Results 11  20
of
20
Boosting probabilistic choice operators
 In Proceedings of Principles and Practices of Constraint Programming, Springer Verlag, LNCS 4741
, 2007
"... Abstract. Probabilistic Choice Operators (PCOs) are convenient tools to model uncertainty in CP. They are useful to implement randomized algorithms and stochastic processes in the concurrent constraint framework. Their implementation is based on the random selection of a value inside a finite domain ..."
Abstract

Cited by 3 (2 self)
 Add to MetaCart
Abstract. Probabilistic Choice Operators (PCOs) are convenient tools to model uncertainty in CP. They are useful to implement randomized algorithms and stochastic processes in the concurrent constraint framework. Their implementation is based on the random selection of a value inside a finite domain according to a given probability distribution. Unfortunately, the probabilistic choice of a PCO is usually delayed until the probability distribution is completely known. This is inefficient and penalizes their broader adoption in realworld applications. In this paper, we associate to PCO a filtering algorithm that prunes the variation domain of its random variable during constraint propagation. Our algorithm runs in O(n) where n denotes the size of the domain of the probabilistic choice. Experimental results show the practical interest of this approach. 1
Concurrency, Time and Constraints
 in "Proc. of the Nineteenth International Conference on Logic Programming (ICLP 2003)", LNCS
"... Abstract Concurrent constraint programming (ccp) is a model of concurrency for systems in which agents (also called processes) interact with one another by telling and asking information in a shared medium. Timed (or temporal) ccp extends ccp by allowing agents to be constrained by time requirements ..."
Abstract

Cited by 2 (0 self)
 Add to MetaCart
Abstract Concurrent constraint programming (ccp) is a model of concurrency for systems in which agents (also called processes) interact with one another by telling and asking information in a shared medium. Timed (or temporal) ccp extends ccp by allowing agents to be constrained by time requirements. The novelty of timed ccp is that it combines in one framework an operational and algebraic view based upon process calculi with a declarative view based upon temporal logic. This allows the model to benefit from two wellestablished theories used in the study of concurrency. This essay offers an overview of timed ccp covering its basic background and central developments. The essay also includes an introduction to a temporal ccp formalism called thentcc calculus. 1
Measure Transformer Semantics for Bayesian Machine Learning
"... Abstract. The Bayesian approach to machine learning amounts to inferring posterior distributions of random variables from a probabilistic model of how the variables are related (that is, a prior distribution) and a set of observations of variables. There is a trend in machine learning towards expres ..."
Abstract

Cited by 2 (2 self)
 Add to MetaCart
Abstract. The Bayesian approach to machine learning amounts to inferring posterior distributions of random variables from a probabilistic model of how the variables are related (that is, a prior distribution) and a set of observations of variables. There is a trend in machine learning towards expressing Bayesian models as probabilistic programs. As a foundation for this kind of programming, we propose a core functional calculus with primitives for sampling prior distributions and observing variables. We define combinators for measure transformers, based on theorems in measure theory, and use these to give a rigorous semantics to our core calculus. The original features of our semantics include its support for discrete, continuous, and hybrid measures, and, in particular, for observations of zeroprobability events. We compile our core language to a small imperative language that has a straightforward semantics via factor graphs, data structures that enable many efficient inference algorithms. We use an existing inference engine for efficient approximate inference of posterior marginal distributions, treating thousands of observations per second for large instances of realistic models. 1
Notes on timed ccp
 In 4th Advanced Course on Petri Nets ICPN’03. LNCS
, 2004
"... Abstract A constraint is a piece of (partial) information on the values of the variables of a system. Concurrent constraint programming (ccp) is a model of concurrency in which agents (also called processes) interact by telling and asking information (constraints) to and from a shared store (a const ..."
Abstract

Cited by 2 (1 self)
 Add to MetaCart
Abstract A constraint is a piece of (partial) information on the values of the variables of a system. Concurrent constraint programming (ccp) is a model of concurrency in which agents (also called processes) interact by telling and asking information (constraints) to and from a shared store (a constraint). Timed (or temporal) ccp (tccp) extends ccp by agents evolving over time. A distinguishing feature of tccp, is that it combines in one framework an operational and algebraic view from process algebra with a declarative view based upon temporal logic. Tccp has been widely used to specify, analyze and program reactive systems. This note provides a comprehensive introduction to the background for and central notions from the theory of tccp. Furthermore, it surveys recent results on a particular tccp calculus, ntcc, and it provides a classification of the expressive power of various tccp languages. 1
Stochastic programs and hybrid automata for (biological) modeling
 In Proceedings of CiE 2009
, 2009
"... Abstract. We present a technique to associate to stochastic programs written in stochastic Concurrent Constraint Programming a semantics in terms of a lattice of hybrid automata. The aim of this construction is to provide a framework to approximate the stochastic behavior by a mixed discrete/continu ..."
Abstract

Cited by 2 (1 self)
 Add to MetaCart
Abstract. We present a technique to associate to stochastic programs written in stochastic Concurrent Constraint Programming a semantics in terms of a lattice of hybrid automata. The aim of this construction is to provide a framework to approximate the stochastic behavior by a mixed discrete/continuous dynamics with a variable degree of discreteness. 1
A Programming Language for Probabilistic Computation
, 2005
"... As probabilistic computations play an increasing role in solving various problems, researchers have designed probabilistic languages to facilitate their modeling. Most of the existing probabilistic languages, however, focus only on discrete distributions, and there has been little effort to develop ..."
Abstract

Cited by 1 (0 self)
 Add to MetaCart
As probabilistic computations play an increasing role in solving various problems, researchers have designed probabilistic languages to facilitate their modeling. Most of the existing probabilistic languages, however, focus only on discrete distributions, and there has been little effort to develop probabilistic languages whose expressive power is beyond discrete distributions. This dissertation presents a probabilistic language, called PTP (ProbabilisTic Programming), which supports all kinds of probability distributions.
Uniform Path Selection of Feasible Paths as a Stochastic Constraint Problem
"... Automatic structural test data generation is a real challenge of Software Testing. Statistical structural testing has been proposed to address this problem. This testing method aims at building an input probability distribution to maximize the coverage of some structural criteria. Under the all path ..."
Abstract
 Add to MetaCart
Automatic structural test data generation is a real challenge of Software Testing. Statistical structural testing has been proposed to address this problem. This testing method aims at building an input probability distribution to maximize the coverage of some structural criteria. Under the all paths testing objective, statistical structural testing aims at selecting each feasible path of the program with the same probability. In this paper, we propose to model a uniform path selector of feasible paths as a stochastic constraint program. Stochastic constraint programming is an interesting framework which combines stochastic decision problem and constraint solving. This paper reports on the translation of uniform feasible path selection problem into a stochastic constraint problem. An implementation which uses the library PCC(FD) of SICStus Prolog designed for this problem is detailed. First experimentations, conducted over a few academic examples, show the interest of our approach. 1
Bayesian Machine Learning
, 2011
"... Abstract. The Bayesian approach to machine learning amounts to inferring posterior distributions of random variables from a probabilistic model of how the variables are related (that is, a prior distribution) and a set of observations of variables. There is a trend in machine learning towards expres ..."
Abstract
 Add to MetaCart
Abstract. The Bayesian approach to machine learning amounts to inferring posterior distributions of random variables from a probabilistic model of how the variables are related (that is, a prior distribution) and a set of observations of variables. There is a trend in machine learning towards expressing Bayesian models as probabilistic programs. As a foundation for this kind of programming, we propose a core functional calculus with primitives for sampling prior distributions and observing variables. We define combinators for measure transformers, based on theorems in measure theory, and use these to give a rigorous semantics to our core calculus. The original features of our semantics include its support for discrete, continuous, and hybrid measures, and, in particular, for observations of zeroprobability events. We compile our core language to a small imperative language that in addition to the measure transformer semantics also has a straightforward semantics via factor graphs, data structures that enable many efficient inference algorithms. We use an existing inference engine for efficient approximate inference of posterior marginal distributions, treating thousands of observations per second for large instances of realistic models. 1