drawtool( gs( :fillbg 0 :fgcolor "DkGray",0.503548,0.503548,0.503548 :font "-*-times-medium-r-normal-*-12-*-*-*-*-*-*-*","Times-Roman",12 :graypat 0), gs( :fillbg 0 :fgcolor "Black",0,0,0 :font "-*-helvetica-bold-r-normal-*-14-*-*-*-*-*-*-*","Helvetica-Bold",14 :graypat 0), gs( :fillbg 0 :fgcolor "Black",0,0,0 :font "-*-helvetica-medium-r-normal-*-14-*-*-*-*-*-*-*","Helvetica",14 :graypat 0), gs( :fillbg 0 :fgcolor "Black",0,0,0 :font "-*-times-medium-r-normal-*-14-*-*-*-*-*-*-*","Times-Roman",14 :graypat 0), gs( :fillbg 0 :fgcolor "Black",0,0,0 :font "-*-times-bold-r-normal-*-14-*-*-*-*-*-*-*","Times-Bold",14 :graypat 0), text(12,"methods to measure and maximize generalization" :gs 0 :transform 0.978079,-0.208233,0.208233,0.978079,308.911,588.629), text(12,"Hebb's rule and PCA" :gs 0 :transform 0.945865,-0.324561,0.324561,0.945865,41.6674,611.985), text(12,"Q-learning" :gs 0 :transform 0.948683,0.316228,-0.316228,0.948683,127.646,551.599), text(12,"generalized expectation/maximization" :gs 0 :transform 0.994916,-0.100712,0.100712,0.994916,292.383,331.497), text(12,"perceptrons and linear programming" :gs 0 :transform 0.975403,-0.22043,0.22043,0.975403,359.083,224.63), text(12,"multilayer perceptrons" :gs 0 :transform 0.9792,0.202897,-0.202897,0.9792,457.353,290.555), text(12,"infomax" :gs 0 :transform 0.96245,-0.27146,0.27146,0.96245,69.3181,44.1433), text(12,"recurrent neural networks" :gs 0 :transform 0.963915,0.26621,-0.26621,0.963915,146.763,307.185), text(12,"biological neurons" :gs 0 :transform 0.932571,-0.360988,0.360988,0.932571,484.978,678.975), text(12,"sensory processing" :gs 0 :transform 0.555666,0.831406,-0.831406,0.555666,147.225,582.899), text(12,"lateral competition" :gs 0 :transform 0.256308,0.966595,-0.966595,0.256308,535.215,395.579), text(12,"trajectory learning in continuous-time dynamical systems" :gs 0 :transform 0.0350318,0.999386,-0.999386,0.0350318,594.923,396.229), text(12,"the structure of the Hessian" :gs 0 :transform 0.798316,0.602239,-0.602239,0.798316,470.81,219.735), text(12,"Gaussian mixture model" :gs 0 :transform 0.940266,0.340441,-0.340441,0.940266,387.252,598.164), text(12,"the method of maximum entropy" :gs 0 :transform 0.939947,0.34132,-0.34132,0.939947,295.439,409.454), text(12,"Helmholtz machines" :gs 0 :transform 0.969015,-0.247004,0.247004,0.969015,229.599,461.72), text(12,"Boltzmann machines" :gs 0 :transform 0.839075,-0.544016,0.544016,0.839075,64.7051,93.6891), text(12,"simulated annealing" :gs 0 :transform 0.99153,-0.129881,0.129881,0.99153,88.9388,228.395), text(12,"ARP" :gs 0 :transform -4.37114e-08,-1,1,-4.37114e-08,578.5,46.5), text(12,"attractor networks" :gs 0 :transform 0.916399,0.400266,-0.400266,0.916399,245.286,212.053), text(12,"hidden units" :gs 0 :transform 0.973841,0.22723,-0.22723,0.973841,392.876,660.578), text(12,"hidden structure" :gs 0 :transform 0.973575,0.228369,-0.228369,0.973575,237.157,544.645), text(12,"feature detectors" :gs 0 :transform 0.987299,-0.158876,0.158876,0.987299,347.182,769.281), text(12,"Bayesian interpretation" :gs 0 :transform 0.964931,-0.262503,0.262503,0.964931,50.1015,281.216), text(12,"probabilistic framework" :gs 0 :transform 0.9961,0.0882261,-0.0882261,0.9961,436.877,753.779), text(12,"nonlinear compression" :gs 0 :transform 0.964764,-0.263117,0.263117,0.964764,439.026,590.594), text(12,"autoencoder" :gs 0 :transform 0.887994,-0.459854,0.459854,0.887994,499.256,614.578), text(12,"associative networks" :gs 0 :transform 0.993994,0.109431,-0.109431,0.993994,182.384,415.121), text(12,"Vapnic-Chervonenkis dimension" :gs 0 :transform 0.998821,-0.048539,0.048539,0.998821,159.291,83.9998), text(12,"probably approximately correct" :gs 0 :transform 0.98917,0.146777,-0.146777,0.98917,370.252,83.5815), text(12,"broken symmetry" :gs 0 :transform -0.127752,-0.991806,-0.991806,0.127752,551.526,99.1587), text(12,"stochastic gradient methods" :gs 0 :transform 0.963811,-0.266586,0.266586,0.963811,66.8034,201.651), text(12,"generalized online delta-bar-delta" :gs 0 :transform 0.992011,0.126152,-0.126152,0.992011,444.186,331.056), text(12,"Backwards propagation of error" :gs 0 :transform 0.998818,-0.0486091,0.0486091,0.998818,265.788,199.835), text(12,"mean field theory" :gs 0 :transform 0.910058,0.414482,-0.414482,0.910058,298.451,761.096), text(12,"sigma-pi units" :gs 0 :transform 0.921887,-0.38746,0.38746,0.921887,229.401,778.83), text(12,"course coding" :gs 0 :transform 0.139468,-0.990227,0.990227,0.139468,44.6492,197.721), text(12,"sparse representations" :gs 0 :transform 0.990621,0.136637,-0.136637,0.990621,99.9927,773.606), text(12,"relearning" :gs 0 :transform 0.64204,-0.766671,0.766671,0.64204,37.8367,795.852), text(12,"multitask learning" :gs 0 :transform 0.803947,-0.594701,0.594701,0.803947,520.603,713.68), text(12,"constructuve interference" :gs 0 :transform 0.988194,0.153208,-0.153208,0.988194,72.6487,689.571), text(12,"real-world datasets" :gs 0 :transform 0.907674,-0.419677,0.419677,0.907674,530.295,796.115), text(12,"standard benchmarks" :gs 0 :transform 0.358851,-0.933395,0.933395,0.358851,45.035,394.372), picture( text(14,"Neural Networks and Machine Learning" :gs 1 :transform 2,0,0,2,50,740), text(14,"ECE 547 / CS 591" :gs 2 :transform 2,0,0,2,208,698), text(14,"Fall 1998, Tue/Thur, 12:30-1:45" :gs 3 :transform 1.2,0,0,1.2,202.6,625.1), text(14,"http://www.cs.unm.edu/~bap/teach/F98/ECE547/" :gs 3 :transform 1.2,0,0,1.2,158.4,42.1), text(14,"Professor Barak Pearlmutter" :gs 4 :transform 1.2,0,0,1.2,206,644.1), text(14,"You may be kidnapped and find yourself working", "at a hidden data mine on the shady side of Silicon Valley." :gs 3 :transform 1.2,0,0,1.2,127.2,148.5), text(14,"Beware!" :gs 4 :transform 1.44,0,0,1.44,282.72,171.42), text(14,"style: hands-on, experimental", "goal: intuition and understanding", "difficulty: commensurate with ability" :gs 4 :transform 1.2,0,0,1.2,182.1,306.9), text(14,"Those intrepid undergraduates and grad students who take this", "cross-listed class will, by the end of the term, be qualified to apply", "neural network techniques to a plethora of real-world scientific and", "industrial problems." :gs 3 :transform 1.2,0,0,1.2,91.6,398.3), text(14,"Somewhere between the soft underbelly of theoretical", "neurobiology, the shoals of statistics, the ancient ruins of", "cybernetics, and the firm bedrocks of signal processing and", "control theory lies the newfound territory of neural networks" :gs 3 :transform 1.2,0,0,1.2,134.1,531.3) ), text(12,"minimum mutual information" :gs 0 :transform 0.0412316,0.99915,-0.99915,0.0412316,16.5966,596.725), text(12,"the method of temporal differences" :gs 0 :transform 0.0567267,-0.99839,0.99839,0.0567267,575.53,256.586), text(12,"supervised learning" :gs 0 :transform 0.633238,0.773957,-0.773957,0.633238,26.6921,466.996), text(12,"unsupervised learning" :gs 0 :transform 0.998998,-0.0447558,0.0447558,0.998998,106.234,438.435), text(12,"reinforcement learning" :gs 0 :transform 0.957098,0.289764,-0.289764,0.957098,27.2434,431.602), text(12,"world champion" :gs 0 :transform 0.713336,0.700822,-0.700822,0.713336,194.377,546.521) )