Showing papers presented at "Computational Intelligence in 1993"
1 May 1993
TL;DR: This article presents a new, two‐phase method for influence diagram evaluation, which takes advantage of asymmetry in influence diagrams to avoid unnecessary computation and generates a much smaller decision graph for the same influence diagram.
Abstract: As Influence diagrams become a popular representational tool for decision analysis, influence diagram evaluation attracts more and more research interests. In this article, we present a new, two--phase method for influence diagram evaluation. In our method, an influence diagram is first mapped into a decision graph and then the analysis is carried out by evaluating the decision graph. Our method is more efficient than Howard and Matheson''s two--phase method because, among other reasons, the size of the decision graph generated by our method from an influence diagram can be much smaller than that by Howard and Matheson''s method for the same influence diagram. Like those most recent algorithms reported in the literature, our method can also exploit independence relationship among variables of decision problems, and provides a clean interface between influence diagram evaluation and Bayesian net evaluation, thus, various well--established algorithms for Bayesian net evaluation can be used in influence diagram evaluation. In this sense, our method is as efficient as those algorithms. Furthermore, our method has a few unique merits. First, it can take advantage of asymmetric processing in influence diagram evaluation. Second, by using heuristic search techniques, it provides an explicit mechanism for making use of heuristic information that may be available in a domain--specific form. These additional merits make our method more efficient than the current algorithms in general. Finally, by using decision graphs as an intermediate representation, the value of perfect information can be computed in a more efficient way.
37 citations
1 May 1993
TL;DR: A view of generation in advice‐giving contexts which is different from the straightforward model of a passive selection of responses to questions asked by users is presented, and a procedure that estimates whether the ambiguity matters to the task of formulating a response is provided.
Abstract: Recognizing the plan underlying a query aids in the generation of an appropriate response. In this paper, we address the problem of how to generate cooperative responses when the user's plan is ambiguous. We show that it is not always necessary to resolve the ambiguity, and provide a procedure that estimates whether the ambiguity matters to the task of formulating a response. The procedure makes use of the critiquing of possible plans and identifies plans with the same fault. We illustrate the process of critiquing with examples. If the ambiguity does matter, we propose to resolve the ambiguity by entering into a clarification dialogue with the user and provide a procedure that performs this task. Together, these procedures allow a question-answering system to take advantage of the interactive and collaborative nature of dialogue in order to recognize plans and resolve ambiguity. This work therefore presents a view of generation in advice-giving contexts which is different from the straightforward model of a passive selection of responses to questions asked by users. We also report on a trial implementation in a course-advising domain, which provides insights on the practicality of the procedures and directions for future research.
35 citations
1 Feb 1993
TL;DR: This paper focuses on the semantics of Telos, a language for representing knowledge about information systems, and proposes an ontology of objects based on the property of existence, which will allow exactly what a knowledge base can be ASK‐ed or TELLs about the domain of discourse.
Abstract: This paper focuses on the semantics of Telos, a language for representing knowledge about information systems. Telos is intended to support the development of information systems, especially in the requirements modeling phase. An object-oriented representational framework is supported by Telos. Its features include aggregation, generalization, and classification, the treatment of attributes as first-class objects and the explicit representation of time. Telos also provides an assertion sublanguage for expressing deductive rules and integrity constraints. A possible-worlds semantics is defined for Telos knowledge bases. This semantics is intended to capture the peculiarities involved in the interpretation of temporal expressions. The integration of time has also inspired the treatment of existence in Telos. An ontology of objects based on the property of existence is proposed. In the spirit of KRYPTON, Telos knowledge bases are specified functionally, in terms of the operations provided for querying and updating them. This knowledge-level analysis will allow us to specify exactly what a knowledge base can be ASK-ed or TELL-ed about the domain of discourse. Soundness, consistency, and completeness results have also been proven to complete the specification of Telos knowledge bases. This formal account of the language provides a logical framework that can be used to verify the correctness of any proposed implementation of the system.
19 citations
1 Feb 1993
TL;DR: This paper provides an extension to the study of the clause management system (CMS) proposed by Reiter and de Kleer and shows two logic‐based diagnostic reasoning paradigms aided by the CMS to exemplify the functionality of the CMS.
Abstract: This paper provides an extension to the study of the clause management system (CMS) proposed by Reiter and de Kleer. The CMS is adapted specifically for aiding a reasoning system in explanations generation. The reasoning system transmits propositional formulae representing its knowledge to the CMS and in return, it requests the CMS for minimal and consistent explanations of a query with respect to the CMS knowledge base. The CMS knowledge base is represented by a set of prime implicates. The classification of implicates as minimal, prime, trivial, and minimal trivial is carefully examined. Similarly, the notion of a support for a clause including minimal, prime, trivial, and minimal trivial is also elaborated. The methods to compute these supports from implicates and a preference ordering scheme expressible by logical specificity for the set of supports for a given clause are also presented. The generalization of the notion of a minimal support for a conjunction of clauses is also shown. Finally, two logic-based diagnostic reasoning paradigms aided by the CMS are shown to exemplify the functionality of the CMS.
19 citations
1 Aug 1993
TL;DR: This work reformulates Reiter's default logic as special argument systems, which enables us, among other things, to apply the negation‐as‐failure rule to general default theories.
Abstract: We define an argument system to be a pair consisting of a set of inference rules and a set of completeness conditions. Inference rules are used to build arguments. Completeness conditions are used to define argument structures, which are sets of arguments supporting belief sets. We reformulate Reiter's default logic as special argument systems. This enables us, among other things, to apply the negation-as-failure rule to general default theories. We also speculate on some other potential uses of our argument systems.
15 citations
PARC1
TL;DR: A criterion for determining when average case results reflect typical behavior is introduced which allows the method developed here to be used for investigating other large‐scale behaviors of complex AI systems.
Abstract: Search is fundamental to artificial intelligence (AI) and numerous sophisticated search methods have been developed. We present a general, simple model of search processes and use it to analytically determine some typical behavior when applied to large problems. In particular, this identifies abrupt changes in overall search cost as small improvements are made in the underlying method. We also examine the robustness of this model's predictions in a range of more realistic cases. More generally, we introduce a criterion for determining when average case results reflect typical behavior which allows the method developed here to be used for investigating other large-scale behaviors of complex AI systems.
14 citations
1 Dec 1993
TL;DR: This work shows how the order on the Herbrand base can be used to compute perfect models of a disjunctive stratified finite theory and guarantees that every model generated belongs to the class of models being computed.
Abstract: Ordered model trees were introduced as a normal form for disjunctive deductive databases. They were also used to facilitate the computation of minimal models for disjunctive theories by exploiting the order imposed on the Herbrand base of the theory. In this work we show how the order on the Herbrand base can be used to compute perfect models of a disjunctive stratified finite theory. We are able to compute the stable models of a general finite theory by combining the order on the elements of the Herbrand base with previous results that had shown that the stable models of a theory T can be computed as the perfect models of a corresponding disjunctive theory ɛT resulting from applying the so called evidential transformation to T. While other methods consider many models that are rejected at the end, the use of atom ordering allows us to guarantee that every model generated belongs to the class of models being computed. As for negation-free databases, the ordered tree serves as the canonical representation of the database.
14 citations
1 Aug 1993
TL;DR: It is argued that within the limits of the worst‐case computational complexity, it is possible to improve on the average‐case complexity of reason maintenance and query processing by eliminating computation that is of no relevance to the problem solver's performance.
Abstract: The basic algorithms involved in reason maintenance in the standard ATMS is known to have a computational complexity that is exponential in the worst case. Yet, also in average-case problem solving, the ATMS often lays claim to a major part of the computational effort spent by a problem solver/ATMS system. In this paper, we argue that within the limits of the worst-case computational complexity, it is possible to improve on the average-case complexity of reason maintenance and query processing by eliminating computation that is of no relevance to the problem solver's performance. To this purpose, we present a set of algorithms designed to control the effort spent by the ATMS on label updating. The basic idea underlying these algorithms is that of lazy evaluation: labels are not automatically maintained on all datums but are computed only when needed (either directly or indirectly) by the problem solver. The algorithms have been implemented in the LazyRMS with which we have experimented in the context of model-based diagnosis; our experiments show a substantial saving in the computational effort spent on reason maintenance.
13 citations
1 Feb 1993
TL;DR: This work suggests that links be interpreted as conditional sentences with appropriate truth conditions rather than uninterpreted “reasons,” and identifies some key differences between the account of inference and those based on the notion of inferential distance with respect to the stability of reasoning.
Abstract: Inheritance reasoners have traditionally been viewed as argument systems, or algorithms that determine reasonable conclusions by constructing acceptable arguments. While the intended meaning of links in such networks is understood, formal semantic accounts are troublesome, as are semantic accounts of the inference process. We adopt a different perspective, suggesting that links be interpreted as conditional sentences with appropriate truth conditions rather than uninterpreted “reasons.” The conditional logic CT4D is used for this purpose. Furthermore, we characterize inference in our networks in terms of preferred (or minimal) models. In the process, we identify some key differences between our account of inference and those based on the notion of inferential distance, specifically with respect to the stability of reasoning. Key words: nonmonotonic reasoning, inheritance hierarchies, minimal models, conditional logic.
8 citations
1 May 1993
TL;DR: This approach provides a unified formalism needing no preprocessing or postprocessing to the functional language itself: parsing and semantics are declared naturally using function definition and evaluation is done by lambda application along the lines of Montague.
Abstract: This paper describes an original approach to semantics representation based on the use of a non-strict functional programming language with polymorphic typing. This approach provides a unified formalism needing no preprocessing or postprocessing to the functional language itself: parsing and semantics are declared naturally using function definition and evaluation is done by lambda application along the lines of Montague. We show that by changing only the model we can, after parsing, compute either the truth value of a sentence or its parse tree.
6 citations
1 Feb 1993
TL;DR: The approach grants the ability to a priori estimate worst case response time and memory requirements, and to better predict the effectiveness of knowledge acquisition efforts.
Abstract: A method is presented of establishing bounds on the number of classification rules in such applications as credit worthiness assessment, investment decisions, premium determination, consumer choices, employee selection, and editorial preferences, to name just a few. A function that relates the maximum number of classification rules to the problem space size of such application domains is established. It is shown that in this important class of ordinal classification problems, the maximum possible number of rules is significantly lower than the relative problem space sizes. The approach grants the ability to a priori estimate worst case response time and memory requirements, and to better predict the effectiveness of knowledge acquisition efforts.
1 May 1993
TL;DR: This work gives a learning algorithm that uses examples and membership queries to PAC learn the intersection of K‐nonoverlapping perceptrons, regardless of whether the instance space in Boolean, discrete, or continuous.
Abstract: We investigate the problem of learning two-layer neural nets of nonoverlapping perceptrons where each input unit is connected to one and only one hidden unit. We first show that this restricted problem with no overlap at all between the receptive fields of the hidden units is as hard as the general problem (with total overlap) if the learner uses examples only. However, if membership queries are allowed, the restricted problem is indeed easier to solve. We give a learning algorithm that uses examples and membership queries to PAC learn the intersection of K-nonoverlapping perceptrons, regardless of whether the instance space in Boolean, discrete, or continuous. An extension of this algorithm is proven to PAC learn two-layer nets with K-nonoverlapping perceptrons. The simulations performed indicate that both algorithms are fast and efficient.
1 Aug 1993
TL;DR: This paper presents an approach that allows base algorithms to be combined, giving us new hybrids, and it is shown that FC‐CBJ is by far the best of the algorithms examined.
Abstract: It might be said that there are five basic tree search algorithms for the constraint satisfaction problem (csp), namely, naive backtracking (BT), backjumping (BJ), conflict-directed backjumping (CBJ), backmarking (BM), and forward checking (FC). In broad terms, BT, BJ, and CBJ describe different styles of backward move (backtracking), whereas BT, BM, and FC describe different styles of forward move (labeling of variables). This paper presents an approach that allows base algorithms to be combined, giving us new hybrids. The base algorithms are described explicitly, in terms of a forward move and a backward move. It is then shown that the forward move of one algorithm may be combined with the backward move of another, giving a new hybrid. In total, four hybrids are presented: backmarking with backjumping (BMJ), backmarking with conflict-directed backjumping (BM-CBJ), forward checking with backjumping (FC-BJ), and forward checking with conflict-directed backjumping (FC-CBJ). The performances of the nine algorithms (BT, BJ, CBJ, BM, BMJ, BM-CBJ, FC, FC-BJ, FC-CBJ) are compared empirically, using 450 instances of the ZEBRA problem, and it is shown that FC-CBJ is by far the best of the algorithms examined.
1 May 1993
TL;DR: In this article, the authors derive reasonable constraints that enable a natural partition of a domain and its representation by separate Bayesian subnets, such that evidential reasoning takes place at only one of them at a time; and marginal probabilities obtained are identical to those that would be obtained from the homogeneous network.
Abstract: Bayesian networks provide a natural, concise knowledge representation method for building knowledge-based systems under uncertainty. We consider domains representable by general but sparse networks and characterized by incremental evidence where the probabilistic knowledge can be captured once and used for multiple cases. Current Bayesian net representations do not consider structure in the domain and lump all variables into a homogeneous network. In practice, one often directs attention to only part of the network within a period of time; i.e., there is “localization” of queries and evidence. In such case, propagating evidence through a homogeneous network is inefficient since the entire network has to be updated each time. This paper derives reasonable constraints, which can often be easily satisfied, that enable a natural {localization preserving) partition of a domain and its representation by separate Bayesian subnets. The subnets are transformed into a set of permanent junction trees such that evidential reasoning takes place at only one of them at a time; and marginal probabilities obtained are identical to those that would be obtained from the homogeneous network. We show how to swap in a new junction tree, and absorb previously acquired evidence. Although the overall system can be large, computational requirements are governed by the size of one junction tree.