Showing papers in "Artificial Intelligence in 1994"

TL;DR: The transferable belief model is described, a model for representing quantified beliefs based on belief functions that can be held at two levels: a credal level where beliefs are entertained and quantified by belief functions, and a pignisticlevel where beliefs can be used to make decisions and are quantification by probability functions.

TL;DR: For these types of restrictions, it is shown when planning is tractable (polynomial) and intractable (NP-hard) and PSPACE-complete to determine if a given planning instance has any solutions.

TL;DR: These tests show that the networks created by KBANN generalize better than a wide variety of learning systems, as well as several techniques proposed by biologists.

TL;DR: Five algorithms that identify a subset of features sufficient to construct a hypothesis consistent with the training examples are presented and it is shown that any learning algorithm implementing the MIN-FEATURES bias requires ⊖(( ln ( l δ ) + [2 p + p ln n])/e) training examples to guarantee PAC-learning a concept having p relevant features out of n available features.

TL;DR: A class of ML systems which use a hierarchy of first-order languages, each language containing names for the language below, are introduced and it is proved that the set of theorems of the most common modal logics can be embedded into that of the corresponding ML systems.

TL;DR: It is shown that finding the maximum a-posteriori probability (MAP) instantiation of all the random variables given the evidence is NP-hard in the general case when graph representations are used, even if the size of the representation happens to be linear in n.

TL;DR: A completely automated approach to generating abstractions for planning using a tractable, domain-independent algorithm whose only input is the definition of a problem to be solved and whose output is an abstraction hierarchy that is tailored to the particular problem.

TL;DR: In this article, the authors proposed a new algorithm, AC-6, which keeps the optimal worst-case time complexity of AC-4 while working out the drawback of space complexity.

TL;DR: This paper investigates the planning of assembly algorithms specifying (dis) assembly operations on the components of a product and the ordering of these operations and presents measures to evaluate the complexity of these algorithms and techniques to estimate the inherent complexity of aproduct.

TL;DR: It is shown how nonmonotonic inferences may elegantly be interpreted in terms of underlying expectations, and it is shown that by using the notion of expectation, one can unify the treatment of the theory of belief revision and that of nonMonotonic inference relations.

TL;DR: An approach to designing systems that are capable of taking their own computational resources into consideration during planning and problem solving by using expectations about the performance of decision-making procedures and preferences over the outcomes resulting from applying those procedures.

TL;DR: This paper connects both simulated and real robots to Alecsys, a parallel implementation of a learning classifier system with an extended genetic algorithm to demonstrate that classifier systems with genetic algorithms can be practically employed to develop autonomous agents.

TL;DR: A family of modal logics is presented in which a conditional connective is defined for statements of normality and its properties are examined, and it is demonstrated that two of the most important conditional approaches are equivalent to fragments of the authors' conditional logics ofnormality.

TL;DR: An operational definition of causal ordering is presented that allows us to extract causal dependency relations among variables implicit in a model of a system, when a model is represented as a set of acausal, mathematical relations.

TL;DR: This paper proposes to model the uncertainty due to noise, e.g. the error in an object's position, by conventional covariance matrices, independent of the sensing modality, being applicable to most temporal data association problems.

TL;DR: It is proved that a constraint satisfaction problem may be decomposed into a number of subproblems precisely when the corresponding hypergraph satisfies a simple condition.

TL;DR: A comprehensive system for automatic theory (knowledge base) refinement combining analytical and empirical methods that applies to classification tasks employing a propositional Hornclause domain theory.

TL;DR: The paper focuses particularly on default logic, circumscription, and the author's own argument-based approach to defeasible reasoning, which is able to differentiate, in a congenial way, between cases having the structure of the lottery paradox and cases having a similar structure to that of the preface.

TL;DR: Positive PAC-learning results for the nonmonotonic inductive logic programming setting are presented and first-order range-restricted clausal theories that consist of clauses with up to k literals of size at most j each are shown to be PAC-learnable with one-sided error from positive examples only.

TL;DR: This paper characterizes the types of domains that offer performance differentiation and the features that distinguish the relative overhead of three planning algorithms, arguing that the observed performance differences are best understood with an extension of Korf's taxonomy of subgoal collections.

TL;DR: The distribution of hard graph coloring problems as a function of graph connectivity is shown to have two distinct transition behaviors, including a peak in the median search cost near the connectivity at which half the graphs have solutions.

TL;DR: Pn-search has been used to establish the game-theoretical values of Connect-Four, Qubic, and Go-Moku and was able to find a forced win for the player to move first.

TL;DR: A technique for analyzing the behavior of sophisticated AI search programs working on realistic, large-scale problems is introduced and it is suggested that this type of analysis can be generalized to other kinds of AI problems.

TL;DR: It is demonstrated that problem classes and regions of the phase transition previously thought to be easy can sometimes be orders of magnitude more difficult than the worst problems in problem classesand regions ofThe phase transition considered hard.

TL;DR: A set of constraints is identified which gives rise to a class of tractable problems and given polynomial time algorithms for solving such problems, and it is proved that the class of problems generated by any set of constraint not contained in this restricted set is NP-complete.

TL;DR: The deduction method presented here contrasts with other methods whose ability to perform logical reasoning is either limited or requires finding all truth assignments consistent with the given sentences.

TL;DR: A simplified version of that propositional nonmonotonic logic is described, it is shown how quantifiers can be included in it, and its relation to circumscription and default logic, to logic programming, and to the theory of epistemic queries developed by Hector Levesque and Ray Reiter is studied.

TL;DR: The case when the DRP fails is examined, and an analytical model of search complexity parameterized by the probability of an abstract solution being refinable is provided, which provides a more accurate picture of the effectiveness of hierarchical problem solving.

TL;DR: The semantics of cost-based abduction for complete models are then generalized to handle negation, which allows the best-first search algorithm to apply as a novel way of computing MAP assignments to belief networks that can enumerate assignments in order of decreasing probability.

TL;DR: A number of variants of default logic are developed to address differing intuitions arising from the original and subsequent formulations, and it is argued that in some situations the requirement of proving the antecedent of a default is too strong.