Package jdistlib.inference

Interface DifferentiableLogDensity

All Superinterfaces:
LogDensity
All Known Subinterfaces:
BatchedDifferentiableLogDensity, ComputeBackedLogDensity
All Known Implementing Classes:
AcceleratedLogisticRegression, BayesianModel, EvaluationCounter, ModelEvaluator, ReverseModeLogDensity
public interface DifferentiableLogDensity extends LogDensity
A log density capable of adding its gradient to caller-owned storage.

  • Method Summary

    Modifier and Type
    Method
    Description
    default double
    logDensity(double[] state)
     
    double
    logDensityAndGradient(double[] state, double[] gradient)
    Evaluates the log density and replaces gradient with its gradient.

  • Method Details

    • logDensityAndGradient

      double logDensityAndGradient(double[] state, double[] gradient)
      Evaluates the log density and replaces gradient with its gradient.

    • logDensity

      default double logDensity(double[] state)
      Specified by:
      logDensity in interface LogDensity