We describe an algorithm and representation level theory of illusory contour shape and salience. Unlike previous theories, our model is derived from a single assumption--- namely, that the prior probability distribution of boundary completion shape can be modeled by a random walk in a lattice whose points are positions and orientations in the image plane (i.e., the space which one can reasonably assume is represented by neurons of the mammalian visual cortex). Our model does not employ numerical relaxation or other explicit minimization, but instead relies on the fact that the probability that a particle following a random walk will pass through a given position and orientation on a path joining two boundary fragments can be computed directly as the product of two vector-field convolutions. We show that for the random walk we define, the maximum likelihood paths are curves of least energy, that is, on average, random walks follow paths commonly assumed to model the shape of illusory co...

In this paper we introduce a new class of labelled transition systems - Labelled Markov Processes - and define bisimulation for them.

Principal curves have been defined as ``self consistent'' smooth curves which pass through the ``middle'' of a $d$-dimensional probability distribution or data cloud. They give a summary of the data and also serve as an efficient feature extraction tool. We take a new approach by defining principal curves as continuous curves of a given length which minimize the expected squared distance between the curve and points of the space randomly chosen according to a given distribution. The new definition makes it possible to theoretically analyze principal curve learning from training data and it also leads to a new practical construction. Our theoretical learning scheme chooses a curve from a class of polygonal lines with $k$ segments and with a given total length, to minimize the average squared distance over $n$ training points drawn independently. Convergence properties of this learning scheme are analyzed and a practical version of this theoretical algorithm is implemented. In each iteration of the algorithm a new vertex is added to the polygonal line and the positions of the vertices are updated so that they minimize a penalized squared distance criterion. Simulation results demonstrate that the new algorithm compares favorably with previous methods both in terms of performance and computational complexity, and is more robust to varying data models.

We consider policy evaluation algorithms within the context of infinite-horizon dynamic programming problems with discounted cost. We focus on discrete-time dynamic systems with a large number of states, and we discuss two methods, which use simulation, temporal differences, and linear cost function approximation. The first method is a new gradient-like algorithm involving least-squares subproblems and a diminishing stepsize, which is based on the #-policy iteration method of Bertsekas and Ioffe. The second method is the LSTD(#) algorithm recently proposed by Boyan, which for # =0coincides with the linear least-squares temporal-difference algorithm of Bradtke and Barto. At present, there is only a convergence result by Bradtke and Barto for the LSTD(0) algorithm. Here, we strengthen this result by showing the convergence of LSTD(#), with probability 1, for every # [0, 1].

. The nonparametric and the nuisance parameter approaches to consistently testing statistical models are both attempts to estimate topological measures of distance between a parametric and a nonparametric fit, and neither dominates in experiments. This topological unification allows us to greatly extend the nuisance parameter approach. How and why the nuisance parameter approach works and how it can be extended bears closely on recent developments in artificial neural networks. Statistical content is provided by viewing specification tests with nuisance parameters as tests of hypotheses about Banach-valued random elements and applying the Banach Central Limit Theorem and Law of Iterated Logarithm, leading to simple procedures that can be used as a guide to when computationally more elaborate procedures may be warranted. 1. Introduction In testing whether or not a parametric statistical model is correctly specified, there are a number of apparently distinct approaches one might take. T...

Abstract. We consider the problem of automatically verifying realtime systems with continuously distributed random delays. We generalise probabilistic timed automata introduced in [19], an extension of the timed automata model of [4], with clock resets made according to continuous probability distributions. Thus, our model exhibits nondeterministic and probabilistic choice, the latter being made according to both discrete and continuous probability distributions. To facilitate algorithmic verification, we modify the standard region graph construction by subdividing the unit intervals in order to approximate the probability to within an interval. We then develop a model checking method for continuous probabilistic timed automata, taking as our specification language Probabilistic Timed Computation Tree Logic (PTCTL). Our method improves on the previously known techniques in that it allows the verification of quantitative probability bounds, as opposed to qualitative properties which can only refer to bounds of probability 0 or 1. 1

) Vineet Gupta Radha Jagadeesan Prakash Panangaden y vgupta@mail.arc.nasa.gov radha@cs.luc.edu prakash@cs.mcgill.ca Caelum Research Corporation Dept. of Math. and Computer Sciences School of Computer Science NASA Ames Research Center Loyola University--Lake Shore Campus McGill University Moffett Field CA 94035, USA Chicago IL 60626, USA Montreal, Quebec, Canada Abstract This paper describes a stochastic concurrent constraint language for the description and programming of concurrent probabilistic systems. The language can be viewed both as a calculus for describing and reasoning about stochastic processes and as an executable language for simulating stochastic processes. In this language programs encode probability distributions over (potentially infinite) sets of objects. We illustrate the subtleties that arise from the interaction of constraints, random choice and recursion. We describe operational semantics of these programs (programs are run by sampling random choices), deno...

This paper therefore builds on [3], in which a model of consistent aspirations-based learning was introduced

We generalize the notion of nuclear maps from functional analysis by defining nuclear ideals in tensored -categories. The motivation for this study came from attempts to generalize the structure of the category of relations to handle what might be called "probabilistic relations". The compact closed structure associated with the category of relations does not generalize directly, instead one obtains nuclear ideals. Most tensored -categories have a large class of morphisms which behave as if they were part of a compact closed category, i.e. they allow one to transfer variables between the domain and the codomain. We introduce the notion of nuclear ideals to analyze these classes of morphisms. In compact closed tensored -categories, all morphisms are nuclear, and in the tensored -category of Hilbert spaces, the nuclear morphisms are the Hilbert-Schmidt maps. We also introduce two new examples of tensored -categories, in which integration plays the role of composition. In the first, mor...

We present a general approach to model selection and regularization that exploits unlabeled data to adaptively control hypothesis complexity in supervised learning tasks. The idea is to impose a metric structure on hypotheses by determining the discrepancy between their predictions across the distribution of unlabeled data. We show how this metric can be used to detect untrustworthy training error estimates, and devise novel model selection strategies that exhibit theoretical guarantees against over-tting (while still avoiding under- tting). We then extend the approach to derive a general training criterion for supervised learning|yielding an adaptive regularization method that uses unlabeled data to automatically set regularization parameters. This new criterion adjusts its regularization level to the specic set of training data received, and performs well on a variety of regression and conditional density estimation tasks. The only proviso for these methods is that s...