BinaryReward

class coba.primitives.BinaryReward

A reward function with two values.

Constructors

__init__(argmax: str | Number | Sequence | Mapping, value: float = 1.0) → None

Instantiate BinaryReward.

Parameters:

• argmax – The location where reward value 1 is returned. At all other actions a reward value of 0 is returned.

• value – The value returned at the argmax.

Methods

__call__(action: str | Number | Sequence | Mapping) → float

Get reward for action.

Parameters:

action – An action taken.

Returns:

Reward received for the action.