torasso with respect to MAB diagnosis, which will be discussed in the following section. The Theorist framework is discussed in Section 5. ....

D. Poole (1990). Normality and faults in logic-based diagnosis. In Proceedings of the 11th International Joint Conference on Artificial Intelligence, 1304--1310.

....as a knowledge representation formalism in building expert systems. Although there seems to be ample reasons to study the suitability of logic as a knowledge representation language for building expert systems, only few researchers have actually investigated expert systems from this perspective [15, 18, 34]. The only expert systems having a basis akin to logic are those developed in the field of logic programming, usually applying PROLOG. It should, however, be noted that logic programming is only concerned with a subset of first order logic (Horn clause logic) and that PROLOG is a programming ....

.... we focus on the logical representation of diagnostic reasoning in the spirit of MYCIN like rule based expert systems [4, 9] In this formalization, diagnostic reasoning is viewed as a deductive process instead of as an abductive process, the other frequently adopted view of diagnostic reasoning [34]. We describe our attempt to reformulate the HEPAR system, a rule based expert system for the diagnosis of disorders of the liver and biliary tract, to first order predicate logic (actually, many sorted predicate logic, but we disregard the sorts in this article) 25] The problems encountered in ....

D. Poole, Normality and faults in logic-based diagnosis, Proceedings of the 11th International Joint Conference on Artificial Intelligence, (Morgan-Kaufman, Los Altos, 1990) 1304-1310.

....Subrahmanian [Sub89] Our meta programs are written in typewriter style and object programs in math style. All the results presented assume the consistency based approach to diagnosis (e.g. de Kleer and Williams [dKW87, dKW89] Reiter [Rei87] and Struss [SD89] For the abductive view of Poole [Poo89], Console and Torasso [CT90, CT91] and Friedrich, Gottlob, and Nejdl [FGN90] the basic approach of this paper remains valid but will require some extensions modifications. In the remainder we often use variables as shorthand for their ground instances, which are covered by the underlying theory. ....

David Poole. Normality and faults in logic-based diagnosis. In Proceedings of the International Joint Conference on Artificial Intelligence (IJCAI), pages 1304--1310, Detroit, August 1989. Morgan Kaufmann Publishers, Inc.

....of inductive and abductive reasonings led and leads to numerous debates. Many works are based on these reasonings, for example in machine learning (e.g. 7, 11] inductive logic programming (see [14] for a survey) abductive logic programming (e.g. 16] or resolution of diagnosis problems (see [4, 17, 3] for various approaches to abduction) However many of these works just deal with one of these reasonings. We situate our paper within a very recent tendency consisting of the analysis and characterization of relations between induction and abduction. Many definitions are proposed for inductive ....

D. Poole, 1989. Normality and faults in logic-based diagnosis in proc. IJCAI'89, pp1304-1310

....methods. In abduction, the hypothesis is a specific set of assumptions that explain the observations of a particular case; while in induction, the hypothesis is a general theory that explains the observations across a number of cases. The body of logical work on abduction, e.g. Pople, 1973; Poole, 1989; Levesque, 1989; Ng Mooney, 1991, 1992; Kakas, Kowalski, Toni, 1993) generally fits this definition of abduction and several diagnostic models (Reiter, 1987; Peng Reggia, 1990) can be shown to be equivalent or a special case of it (Poole, 1989; Ng, 1992) The work on inductive logic ....

.... of logical work on abduction, e.g. Pople, 1973; Poole, 1989; Levesque, 1989; Ng Mooney, 1991, 1992; Kakas, Kowalski, Toni, 1993) generally fits this definition of abduction and several diagnostic models (Reiter, 1987; Peng Reggia, 1990) can be shown to be equivalent or a special case of it (Poole, 1989; Ng, 1992) The work on inductive logic programming (ILP) Muggleton, 1992; Lavrac Dzeroski, 1994) employs this definition of induction, and most machine learning work on induction can also be seen as fitting this paradigm (Michalski, 1983) 3 Abduction in Theory Revision Theory revision or ....

Poole, D. (1989). Normality and faults in logic-based diagnosis. In Proceedings of the Eleventh International Joint Conference on Artificial Intelligence, pp. 1304--1310 Detroit, MI.

.... models [Struss 92] and multiple component modes [deKleer, Williams 89] Dressler, Struss 92] diagnostic inference patterns, e.g. alibis [Raiman 89] physical negation [Struss, Dressler 89] and physical impossibility [Friedrich et al. 90] integration of abduction [Console, Torasso 90] Poole 89] Ayeb et al. 90] use of non intermittency [Raiman et al. 91] and . inferential power [de Kleer 91] Nevertheless,a main obstacle for building large diagnostic systems remains. How can we control when to use which inference pattern or representation Consider a ballast tank system as shown ....

D. Poole, Normality and Faults in Logic-based Diagnosis, Proc. IJCAI 89

No context found.

D. Poole. Normality and faults in logic-based diagnosis. In IJCAI '89, pages 1304--1310, 1989.

No context found.

D. Poole. Normality and faults in logic-based diagnosis. In Proceedings of the Eleventh International Joint Conference on Artificial Intelligence (IJCAI-89), pages 1304--1310, 1989.

No context found.

D. Poole. Normality and faults in logic-based diagnosis. In Proceedings of the Eleventh International Joint Conference on Artificial Intelligence (IJCAI-89), pages 1304--1310, 1989. 39

No context found.

David Poole. Normality and faults in logic-based diagnosis. In IJCAI-89, pages 1304--1310, 1989.

No context found.

David Poole. Normality and faults in logicbased diagnosis. In IJCAI-89, pages 1304--1310, 1989.

No context found.

D. Poole. Normality and faults in logic-based diagnosis. In Proceedings of the 11th International Joint ConferenceonArti#cial Intelligence #IJCAI'98#, 1989.

No context found.

D. L. Poole, "Normality and Faults in Logic-Based Diagnosis" Proceedings of IJCAI-89, Morgan Kaufman, San Mateo, CA, pp. 1304-1310, 1989.

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