Pierre Bessiere, Olivier Lebeltel, Olivier Lebeltel et al. · 2000
We propose a new method to program robots based on Bayesian inference and learning. The capacities of this programming method are demonstrated through a succession of increasingly complex experiments. Starting from the learning of simple reactive behaviors, we present instances of behavior combinations, sensor fusion, hierarchical behavior composition, situation recognition and temporal sequencing. This series of experiments comprises the steps in the incremental development of a complex robot program. The advantages and drawbacks of this approach are discussed along with these different experiments and summed up as a conclusion. These different robotics programs may be seen as an illustration of probabilistic programming applicable whenever one must deal with problems based on uncertain or incomplete knowledge. The scope of possible applications is obviously much broader than robotics.
Cited by 0
Evandro Nunes Regolin, Aurora Pozo · Lecture notes in computer science · 2005
Cited by 26
Ulrich Rieder · Advances in Applied Probability · 1975
We consider a non-stationary Bayesian dynamic decision model with general state, action and parameter spaces. It is shown that this model can be reduced to a non-Markovian (resp. Markovian) decision model with completely known transition probabilities. Under rather weak convergence assumptions on the expected total rewards some general results are presented concerning the restriction on deterministic generalized Markov policies, the criteria of optimality and the existence of Bayes policies. These facts are based on the above transformations and on results of Hindererand Schäl.
Cited by 8
Olivier Lebeltel, Pierre Bessiere, Julien Diard et al. · 2000
We propose a new method to program robots based on Bayesian inference and learning. The capacities of this programming method are demonstrated through a succession of increasingly complex experiments. Starting from the learning of simple reactive behaviors, we present instances of behavior combinations, sensor fusion, hierarchical behavior composition, situation recognition and temporal sequencing. This series of experiments comprises the steps in the incremental development of a complex robot program. The advantages and drawbacks of this approach are discussed along with these different experiments and summed up as a conclusion. These different robotics programs may be seen as an illustration of probabilistic programming applicable whenever one must deal with problems based on uncertain or incomplete knowledge. The scope of possible applications is obviously much broader than robotics. 1.
Cited by 0
Gilles Barthe, Gian Pietro Farina, Marco Gaboardi et al. · 2016
We present PrivInfer, an expressive framework for writing and verifying differentially private Bayesian machine learning algorithms. Programs in PrivInfer are written in a rich functional probabilistic programming language with constructs for performing Bayesian inference. Then, differential privacy of programs is established using a relational refinement type system, in which refinements on probability types are indexed by a metric on distributions. Our framework leverages recent developments in Bayesian inference, probabilistic programming languages, and in relational refinement types. We demonstrate the expressiveness of PrivInfer by verifying privacy for several examples of private Bayesian inference.
Cited by 23Open access
João Filipe Ferreira, Jorge Dias · Springer tracts in advanced robotics · 2013
Cited by 0
MorawskiP. · 1989
Cited by 7
2013
Cited by 0
Scott Sanner, Robby Goetschalckx, Kurt Driessens et al. · Lirias (KU Leuven) · 2009
Real-time dynamic programming (RTDP) solves Markov decision processes (MDPs) when the initial state is restricted, by focusing dynamic programming on the envelope of states reachable from an initial state set. RTDP often provides performance guarantees without visiting the entire state space. Building on RTDP, recent work has sought to improve its efficiency through various optimizations, including maintaining upper and lower bounds to both govern trial termination and prioritize state exploration. In this work, we take a Bayesian perspective on these upper and lower bounds and use a value of perfect information (VPI) analysis to govern trial termination and exploration in a novel algorithm we call VPI-RTDP. VPI-RTDP leads to an improvement over state-of-the-art RTDP methods, empirically yielding up to a three-fold reduction in the amount of time and number of visited states required to achieve comparable policy performance. 1
Cited by 40Open access
Ugo Dal Lago, Naohiko Hoshino · SPIRE - Sciences Po Institutional REpository · 2019
We give a geometry of interaction model for a typed λ -calculus endowed with operators for sampling from a continuous uniform distribution and soft conditioning, namely a paradigmatic calculus for higher-order Bayesian programming. The model is based on the category of measurable spaces and partial measurable functions, and is proved adequate with respect to both a distribution-based and a sampling-based operational semantics.
Cited by 8Open access
Pierre Bessìère, Olivier Lebeltel · Springer tracts in advanced robotics · 2008
Cited by 6
Stephen Muggleton · 2015
Inductive Logic Progrdng (ILP) involves the construction of first-order definite clause theories from examples and background knowledge, Un-like both traditional Machine Learning and Com-putational Learning Theory, ILP is based on lock-step development of Theory, Implementations and Applications. ILP systems have successful appli-cations in the learning of structure-activity rules for drug design, semantic grammars rules, finite element mesh design rules and rules for prediction of protein structure and mutagenic molecules. The strong applications in ILP can be contrasted with relatively weak PAC-learning results (even highly-restricted forms of logic programs are known to be prediction-hard). It has been recently argued that the mismatch is due to distributional assump-tions made in application domains. These assump-tions can be modelled as a Bayesiau prior prob-ability representing subjective degrees of belief. Other authors have argued for the use of Bayesian prior distributions for reasons different to those here, though this has not lead to a new model of polynomial-time learnability. Incorporation of Bayesian prior distributions over time-bounded hy-potheses in PAC leads to a new model called U-Learnability. It is argued that U-learnability is more appropriate than PAC for Universal (Tur-ing computable) languages. Time-bounded logic programs have been shown to be polynomially U-learnable under certain distributions. The use of time-bounded hypotheses enforces decidability and allows a unified characterisation of speed-up learning and inductive learning. U-learnability has as special cases PAC and Natarajan’s model of speed-up learning. Permission to copy without fee all or part of this material is granted provided that the copies are not made or distributed for direct commercial advantage, the ACM copyright notice and the title of the publication and its date appear, and notice is given
Cited by 0
GARCIA RAMIREZ, Ruben Senen, MAZER, Emmanuel · OpenGrey (Institut de l'Information Scientifique et Technique) · 2003
Cette thèse porte sur la programmation bayesienne de bras manipulateurs équipés d'un système de vision stéréoscopique. L'implantation d'une tâche de prise et de dépose d'un objet est donnée en exemple pour évaluer l'approche. Cette tâche met en jeu des modèles géométriques du bras, du système de vision stéréoscopique et des objets manipulés. L'incertitude associée aux modèles géométriques, aux capteurs des robots et au système de vision est prise en compte.La méthode de programmation est formelle et systématique. Elle consiste en trois parties : La description où sont définies les variables pertinentes et la distribution de probabilité conjointe. L'inférence qui vise à construire une distribution de probabilité sur les commandes motrices sachant les informations sensorielles, et enfin le choix d'une commande à partir de la distribution précédente. Une nouvelle méthode de programmation de robot dite " inverse " est introduite. Des méthodes de résolutions approchées sont présentées visant à combattre la complexité du calcul liée à cette approche. Une expérimentation qualitative du système de programmation est présentée.
Cited by 0
Byoung-tak Zhang · 2000
. A Bayesian framework for genetic programming GP is presented. This is motivated by the observation that genetic programming iteratively searches populations of fitter programs and thus the information gained in the previous generation can be used in the next generation. The Bayesian GP makes use of Bayes theorem to estimate the posterior distribution of programs from their prior distribution and likelihood for the fitness data observed. Offspring programs are then generated by sampling from the posterior distribution by genetic variation operators. We present two GP algorithms derived from the Bayesian GP framework. One is the genetic programming with the adaptive Occam's Z. razor AOR designed to evolve parsimonious programs. The other is the genetic programming with Z. incremental data inheritance IDI designed to accelerate evolution by active selection of fitness cases. A multiagent learning task is used to demonstrate the effectiveness of the presented methods. In a series of experiments, AOR reduced solution complexity by 20% and IDI doubled evolution speed, both without loss of solution accuracy. Keywords: Bayesian genetic programming, probabilistic evolution, adaptive Occam's razor, incremental data inheritance, parsimony pressure, data subset selection 1.
Cited by 0
2008
IP has been used to learn various kinds of statistical models from given data including multivariate mixture models (clustering), segmentation models [10],
Cited by 0
Pierre Bessìère, Juan-Manuel Ahuactzin, Emanuel Mazer et al. · HAL (Le Centre pour la Communication Scientifique Directe) · 2008
Cited by 0
Int Workshop On, O. Lebeltel, J. Diard et al. · 2008
We propose an original method for programming robots based on bayesian inference and learning. This method formally deals with problems of uncertainty and incomplete information that are inherent to the field. Indeed, the principal difficulties of robot programming comes from the unavoidable incompleteness of the models used. We present the formalism for describing a robotic task as well as the resolution methods. This formalism is inspired by the theory of probability, suggested by the physicist E T Jaynes: "Probability as Logic"[1]. Learning and maximum entropy principle translate incompleteness into uncertainty. Bayesian inference offers a formal framework for reasoning with this uncertainty. The main contribution of this paper is the definition of a generic system of robotic programming and its experimental application. We illustrate it by programming a surveillance task with a mobile robot: the Khepera. In order to do this, we use generic programming resources called "descriptions". We show how to define and use these resources in an incremental way (reactive behaviors, sensor fusion, situation recognition and sequences of behaviors) within a systematic and unified framework.
Cited by 0
Carla Koike A, Cédric Pradalier, Pierre Bessière et al. · 2003
bInstitut National de Recherche en Informatique et en Automatique. cLab. Graphisme, Vision et Robotique. Abstract — Unexpected events and not modeled properties of the robot environment are some of the challenges presented by situated robotics research field. Collision avoidance is a basic security requirement and this paper proposes a probabilistic approach called Bayesian Programming, which aims to deal with the uncertainty, imprecision and incompleteness of the information handled to solve the obstacle avoidance problem. Some examples illustrate the process of embodying the programmer preliminary knowledge into a Bayesian program and experimental results of these examples implementation in an electrical vehicle are described and commented. A video illustration of the developed experiments can be found at
Cited by 0
Ruwantha Perera, Dasindu Dilshan · 2024
Cited by 0Open access
Ilya O. Ryzhov, Martijn Mes, Warren B. Powell et al. · Operations Research · 2019
Approximate dynamic programming has been applied to solve large-scale resource allocation problems in many domains, including transportation, energy, and healthcare. The main algorithmic challenge in these applications is to accurately estimate the value of a resource in a certain state (for example, a battery that is half full, or a truck en route to a certain destination). This “value function” must be learned as part of the decision-making process, giving rise to the “exploration/exploitation” challenge: we may, and often should, choose to experiment with seemingly suboptimal actions because they have high potential to be better than we believe. In “Bayesian Exploration for Approximate Dynamic Programming,” Ilya O. Ryzhov, Martijn R. K. Mes, Warren B. Powell and Gerald van den Berg present a principled framework for modeling uncertainty about the value function and measuring the potential of a state to improve a decision-making policy. This method is scalable and performs well in experiments.
Cited by 7Open access