Human reaction times during sensory-motor tasks vary considerably.

Hierarchical learning machines are non-regular and non-identifiable statistical models, whose true parameter sets are analytic sets with singularities. Using algebraic analysis, we rigorously prove that the stochastic complexity of a non-identifiable learning machine is asymptotically equal to '1 log n - (ml - 1) log log n

Fishers linear discriminant analysis (LDA) is a classical multivariate technique both for dimension reduction and classification. The data vectors are transformed into a low dimensional subspace such that the class centroids are spread out as much as possible. In this subspace LDA works as a simple prototype classifier with linear decision boundaries. However, in many applications the linear boundaries do not adequately separate the classes. We present a nonlinear generalization of discriminant analysis that uses the kernel trick of representing dot products by kernel functions.

We analyze the conditions under which synaptic learning rules based by learning rules basedon action potential timing can be approximated of plasticity in whichon firing rates. In particular, we consider a form synapses depress when a presynaptic spike is followed by a postsynaptic differentialspike, and potentiate with the opposite temporal ordering.

Recently,a number of authorshave proposedtreating dialogue systems as Markov decision processes(MDPs). However,the practicalapplicationofMDP algorithms to dialogue systems faces a numberof severe technicalchallenges.We have built a general software tool (RLDS, for ReinforcementLearning for Dialogue Systems) on the MDP framework, and have applied it to dialogue corpora gatheredbased from two dialoguesystemsbuilt at AT&T Labs. Our experimentsdemonstratethat RLDS holds promise as a tool for "browsing" and understandingcorrelationsin complex, temporallydependentdialogue corpora.

We provide an analysis of the turbo decoding algorithm (TDA) setting involving Gaussian densities. In this context, we arein a able to show that the algorithm converges and that - somewhat the density generated by the TDA may differsurprisingly - though significantly from the desired posterior density, the means of these two densities coincide.

Decision theory formally solves the problem of rational agents in uncertain worlds if the true environmental probability distribution is known. Solomonoff's theory of universal induction formally solves the problem of sequence prediction for unknown distribution. We unify both theories and give strong arguments that the resulting universal AIXI model behaves optimal in any computable environment. The major drawback of the AIXI model is that it is uncomputable. To overcome this problem, we construct a modified algorithm AIXI^tl, which is still superior to any other time t and space l bounded agent. The computation time of AIXI^tl is of the order t x 2^l.

In this article, we present a last-minute travel application as part of a complete virtual travel agency. Each year, a significant amount of tour packages are sold as last minute tours in Germany. It is impossible for a travel agent to keep track of all the offered tour packages. Electronic-commerce applications might present the best possible tour package for a specific customer request. Traditional database-driven applications, as used by most of the tour operators, are not sufficient enough to implement a sales process with consultation on the World Wide Web. The last-minute travel application presented here uses case-based reasoning to bridge this gap and simulate the sales assistance of a human travel agent. A case retrieval net, as an internal data structure, proved to be efficient in handling the large amount of data. Important for the acceptance by customers is also the integration into the virtual travel agency and the interconnections to other parts of this system, such as background information or the online car rental application.

The researcher will be numerous international societies and IJCAI-01/IJCAI-03 Update. Planning required to sign a statement of restricted conferences. Barbara Grosz agreed to for IJCAI-01 is well under way.

There is a fundamental mismatch between the computational basis of spreadsheets and our knowledge of the real world. In spreadsheets, numeric data are represented as exact numbers and their mutual relations as functions, whose values (output) are computed from given argument values (input). However, in the real world, data are often inexact and uncertain in many ways, and the relationships, that is, constraints, between input and output are far more complicated. This article shows that interval constraint solving, an emerging AI-based technology, provides a more versatile and useful foundation for spreadsheets. The new computational basis is 100-percent downward compatible with the traditional spreadsheet paradigm. The idea has been successfully integrated with Microsoft excel as the add-in interval solver that seamlessly upgrades the arithmetic core of excel into interval constraint solving. The product has been downloaded by thousands of end users all over the world and has been used in various applications in business computing, engineering, education, and science. There is an intriguing chance for a major breakthrough of the AI technology on the spreadsheet platform: Tens of millions of excel users are making important decisions based on spreadsheet calculations.