framework

Type Members

1. trait AnnotatingInference[Datum] extends Inference[Datum]

   AnnotatingInference adds the ability to take a Marginal and attach (or replace) new information to the Datum.

2. trait AugmentableInference[Datum, Augment] extends Inference[Datum]

   AugmentableInference is an epic.framework.Inference that can support injecting additional information into the structure computation.

3. case class ComponentFeature[T](index: Int, feature: T) extends Feature with Product with Serializable

4. case class EPExpectedCounts(loss: Double, counts: IndexedSeq[ExpectedCounts[_]]) extends ExpectedCounts[EPExpectedCounts] with Product with Serializable

5. class EPInference[Datum, Augment <: AnyRef] extends ProjectableInference[Datum, Augment] with SafeLogging with Serializable

6. case class EPMarginal[Augment, Marginal](logPartition: Double, q: Augment, marginals: IndexedSeq[Marginal]) extends Marginal with Product with Serializable

7. class EPModel[Datum, Augment <: AnyRef] extends Model[Datum] with SafeLogging

8. case class EPScorer[Scorer](scorers: IndexedSeq[Scorer]) extends Product with Serializable

9. trait EvaluableModel[Datum] extends Model[Datum]

   A model that has some kind of evaluation function.

10. trait EvaluationResult[R <: EvaluationResult[R]] extends AnyRef

    Marker for the output of an evaluation routine.

11. trait ExpectedCounts[Self <: ExpectedCounts[Self]] extends AnyRef

12. trait Feature extends AnyRef

    Just a marker for features.

13. trait Inference[Datum] extends Serializable

    Inference is the core interface in Epic.

14. trait LossAugmentation[Datum, Augment] extends (Datum) ⇒ Augment

    A class that returns an augment that gives higher scores to spans that are wrong.

15. trait Marginal extends AnyRef

    Marginals are created by epic.framework.Inference objects and used for expected counts and decoding, where applicable.

16. trait Model[Datum] extends SafeLogging

    A Model represents a class for turning weight vectors into epic.framework.Inferences.

17. trait ModelFactory[Datum] extends AnyRef

    Interface for producing Models from training data.

18. class ModelObjective[Datum] extends BatchDiffFunction[DenseVector[Double]] with SafeLogging

    The objective function for training a epic.framework.Model.

19. class OneBestInferenceAdaptor[Datum] extends Inference[Datum]

    TODO

20. class OneBestModelAdaptor[Datum] extends Model[Datum]

21. trait ProjectableInference[Datum, Augment] extends AugmentableInference[Datum, Augment]

    A ProjectableInference is an epic.framework.AugmentableInference that can also create a new Augment from the marginal and the old augment.

22. case class StandardExpectedCounts[F](loss: Double, counts: DenseVector[Double], index: Index[F]) extends ExpectedCounts[StandardExpectedCounts[F]] with Product with Serializable

    This is a standard expected counts class that most models will use.

23. class StructuredPerceptron[Datum] extends LazyLogging

    TODO

24. trait VisitableMarginal[Visitor] extends Marginal