Package | Description |
---|---|
com.ibm.bluej.consistency | |
com.ibm.bluej.consistency.inference | |
com.ibm.bluej.consistency.learning |
Modifier and Type | Method and Description |
---|---|
IDeltaWeight |
CRFState.getWorldWeight() |
Modifier and Type | Method and Description |
---|---|
static SavedWorld |
SavedWorld.instance(IDeltaWeight worldWeight,
java.util.Iterator<IndexEntry> truth,
com.ibm.bluej.util.common.FunST<ScanTerm,java.lang.Boolean> shouldSave) |
IWorldWeight |
NBestWorldSaver.save(IDeltaWeight worldWeight,
java.util.Iterator<IndexEntry> truth) |
IWorldWeight |
SingleWorldSaver.save(IDeltaWeight worldWeight,
java.util.Iterator<IndexEntry> truth) |
IWorldWeight |
MAPWorldSaver.save(IDeltaWeight worldWeight,
java.util.Iterator<IndexEntry> truth) |
Modifier and Type | Class and Description |
---|---|
class |
DeltaOnly |
class |
FixedDeltaWeight |
class |
SingleDelta
Only makes sense if doing MAP inference and learning at the same time
Otherwise either DeltaOnly or FixedDeltaWeight is the right choice
|
Modifier and Type | Method and Description |
---|---|
void |
TrainingAnalysis.recordDelta(WorldChange delta,
double objectiveScore,
double prevObjectiveScore,
IDeltaWeight deltaStr) |