Monotonicity Extraction for Monotonic Bayesian Networks Parameter Learning

We propose a logical/mathematical framework for statistical parameter learning of parameterized logic programs, i.e. definite clause programs containing probabilistic facts with a parameterized distribution. It extends the traditional least Herbrand model semantics in logic programming to distribution semantics, possible world semantics with a probability distribution which is unconditionally applicable to arbitrary logic programs including ones for HMMs, PCFGs and Bayesian networks. We also propose a new EM algorithm, the graphical EM algorithm, that runs for a class of parameterized logic programs representing sequential decision processes where each decision is exclusive and independent. It runs on a new data structure called support graphs describing the logical relationship between observations and their explanations, and learns parameters by computing inside and outside probability generalized for logic programs. The complexity analysis shows that when combined with OLDT search for all explanations for observations, the graphical EM algorithm, despite its generality, has the same time complexity as existing EM algorithms, i.e. the Baum-Welch algorithm for HMMs, the Inside-Outside algorithm for PCFGs, and the one for singly connected Bayesian networks that have been developed independently in each research field. Learning experiments with PCFGs using two corpora of moderate size indicate that the graphical EM algorithm can significantly outperform the Inside-Outside algorithm.

Download Full-text

Compiling Bayesian Networks for Parameter Learning Based on Shared BDDs

AI 2011: Advances in Artificial Intelligence - Lecture Notes in Computer Science ◽

10.1007/978-3-642-25832-9_21 ◽

2011 ◽

pp. 203-212 ◽

Cited By ~ 1

Author(s):

Masakazu Ishihata ◽

Taisuke Sato ◽

Shin-ichi Minato

Keyword(s):

Bayesian Networks ◽

Parameter Learning

Download Full-text

Recognition of Human Intentions Using Bayesian Artificial Intelligence

Volume 7: Engineering Education and Professional Development ◽

10.1115/imece2007-43325 ◽

2007 ◽

Cited By ~ 1

Author(s):

R. Tse ◽

G. Seet ◽

S. K. Sim

Keyword(s):

Artificial Intelligence ◽

Maximum Likelihood ◽

Bayesian Networks ◽

Success Rate ◽

Dynamic Bayesian Networks ◽

Laser Sensor ◽

Parameter Learning ◽

Complete Set ◽

Bayesian Artificial Intelligence

Controlling a robot to perform a task is more difficult than commanding a human. A robot needs to be preprogrammed to perform a task. This is achieved by providing the robot with a complete set of step-by-step commands from the beginning till the end. In contrast, to a human, recalling an experience when he was instructed with the same command in a similar situation, a human would be able to guess what intention behind such a command is and could then behave cooperatively. Our objective is to equip the robot with such a capability of recognizing some simple human intentions required of a robot, such as: moving around a corner, moving parallel to the wall, or moving towards an object. The cues used by the robot to make an inference were: the odometer and laser sensor readings, and the human operator’s commands given. Using the Maximum-Likelihood (ML) parameter learning on Dynamic Bayesian Networks, the correlations between these cues and the intentions were modeled and used to infer the human intentions in controlling the robot. From the experiments, the robot was able to learn and infer the above mentioned intentions of the human user with a satisfying success rate.

Download Full-text