Showing papers in "Artificial Intelligence in 1989"

TL;DR: SME has been built to explore Gentner's structure-mapping theory of analogy, and provides a "tool kit" for constructing matching algorithms consistent with this theory, making it a useful component in machine learning systems as well.

TL;DR: In this paper, each connection computes the derivative, with respect to the connection strength, of a global measure of the error in the performance of the network, and the strength is then adjusted in the direction that decreases the error.

TL;DR: The definition, theory, and extant applications of classifier systems are reviewed, comparing them with other machine learning techniques, and closing with a discussion of advantages, problems, and possible extensions of classifiers.

TL;DR: A systematic way of regrouping constraints into hierarchical structures capable of supporting search without backtracking is offered.

TL;DR: This paper shows how formal characterizations of causality and of the method of comparative statics, long used in econometrics, thermodynamics and other domains, can be applied to clarify and make rigorous the qualitative causal calculus recently proposed by de Kleer and Brown.

TL;DR: In this paper, an explanation-based learning (EBL) approach is proposed to improve problem solving performance through experience, where the explanations are converted into operational recognition rules, which are then converted into explanations for a particular instance of a concept.

TL;DR: It is argued that computational studies of analogy are in a state of adolescence: looking to more mature research areas in artificial intelligence for robust accounts of basic reasoning processes and drawing upon a long tradition of research in other disciplines.

TL;DR: This work suggests a new, more powerful form of the negation as failure rule and describes an important class of theories for which this form of negations as failure is equivalent to particular forms of circumscription, establishing a close relationship between the two important formalizations of nonmonotonic reasoning.

TL;DR: A logical framework for depiction and interpretation that formalizes image domain knowledge, scene domain knowledge and the depiction mapping between the image and scene domains is proposed and can be used as a foundation for the specification, design and implementation of vision and graphics systems that are correct with respect to the task and algorithm levels.

TL;DR: This algorithm is a solution to the consistent labeling problem with almost optimal worst-case time and space complexity.

TL;DR: A class of connectionist networks is described that has learned to play backgammon at an intermediate-to-advanced level and is suggested how these techniques might be used in applications of network learning to general large-scale, difficult “real-world” problem domains.

TL;DR: A query answering algorithm for circumscriptive theories using McCarthy's theory of circumscription, which can be used to answer queries in theories circumscribed by prioritized circumscription.

TL;DR: These parameterized algorithms are found to encompass a wide class of best first search algorithms and guarantee admissible solutions within specified memory limitations (above the minimum required).

TL;DR: Derivational analogy as mentioned in this paper solves a problem by replaying the plan used to solve a previous problem, modifying it where necessary, and reuse partial plans to adapt to a new problem, and show how each system's approach to these seven issues affects the scope of problems it can solve, its ability to solve new problems, the QUALITY of its solutions, the EFFICIENCY of its computation and its AUTONOMY from the user.

TL;DR: The 3D shape of a textured surface is recovered from its projected image on the assumption that the texture is homogeneously distributed, and the 3D recovery equations for determining the surface shape are derived for both planar and curved surfaces.

TL;DR: The trade-off between expressive power and computational tractability which plagues terminological logics based on standard, two-valued semantics can be avoided while still retaining a useful and semantically supported set of subsumptions.

TL;DR: This work describes explanation patterns (XPs), knowledge structures that package the reasoning underlying explanations, using the SWALE system as a base, and discusses the retrieval and modification process, and the criteria used when deciding which explanation to accept.

TL;DR: Experimental studies are presented of analogical models, which bear a structural similarity to what is being modeled, which embody a small number of local constraints derived from naive observation, such as material continuity and noncopenetrability.

TL;DR: It is shown that first-order logic ideas can be used to construct a theorem prover for a wide class of circumscriptive theories, and the resulting system has been applied to the canonical birds flying example, to a nonseparable circumscription, and to the Yale shooting problem.

TL;DR: The approach described here unifies symbolic and numeric approaches to uncertainty management, thus facilitating dynamic construction of quantitative belief arguments, explanation of beliefs, and resolution of conflicts.

TL;DR: The PUPS theory and its ACT∗ predecessor are computational embodiments of psychology's effort to develop a theory of the origins of knowledge as discussed by the authors, which contains proposals for extraction of knowledge from the environment, a strength-based prioritization of knowledge, knowledge compilation mechanisms for forming use-specific versions of knowledge and induction mechanisms for extending knowledge.

TL;DR: It is shown that subsumption in the terminological logic of NikL is undecidable and thus that there are no complete algorithms for subsumption or classification in NIKL.

TL;DR: An improved algorithm BS∗ is described which expands significantly less nodes on average than any other algorithm in the same class of non-wave-shaping admissible bidirectional algorithms, and is the first staged search algorithm which preserves admissibility.

TL;DR: A method for responding to misconceptions in a domain-independent and context-sensitive fashion is discussed, which calls for reasoning about possible sources of the misconception using a model of the user and generating a response based on this reasoning.

TL;DR: In this paper, a new methodology for the construction of world models is presented, which makes possible the modeling of simultaneous, interactive processes, processes characterized by a continuum of gradual change, involuntarily activated processes (such as the growing of grass) and time as a continuous phenomenon.

TL;DR: One can identify four major paradigms under active investigation: inductive learning, analytic learning, genetic algorithms, and connectionist learning methods as discussed by the authors, and they can be classified into four major classes:

TL;DR: This paper presents a formalization of inheritance and recognition based on the principle of maximum entropy, and it can be shown that the proposed formalization may be realized as a massively parallel network of simple processing elements that can solve an interesting class of inherited and recognition problems extremely fast—in time proportional to the depth of the conceptual hierarchy.

TL;DR: This paper focuses on four machine discovery systems that share a number of features: the use of data-driven heuristics to constrain the search for numeric laws; a reliance on theoretical terms; and the recursive application of a few general discovery methods.