public abstract class ExpectationMaximization extends VotedPerceptron
Modifier and Type | Field and Description |
---|---|
static String |
CONFIG_PREFIX
Prefix of property keys used by this class.
|
protected int |
emIteration |
static int |
ITER_DEFAULT |
static String |
ITER_KEY
Key for positive int property for the number of iterations of expectation
maximization to perform
|
protected int |
iterations |
protected double |
tolerance |
static double |
TOLERANCE_DEFAULT |
static String |
TOLERANCE_KEY
Key for positive double property for the minimum absolute change in weights
such that EM is considered converged
|
AVERAGE_STEPS_DEFAULT, AVERAGE_STEPS_KEY, averageSteps, baseStepSize, CLIP_NEGATIVE_WEIGHTS_DEFAULT, CLIP_NEGATIVE_WEIGHTS_KEY, clipNegativeWeights, CUT_OBJECTIVE_DEFAULT, CUT_OBJECTIVE_KEY, cutObjective, inertia, INERTIA_DEFAULT, INERTIA_KEY, L1_REGULARIZATION_DEFAULT, L1_REGULARIZATION_KEY, l1Regularization, L2_REGULARIZATION_DEFAULT, L2_REGULARIZATION_KEY, l2Regularization, maxNumSteps, NUM_STEPS_DEFAULT, NUM_STEPS_KEY, numSteps, SCALE_GRADIENT_DEFAULT, SCALE_GRADIENT_KEY, SCALE_STEP_SIZE_DEFAULT, SCALE_STEP_SIZE_KEY, scaleGradient, scaleStepSize, STEP_SIZE_DEFAULT, STEP_SIZE_KEY, ZERO_INITIAL_WEIGHTS_DEFAULT, ZERO_INITIAL_WEIGHTS_KEY, zeroInitialWeights
allRules, atomManager, evaluator, EVALUATOR_DEFAULT, EVALUATOR_KEY, expectedIncompatibility, GROUND_RULE_STORE_DEFAULT, GROUND_RULE_STORE_KEY, groundRuleStore, inLatentMPEState, inMPEState, latentGroundRuleStore, latentTermStore, MAX_RANDOM_WEIGHT, MIN_ADMM_STEPS, mutableRules, observedDB, observedIncompatibility, RANDOM_WEIGHTS_DEFAULT, RANDOM_WEIGHTS_KEY, reasoner, REASONER_DEFAULT, REASONER_KEY, rvDB, supportsLatentVariables, TERM_GENERATOR_DEFAULT, TERM_GENERATOR_KEY, TERM_STORE_DEFAULT, TERM_STORE_KEY, termGenerator, termStore, trainingMap
Constructor and Description |
---|
ExpectationMaximization(List<Rule> rules,
Database rvDB,
Database observedDB) |
Modifier and Type | Method and Description |
---|---|
protected void |
doLearn()
Do the actual learning procedure.
|
protected void |
eStep()
The Expectation step in the EM algorithm.
|
protected void |
mStep()
The Maximization step in the EM algorithm.
|
computeRegularizer, computeScalingFactor, getLoss, setBudget
close, computeExpectedIncompatibility, computeLatentMPEState, computeLoss, computeMPEState, computeObservedIncompatibility, createAtomManager, getWLA, initGroundModel, initGroundModel, initGroundModel, initLatentGroundModel, learn, postInitGroundModel, setDefaultRandomVariables, setLabeledRandomVariables
public static final String CONFIG_PREFIX
public static final String ITER_KEY
public static final int ITER_DEFAULT
public static final String TOLERANCE_KEY
public static final double TOLERANCE_DEFAULT
protected final int iterations
protected final double tolerance
protected int emIteration
protected void doLearn()
WeightLearningApplication
doLearn
in class VotedPerceptron
protected void eStep()
protected void mStep()
Copyright © 2018 University of California, Santa Cruz. All rights reserved.