We investigate a family of inference problems on Markov models, where many sample paths are drawn from a Markov chain and partial information is revealed to an observer who attempts to reconstruct the sample paths. We present algorithms and hardness results for several variants of this problem which arise by revealing different information to the observer and imposing different requirements for the reconstruction of sample paths. Our algorithms are analogous to the classical Viterbi algorithm for Hidden Markov Models, which finds the single most probable sample path given a sequence of observations. Our work is motivated by an important application in ecology: inferring bird migration paths from a large database of observations.

We investigate a family of inference problems on Markov models, where many sample paths are drawn from a Markov chain and partial information is revealed to an observer who attempts to reconstruct the sample paths. We present algorithms and hardness results for several variants of this problem which arise by revealing different information to the observer and imposing different requirements for the reconstruction of sample paths. Our algorithms are analogous to the classical Viterbi algorithm for Hidden Markov Models, which finds the single most probable sample path given a sequence of observations. Our work is motivated by an important application in ecology: inferring bird migration paths from a large database of observations.

Natural language generation (NLG) systems are computer software systems that produce texts in English and other human languages, often from non-linguistic input data. NLG systems, like most AI systems, need substantial amounts of knowledge. However, our experience in two NLG projects suggests that it is difficult to acquire correct knowledge for NLG systems; indeed, every knowledge acquisition (KA) technique we tried had significant problems. In general terms, these problems were due to the complexity, novelty, and poorly understood nature of the tasks our systems attempted, and were worsened by the fact that people write so differently. This meant in particular that corpus-based KA approaches suffered because it was impossible to assemble a sizable corpus of high-quality consistent manually written texts in our domains; and structured expert-oriented KA techniques suffered because experts disagreed and because we could not get enough information about special and unusual cases to build robust systems. We believe that such problems are likely to affect many other NLG systems as well. In the long term, we hope that new KA techniques may emerge to help NLG system builders. In the shorter term, we believe that understanding how individual KA techniques can fail, and using a mixture of different KA techniques with different strengths and weaknesses, can help developers acquire NLG knowledge that is mostly correct.

Speedup learning seeks to improve the computational efficiency of problem solving with experience. In this paper, we develop a formal framework for learning efficient problem solving from random problems and their solutions. We apply this framework to two different representations of learned knowledge, namely control rules and macro-operators, and prove theorems that identify sufficient conditions for learning in each representation. Our proofs are constructive in that they are accompanied with learning algorithms. Our framework captures both empirical and explanation-based speedup learning in a unified fashion. We illustrate our framework with implementations in two domains: symbolic integration and Eight Puzzle. This work integrates many strands of experimental and theoretical work in machine learning, including empirical learning of control rules, macro-operator learning, Explanation-Based Learning (EBL), and Probably Approximately Correct (PAC) Learning.

Because of the complexity of the application and the large amount of data, the problem cannot be solved by using a single method. The solution we propose is to build multimodules NN architectures where several NN cooperate together. Such system suffer from generic problem for whom we propose solutions. They allow to reach accurate performances for multi-valued function approximations and probability estimations. The results are compared with six other methods which have been used for this problem. We show that the methodology we have developed is general and can be used for a large variety of applications.

Because of the complexity of the application and the large amount of data, the problem cannot be solved by using a single method. The solution we propose is to build multimodules NN architectures where several NN cooperate together. Such system suffer from generic problem for whom we propose solutions. They allow to reach accurate performances for multi-valued function approximations and probability estimations. The results are compared with six other methods which have been used for this problem. We show that the methodology we have developed is general and can be used for a large variety of applications.

The Department of Computer Science at Louisiana State University (LSU) has been involved in robotics research since 1992 when the Robotics Research Laboratory (RRL) was established as a research and teaching program specializing in autonomous mobile robots (AMRS). Researchers at RRL are conducting high-quality research in amrs with the goal of identifying the computational problems and the types of knowledge that are fundamental to the design and implementation of autonomous mobile robotic systems. In this article, we overview the projects that are currently under way at LSU's RRL.

This article picks up the call for a reflective examination of the prevailing computational metaphor of AI (and philosophical presuppositions behind it) by sketching alternatives that might serve as seeds for discussion-specifically, the seven alternatives introduced in our previous article (see "AI Magazine, spring 1991). The relative strengths and weaknesses of the alternatives are contrasted with those of the computational metaphor.

It (McDermott 1988), and the idea of generic also helps to explicitly focus on how to go tasks and task-specific architectures (Chandrasekaran from the knowledge level to the symbol or 1983). These various proposals are program level. I call this in-between level the obviously related to each other, which makes knowledge-use level. At the knowledge-use it desirable to construct a synthesis that combines level, we focus on issues such as how the their strengths. Such a synthesis is presented overall task will be decomposed into manageable here in the form of a componential subtasks, what ordering will be imposed framework. The framework stresses modularity on the tasks, what kind of access to knowledge and consideration of the pragmatic constraints will be needed (and, consequently, what of the domain.

The initial results from a few AI research projects in statistics have been quite interesting to statisticians: Feasibility demonstration systems have been built at Stanford University, AT-T bell Laboratories, and the University of Edinburgh. Several more design studies have been completed. A conference devoted to expert systems in statistics was sponsored by the Royal Statistical Society. On the other hand, statistic as a domain may be of particular interest to AI researchers, for it offers both tasks well suited to current AI capabilities and tasks requiring development of new AI techniques.