org.linqs.psl.application.learning.weight.search

Class InitialWeightHyperband

java.lang.Object

org.linqs.psl.application.learning.weight.WeightLearningApplication

org.linqs.psl.application.learning.weight.search.Hyperband

org.linqs.psl.application.learning.weight.search.InitialWeightHyperband

All Implemented Interfaces:

ModelApplication

public class InitialWeightHyperband
extends Hyperband

Hyperband, but the weights chosen are used as initial weights for further weight learning.

Field Summary

Fields

Modifier and Type	Field and Description
static String	CONFIG_PREFIX Prefix of property keys used by this class.
static String	INTERNAL_WLA_DEFAULT
static String	INTERNAL_WLA_KEY The internal weight learning application (WLA) to use.

Fields inherited from class org.linqs.psl.application.learning.weight.search.Hyperband

BASE_BRACKET_SIZE_DEFAULT, BASE_BRACKET_SIZE_KEY, MEAN, MIN_BRACKET_SIZE, MIN_BUDGET_PROPORTION, NUM_BRACKETS_DEFAULT, NUM_BRACKETS_KEY, SURVIVAL_DEFAULT, SURVIVAL_KEY, VARIANCE

Fields inherited from class org.linqs.psl.application.learning.weight.WeightLearningApplication

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 Summary

Constructors

Constructor and Description
InitialWeightHyperband(List<Rule> rules, Database rvDB, Database observedDB)
InitialWeightHyperband(Model model, Database rvDB, Database observedDB)

Method Summary

Modifier and Type	Method and Description
void	close() Releases all resources used by this ModelApplication.
protected void	postInitGroundModel() A convenient place for children to do additional ground model initialization.
protected double	run(double[] weights) Run and eval on the given weights using the given budget (ratio of max resources) and give back its score (lower is better).
void	setBudget(double budget) Set a budget (give as a proportion of the max budget).

Methods inherited from class org.linqs.psl.application.learning.weight.search.Hyperband

doLearn

Methods inherited from class org.linqs.psl.application.learning.weight.WeightLearningApplication

computeExpectedIncompatibility, computeLatentMPEState, computeLoss, computeMPEState, computeObservedIncompatibility, createAtomManager, getWLA, initGroundModel, initGroundModel, initGroundModel, initLatentGroundModel, learn, setDefaultRandomVariables, setLabeledRandomVariables

Methods inherited from class java.lang.Object

clone, equals, finalize, getClass, hashCode, notify, notifyAll, toString, wait, wait, wait

Field Detail

CONFIG_PREFIX

public static final String CONFIG_PREFIX

Prefix of property keys used by this class.

See Also:

Constant Field Values

INTERNAL_WLA_KEY

public static final String INTERNAL_WLA_KEY

The internal weight learning application (WLA) to use. Should actually be a VotedPerceptron.

See Also:

Constant Field Values

INTERNAL_WLA_DEFAULT

public static final String INTERNAL_WLA_DEFAULT

Constructor Detail

InitialWeightHyperband

public InitialWeightHyperband(Model model,
                              Database rvDB,
                              Database observedDB)

InitialWeightHyperband

public InitialWeightHyperband(List<Rule> rules,
                              Database rvDB,
                              Database observedDB)

Method Detail

postInitGroundModel

protected void postInitGroundModel()

Description copied from class: WeightLearningApplication

A convenient place for children to do additional ground model initialization.

Overrides:

postInitGroundModel in class WeightLearningApplication

setBudget

public void setBudget(double budget)

Description copied from class: WeightLearningApplication

Set a budget (give as a proportion of the max budget). Child implementations should make sure to override this and call up the super chain.

Overrides:

setBudget in class WeightLearningApplication

run

protected double run(double[] weights)

Description copied from class: Hyperband

Run and eval on the given weights using the given budget (ratio of max resources) and give back its score (lower is better). This method may modify weights if it wants to store a different set of weights than those initially passed in. The rules have already been set with the given weights, they are only passed in so the method has a chance to modify them before the result is stored. This is a prime method for child classes to override. Implementers should make sure to correct (negate) the value that comes back from the Evaluator if lower is better for that evaluator.

Overrides:

run in class Hyperband

close

public void close()

Description copied from interface: ModelApplication

Releases all resources used by this ModelApplication.

Specified by:

close in interface ModelApplication

Overrides:

close in class WeightLearningApplication