http://www.syntheticsapien.com/ 2012-05-27T20:47:52+00:00 http://www.syntheticsapien.com/ http://www.syntheticsapien.com/lib/tpl/roundbox/images/favicon.ico text/html 2011-02-18T06:15:15+00:00 http://www.syntheticsapien.com/doku.php/goplayer:bnstatespace?rev=1298009715&do=diff Since a Bayesian network is a directed acyclical graph, it can be represented as a series of ordered nodes connected by directed arcs. The nodes are ordered in a hierarchy where a node cannot be the parent of another node higher in the hierarchy. In other words, the nodes are topologically ordered. text/html 2011-02-18T07:10:32+00:00 http://www.syntheticsapien.com/doku.php/goplayer:cptwrappers?rev=1298013032&do=diff class CPTWrapper(object): pass; class SimpleDiscreteCPTBN(CPTWrapper): def __init__(self, nodeOrder, map, stateSpace, jft): ''' nodeOrder: [node1, node2, node3, ...] map: {node1:[], node2:[node1], node3:[node1, node2], ...} stateSpace: {node1:['a', 'b', 'c'], node2:[1, 3, 5, 7], node3:['high', 'low'], ...} jft: {node1:[0, 2, 3, 1, 1, ...], node2:[3, 3, 1, 1, 2, ...], node3:[1, 1, 0, 0, 0, ...], ...} ''' self.__nodeMap = map; self.__orderedNodeList = nodeOrder; self.__state… text/html 2011-09-14T00:31:43+00:00 http://www.syntheticsapien.com/doku.php/goplayer:dbconnector?rev=1315960303&do=diff /* ============================================================================ Name : DBConnector.h Author : Stephen Cannon Version : 0.1 Copyright : Copyright 2011 Stephen Cannon Description : ============================================================================ * * This file is part of LikelihoodWeighting. * * LikelihoodWeighting is free software: you can redistribute it and/or modify * it under the terms of the GNU Lesser General Public License as publish… text/html 2011-09-08T13:00:53+00:00 http://www.syntheticsapien.com/doku.php/goplayer:dbrv?rev=1315486853&do=diff /* ============================================================================ Name : dbrv.h Author : Stephen Cannon Version : 0.1 Copyright : Copyright 2011 Stephen Cannon Description : ============================================================================ * * This file is part of LikelihoodWeighting. * * LikelihoodWeighting is free software: you can redistribute it and/or modify * it under the terms of the GNU Lesser General Public License as published by … text/html 2011-02-13T03:54:08+00:00 http://www.syntheticsapien.com/doku.php/goplayer:goplayer?rev=1297569248&do=diff So now that I’ve got my Master’s degree, I am looking to expand and test my knowledge in AI as well as see what I’m capable of. I think one of the best ways to achieve this goal is to enter one or more AI competitions. This way I can best test and hone my abilities against others more capable than me as well as get direct feedback regarding where my abilities lie in comparison to others. Basically, I feel any competition that forces me to improve my skills in AI are good for me at this stage … text/html 2011-09-08T13:04:48+00:00 http://www.syntheticsapien.com/doku.php/goplayer:likelihoodweighting?rev=1315487088&do=diff cdef extern from "dbrv.h": char dbrv__setup(DBRV *self, char *connectionString); char dbrv__logP(DBRV *self, char *nodeName, PyObject *parentValues, double *prob); char dbrv__randomSample(DBRV *self, char *nodeName, PyObject *parentValues, unsigned int *sample); char dbrv__tearDown(DBRV *self); cdef class NodeSample: cdef node cdef int stateIndex cdef list states cdef list isSet cdef list parents cdef class LikelihoodWeighting: cdef list __orderedNodes cdef size_t __numSample … text/html 2011-12-16T20:29:41+00:00 http://www.syntheticsapien.com/doku.php/goplayer:mcffnngmmdist?rev=1324067381&do=diff The Monte Carlo Multilayer Perceptron Gaussian Mixture Model Distribution (MCMPGMMDist). A Bayesian network node conditional probability distribution is represented by a series of Gaussian mixture models. A Gaussian mixture model is defined for every combination of states the parent nodes of the given node can take on. The parameters for each Gaussian mixture model is determined from a multilayer perceptron. As the mixture models for each set of parent states are not independent of each othe… text/html 2012-04-17T04:09:00+00:00 http://www.syntheticsapien.com/doku.php/goplayer:mcmpgmmdist?rev=1334635740&do=diff The Monte Carlo Multilayer Perceptron Gaussian Mixture Model Distribution (MCMPGMMDist). A Bayesian network node conditional probability distribution is represented by a series of Gaussian mixture models. A Gaussian mixture model is defined for every combination of states the parent nodes of the given node can take on. The parameters for each Gaussian mixture model is determined from a multilayer perceptron. As the mixture models for each set of parent states are not independent of each othe… text/html 2011-02-13T09:06:59+00:00 http://www.syntheticsapien.com/doku.php/goplayer:rv?rev=1297588019&do=diff I wanted to use Python to define a random variable in a pythonic way. That way, I could intuitively work with random variables from within python and potentially even use some of Python's more powerful language features in conjunction with these random variables. text/html 2011-12-19T21:17:05+00:00 http://www.syntheticsapien.com/doku.php/goplayer:sidebar?rev=1324329425&do=diff * Main * Go Player * Python Representation of RVs * Python CPT Wrappers * Monte Carlo Multilayer Perceptron Gaussian Mixture Model Distribution * BN State Space * DB Connector * DB RV * Likelihood Weighting * Using DokuWiki * About Me