Package samer.models
Class ICAWithScaler
- java.lang.Object
-
- Viewable
-
- samer.models.ICA
-
- samer.models.ICAWithScaler
-
-
Nested Class Summary
Nested Classes Modifier and Type Class and Description classICAWithScaler.DifferentialTrainerclassICAWithScaler.ScalerTrainerThis one trains ONLY the scaler part, not the ICA part, so is a lot faster than using the differential trainer with a zero learning rate.-
Nested classes/interfaces inherited from class samer.models.ICA
ICA.NewtonTrainer, ICA.ON2DecayWhenActive, ICA.ON2Trainer, ICA.ON3Trainer
-
Nested classes/interfaces inherited from interface samer.models.Model
Model.Trainer
-
-
Constructor Summary
Constructors Constructor and Description ICAWithScaler(int N)ICAWithScaler(Vec input)
-
Method Summary
All Methods Instance Methods Concrete Methods Modifier and Type Method and Description voidcompute()contract is that getEnergy and getGradient must return correct values for current x after infer and compute has been called, but not necessarily before.voiddispose()voidfold()Functionxfunctionx()This presents a more functional interface to the model so that it can be driven by an optimiser.Model.TrainergetDiffTrainer()Model.TrainergetScaleTrainer()voidinfer()should infer values latent variables-
Methods inherited from class samer.models.ICA
execute, getAltTrainer, getBasisMatrix, getCommands, getDecayWhenActiveTrainer, getEnergy, getGradient, getOutputModel, getSize, getTrainer, getWeightMatrix, output, run, setInput, setOutputModel, starting, stopping
-
-
-
-
Constructor Detail
-
ICAWithScaler
public ICAWithScaler(int N)
-
ICAWithScaler
public ICAWithScaler(Vec input)
-
-
Method Detail
-
dispose
public void dispose()
-
infer
public void infer()
Description copied from interface:Modelshould infer values latent variables
-
compute
public void compute()
Description copied from interface:Modelcontract is that getEnergy and getGradient must return correct values for current x after infer and compute has been called, but not necessarily before. This is to give model an opportunity to cache values of energy and gradient to avoid repeated computations.
-
functionx
public Functionx functionx()
Description copied from interface:ModelThis presents a more functional interface to the model so that it can be driven by an optimiser. See classes Functionx and MinimiserBase in package samer.maths.opt.
-
fold
public void fold()
-
getDiffTrainer
public Model.Trainer getDiffTrainer()
-
getScaleTrainer
public Model.Trainer getScaleTrainer()
-
-