Documentation

Determinize.Spec.RewardModel.Model

Reward models #

Finite models whose transitions are lists of edges, each paying a reward that is added to the eventual output.

An outgoing edge of a state.

  • target : Fin size

    The state the edge leads to.

  • probability : ℚ

    The probability of taking the edge.

  • reward : ℚ

    The reward added to the output when the edge is taken.

Instances For

    Edge rewards translate successful outputs. Rejection and divergence discard them.

    Instances For

      The probability of moving from state i to state j in one step: the total probability of the edges from i to j.

      Equations
      Instances For

        The law of x + reward when x has law law.

        Equations
        Instances For

          Unnormalized output accumulated within steps transitions from a state. Taking an edge adds its reward to the eventual output.

          Equations
          Instances For

            The output law from a state: the limit of outputWithin over the number of steps.

            Equations
            Instances For

              The output law of the model: the output law from its initial state.

              Equations
              Instances For

                Exact source correspondence includes failure freedom and the entire output law.

                Equations
                Instances For