pyphi.substrate_generator.mechanisms.stabilized_sigmoid

pyphi.substrate_generator.mechanisms.stabilized_sigmoid(element, weights, state, *, input_weights, determinism, threshold, modulation, floor=0.0, ceiling=1.0, inputs=None, **kwargs)

Sigmoid whose modulators stabilize the unit’s current state, more strongly the more modulators are active.

modulation is {'modulator': tuple(indices), 'threshold': float, 'determinism': float, 'selectivity': float}. Requires a self-loop.

The modulator inputs are exactly modulation['modulator']; the remaining inputs drive the sigmoid. Results differ from the substrate_modeler library for this mechanism, which treated the input and modulator axes in swapped order.