def build_moja(model, moja, rule):
conn = rule.connection
pre_activities = model.sig[get_pre_ens(conn).neurons]["out"]
post_activities = model.sig[get_post_ens(conn).neurons]["out"]
pre_filtered = build_or_passthrough(model, moja.pre_synapse,
pre_activities)
post_filtered = build_or_passthrough(model, moja.post_synapse,
post_activities)
pos_memristors, \
neg_memristors, \
r_min_noisy, \
r_max_noisy, \
exponent_noisy = initialise_memristors( moja, pre_filtered.shape[ 0 ], post_filtered.shape[ 0 ] )
model.sig[conn]["pos_memristors"] = pos_memristors
model.sig[conn]["neg_memristors"] = neg_memristors
model.add_op(
SimmOja(pre_filtered, post_filtered, moja.beta,
model.sig[conn]["pos_memristors"],
model.sig[conn]["neg_memristors"], model.sig[conn]["weights"],
moja.noise_percentage, moja.gain, r_min_noisy, r_max_noisy,
exponent_noisy, moja.voltage, moja.initial_state))
# expose these for probes
model.sig[rule]["pre_filtered"] = pre_filtered
model.sig[rule]["post_filtered"] = post_filtered
model.sig[rule]["pos_memristors"] = pos_memristors
model.sig[rule]["neg_memristors"] = neg_memristors
def build_stdp(model, stdp, rule):
conn = rule.connection
pre_activities = model.sig[get_pre_ens(conn).neurons]['out']
post_activities = model.sig[get_post_ens(conn).neurons]['out']
pre_trace = Signal(np.zeros(pre_activities.size), name="pre_trace")
post_trace = Signal(np.zeros(post_activities.size), name="post_trace")
model.add_op(SimSTDP(
pre_activities,
post_activities,
pre_trace,
post_trace,
model.sig[conn]['weights'],
model.sig[rule]['delta'],
pre_tau=stdp.pre_tau,
post_tau=stdp.post_tau,
alf_p=stdp.alf_p,
alf_n=stdp.alf_n,
beta_p=stdp.beta_p,
beta_n=stdp.beta_n,
max_weight=stdp.max_weight,
min_weight=stdp.min_weight,
learning_rate=stdp.learning_rate,
))
# expose these for probes
model.sig[rule]['pre_trace'] = pre_trace
model.sig[rule]['post_trace'] = post_trace
model.params[rule] = None # no build-time info to return
def build_mpes(model, mpes, rule):
conn = rule.connection
# Create input error signal
error = Signal(shape=(rule.size_in, ), name="PES:error")
model.add_op(Reset(error))
model.sig[rule]["in"] = error # error connection will attach here
acts = build_or_passthrough(model, mpes.pre_synapse,
model.sig[conn.pre_obj]["out"])
post = get_post_ens(conn)
encoders = model.sig[post]["encoders"]
pos_memristors, neg_memristors, r_min_noisy, r_max_noisy, exponent_noisy = initialise_memristors(
mpes, acts.shape[0], encoders.shape[0])
model.sig[conn]["pos_memristors"] = pos_memristors
model.sig[conn]["neg_memristors"] = neg_memristors
if conn.post_obj is not conn.post:
# in order to avoid slicing encoders along an axis > 0, we pad
# `error` out to the full base dimensionality and then do the
# dotinc with the full encoder matrix
# comes into effect when slicing post connection
padded_error = Signal(shape=(encoders.shape[1], ))
model.add_op(Copy(error, padded_error, dst_slice=conn.post_slice))
else:
padded_error = error
# error = dot(encoders, error)
local_error = Signal(shape=(post.n_neurons, ))
model.add_op(Reset(local_error))
model.add_op(DotInc(encoders, padded_error, local_error, tag="PES:encode"))
model.operators.append(
SimmPES(acts, local_error, model.sig[conn]["pos_memristors"],
model.sig[conn]["neg_memristors"], model.sig[conn]["weights"],
mpes.noise_percentage, mpes.gain, r_min_noisy, r_max_noisy,
exponent_noisy, mpes.initial_state))
# expose these for probes
model.sig[rule]["error"] = error
model.sig[rule]["activities"] = acts
model.sig[rule]["pos_memristors"] = pos_memristors
model.sig[rule]["neg_memristors"] = neg_memristors
def build_stdp(model, stdp, rule):
conn = rule.connection
pre_activities = model.sig[get_pre_ens(conn).neurons]["out"]
post_activities = model.sig[get_post_ens(conn).neurons]["out"]
pre_trace = Signal(np.zeros(pre_activities.size), name="pre_trace")
post_trace = Signal(np.zeros(post_activities.size), name="post_trace")
pre_scale = Signal(np.zeros(model.sig[conn]["weights"].shape),
name="pre_scale")
post_scale = Signal(np.zeros(model.sig[conn]["weights"].shape),
name="post_scale")
model.add_op(
SimSTDP(
pre_activities,
post_activities,
pre_trace,
post_trace,
pre_scale,
post_scale,
model.sig[conn]["weights"],
model.sig[rule]["delta"],
learning_rate=stdp.learning_rate,
pre_tau=stdp.pre_tau,
post_tau=stdp.post_tau,
pre_amp=stdp.pre_amp,
post_amp=stdp.post_amp,
bounds=stdp.bounds,
max_weight=stdp.max_weight,
min_weight=stdp.min_weight,
))
# expose these for probes
model.sig[rule]["pre_trace"] = pre_trace
model.sig[rule]["post_trace"] = post_trace
model.sig[rule]["pre_scale"] = pre_scale
model.sig[rule]["post_scale"] = post_scale
model.params[rule] = None # no build-time info to return
def build_tripletstdp(model, stdp, rule):
conn = rule.connection
pre_activities = model.sig[get_pre_ens(conn).neurons]["out"]
post_activities = model.sig[get_post_ens(conn).neurons]["out"]
pre_trace1 = Signal(np.zeros(pre_activities.size), name="pre_trace1")
post_trace1 = Signal(np.zeros(post_activities.size), name="post_trace1")
pre_trace2 = Signal(np.zeros(pre_activities.size), name="pre_trace2")
post_trace2 = Signal(np.zeros(post_activities.size), name="post_trace2")
model.add_op(
SimTripletSTDP(
pre_activities,
post_activities,
pre_trace1,
post_trace1,
pre_trace2,
post_trace2,
model.sig[rule]["delta"],
learning_rate=stdp.learning_rate,
pre_tau=stdp.pre_tau,
pre_taux=stdp.pre_taux,
post_tau=stdp.post_tau,
post_tauy=stdp.post_tauy,
pre_amp2=stdp.pre_amp2,
pre_amp3=stdp.pre_amp3,
post_amp2=stdp.post_amp2,
post_amp3=stdp.post_amp3,
nearest_spike=stdp.nearest_spike,
))
# expose these for probes
model.sig[rule]["pre_trace1"] = pre_trace1
model.sig[rule]["post_trace1"] = post_trace1
model.sig[rule]["pre_trace2"] = pre_trace2
model.sig[rule]["post_trace2"] = post_trace2
model.params[rule] = None # no build-time info to return
def build_pes(model, pes, rule):
"""
Builds a `nengo.PES` object into a model.
Parameters
----------
model : Model
The model to build into.
pes : PES
Learning rule type to build.
rule : LearningRule
The learning rule object corresponding to the neuron type.
Notes
-----
Does not modify ``model.params[]`` and can therefore be called
more than once with the same `nengo.PES` instance.
"""
conn = rule.connection
# Create input error signal
error = Signal(np.zeros(rule.size_in), name="PES:error")
model.add_op(Reset(error))
model.sig[rule]['in'] = error # error connection will attach here
if LooseVersion(nengo_version) < "2.7.1":
acts = model.build(
Lowpass(pes.pre_tau), model.sig[conn.pre_obj]["out"])
else:
acts = model.build(pes.pre_synapse, model.sig[conn.pre_obj]["out"])
if not conn.is_decoded:
# multiply error by post encoders to get a per-neuron error
post = get_post_ens(conn)
encoders = model.sig[post]["encoders"]
if conn.post_obj is not conn.post:
# in order to avoid slicing encoders along an axis > 0, we pad
# `error` out to the full base dimensionality and then do the
# dotinc with the full encoder matrix
padded_error = Signal(np.zeros(encoders.shape[1]))
model.add_op(Copy(error, padded_error,
dst_slice=conn.post_slice))
else:
padded_error = error
# error = dot(encoders, error)
local_error = Signal(np.zeros(post.n_neurons), name="PES:encoded")
model.add_op(Reset(local_error))
model.add_op(DotInc(encoders, padded_error, local_error,
tag="PES:encode"))
else:
local_error = error
model.operators.append(SimPES(acts, local_error, model.sig[rule]["delta"],
pes.learning_rate))
# expose these for probes
model.sig[rule]["error"] = error
model.sig[rule]["activities"] = acts
def build_mpes(model, mpes, rule):
conn = rule.connection
# Create input error signal
error = Signal(shape=(rule.size_in, ), name="PES:error")
model.add_op(Reset(error))
model.sig[rule]["in"] = error # error connection will attach here
acts = build_or_passthrough(model, mpes.pre_synapse,
model.sig[conn.pre_obj]["out"])
post = get_post_ens(conn)
encoders = model.sig[post]["encoders"]
out_size = encoders.shape[0]
in_size = acts.shape[0]
from scipy.stats import truncnorm
def get_truncated_normal(mean, sd, low, upp):
try:
return truncnorm( (low - mean) / sd, (upp - mean) / sd, loc=mean, scale=sd ) \
.rvs( out_size * in_size ) \
.reshape( (out_size, in_size) )
except ZeroDivisionError:
return np.full((out_size, in_size), mean)
np.random.seed(mpes.seed)
r_min_noisy = get_truncated_normal(mpes.r_min,
mpes.r_min * mpes.noise_percentage[0],
0, np.inf)
np.random.seed(mpes.seed)
r_max_noisy = get_truncated_normal(mpes.r_max,
mpes.r_max * mpes.noise_percentage[1],
np.max(r_min_noisy), np.inf)
np.random.seed(mpes.seed)
exponent_noisy = np.random.normal(
mpes.exponent,
np.abs(mpes.exponent) * mpes.noise_percentage[2], (out_size, in_size))
np.random.seed(mpes.seed)
pos_mem_initial = np.random.normal(1e8, 1e8 * mpes.noise_percentage[3],
(out_size, in_size))
np.random.seed(mpes.seed + 1)
neg_mem_initial = np.random.normal(1e8, 1e8 * mpes.noise_percentage[3],
(out_size, in_size))
pos_memristors = Signal(shape=(out_size, in_size),
name="mPES:pos_memristors",
initial_value=pos_mem_initial)
neg_memristors = Signal(shape=(out_size, in_size),
name="mPES:neg_memristors",
initial_value=neg_mem_initial)
model.sig[conn]["pos_memristors"] = pos_memristors
model.sig[conn]["neg_memristors"] = neg_memristors
if conn.post_obj is not conn.post:
# in order to avoid slicing encoders along an axis > 0, we pad
# `error` out to the full base dimensionality and then do the
# dotinc with the full encoder matrix
# comes into effect when slicing post connection
padded_error = Signal(shape=(encoders.shape[1], ))
model.add_op(Copy(error, padded_error, dst_slice=conn.post_slice))
else:
padded_error = error
# error = dot(encoders, error)
local_error = Signal(shape=(post.n_neurons, ))
model.add_op(Reset(local_error))
model.add_op(DotInc(encoders, padded_error, local_error, tag="PES:encode"))
model.operators.append(
SimmPES(acts, local_error, mpes.learning_rate,
model.sig[conn]["pos_memristors"],
model.sig[conn]["neg_memristors"], model.sig[conn]["weights"],
mpes.noise_percentage, mpes.gain, r_min_noisy, r_max_noisy,
exponent_noisy))
# expose these for probes
model.sig[rule]["error"] = error
model.sig[rule]["activities"] = acts
model.sig[rule]["pos_memristors"] = pos_memristors
model.sig[rule]["neg_memristors"] = neg_memristors
