作者: G Roup , Laplace R Esearch
DOI:
关键词: Probabilistic logic 、 Bayesian network 、 Machine learning 、 Deliverable 、 Bayesian programming 、 Inference 、 Sensor fusion 、 Artificial intelligence 、 Kalman filter 、 Computer science 、 Bayesian probability
摘要: The purpose of this paper is to make a state the art on probabilistic methodology and techniques for artefact conception development. It 8th deliverable BIBA (Bayesian Inspired Brain Artefact) project. We first present incompletness problem as central difficulty that both living creatures artefacts have face: how can they perceive, infer, decide act efficiently with incomplete uncertain knowledge?. then introduce generic formalism called Bayesian Programming . This used review main techniques. organized in 3 parts: models from networks Kalman filters sensor fusion CAD systems, second inference finally learning model acquisition comparison methodologies. conclude perspectives project rise art. 1. P IERRE B ESSIERE , J UAN -M ANUEL A HUACTZIN O LIVIER YCARD D AVID ELLOT F RANCIS C OLAS HRISTOPHE OUE ULIEN IARD R UBEN G ARCIA ARLA K OIKE L EBELTEL ONAN E H Y OL IVIER M ALRAIT MMANUEL AZER AMEL EKHNACHA EDRIC RADALIER NNE S PALANZANI