pyhgf.response.binary_softmax_inverse_temperature

pyhgf.response.binary_softmax_inverse_temperature(hgf: HGF, response_function_inputs=typing.Union[jax.Array, numpy.ndarray, numpy.bool_, numpy.number, bool, int, float, complex], response_function_parameters=typing.Union[jax.Array, numpy.ndarray, numpy.bool_, numpy.number, bool, int, float, complex]) → float

Surprise from a binary sofmax parametrized by the inverse temperature.

The probability of chosing A is given by:

\[P(A|\hat{\mu}^{(k)_{1}, t) = \frac{1}{1+e^{-t\hat{\mu}^{(k)_{1}}}\]

Where \(\hat{mu}^{(k)_{1}\) is the expected probability of A at the firt level, and \(t\) is the temperature parameter.

Parameters:

hgf

An instance of the HGF model.

response_function_inputs

The inputs to the response functions, here containing the decision from the paraticipant at time k [0 or 1].

response_function_parameters

Additionnal parameters for the response function (optional). Here, the inverse temperature is provided.

Returns:

surprise

The surprise under the binary sofmax model.