radiation_objs: List[Rad_damage] = []
for effect_type, rad_settings in radiations.items():
if effect_type == "neuron_death":
for probability_per_t in rad_settings["probability_per_t"]:
rad_damage = Rad_damage(
amplitude=-10**10,
# amplitude=-inf,
effect_type=effect_type,
excitatory=False,
inhibitory=True,
probability_per_t=probability_per_t,
)
radiation_objs.append(rad_damage)
else:
add_generic_exp_config_rad_dict(
effect_type=effect_type,
rad_settings=rad_settings,
radiation_objs=radiation_objs,
)
return radiation_objs
yields NAN error in the counter neurons (results), and in:
# if self.out_pre[self.index] > 0 and not isinstance(self.pre, LIF):
# input(f'{self.out_pre[self.index]}, {self.post.name}, {self.post.I}')
in connections.py.
Bandaid
Currently, a synapse weight of -10**8 is selected, as this is considered enough for the MDSA algorithm neurons to stop spiking throughout any practical remainder of the simulation.
Objective
Either :
[ ] properly compute what the minimum synapse weight should be to ensure every neuron (type) (in the MDSA) algorithm stops spiking for the remainder of the simulation:
or:
[ ] Modify the neurons subjected to neuron_death such that they do not spike anymore.
-inf
at:yields NAN error in the counter neurons (results), and in:
in connections.py.
Bandaid
Currently, a synapse weight of
-10**8
is selected, as this is considered enough for the MDSA algorithm neurons to stop spiking throughout any practical remainder of the simulation.Objective
Either :
neuron_death
such that they do not spike anymore.