def __init__(
self, data={}, n_bits_in=16, n_bits_out=16, n_ones_in=3, n_ones_out=3, n_samples=-1, algorithm="balanced"
):
"""
Fills the network structure with the given parameters.
:param n_bits_in: number of input bits.
:param n_bits_out: number of output bits.
:param n_ones_in: number of ones in the input per sample.
:param n_ones_out: number of ones in the output per sample.
:param n_samples: number of samples. If negative, the number of samples
with the maximum information are chosen.
"""
# If "n_bits" is specified, set both n_bits_in and n_bits_out
if ("n_bits" in data) and (data["n_bits"] > 0):
self["n_bits_in"] = int(data["n_bits"])
self["n_bits_out"] = int(data["n_bits"])
self["n_bits"] = int(data["n_bits"])
else:
utils.init_key(self, data, "n_bits_in", n_bits_in, _type=int)
utils.init_key(self, data, "n_bits_out", n_bits_out, _type=int)
utils.init_key(
self, data, "n_bits", self["n_bits_in"] if self["n_bits_in"] == self["n_bits_out"] else -1, _type=int
)
utils.init_key(self, data, "n_ones_in", n_ones_in, _type=int)
utils.init_key(self, data, "n_ones_out", n_ones_out, _type=int)
utils.init_key(self, data, "n_samples", n_samples, _type=int)
utils.init_key(self, data, "algorithm", algorithm, _type=str)
if not self["algorithm"] in ["random", "balanced", "unique"]:
raise Exception(
'Invalid data generation algorithm "'
+ self["algorithm"]
+ '", must be one of {"random", "balanced", "unique"}!'
)
# If n_ones_in and n_ones_out is not given, automatically calculate
# n_ones_in, n_ones_out and n_samples
if (self["n_ones_in"] <= 0) and (self["n_ones_out"] <= 0):
params = entropy.optimal_parameters(
n_samples=self["n_samples"], n_bits_in=self["n_bits_in"], n_bits_out=self["n_bits_out"]
)
self["n_samples"] = params["n_samples"]
self["n_ones_in"] = params["n_ones_in"]
self["n_ones_out"] = params["n_ones_out"]
# Automatically choose the optimal number of samples
if self["n_samples"] <= 0:
self["n_samples"] = entropy.optimal_sample_count(
n_bits_out=self["n_bits_out"],
n_bits_in=self["n_bits_in"],
n_ones_out=self["n_ones_out"],
n_ones_in=self["n_ones_in"],
)
def __init__(self, data={}, n_bits_in=16, n_bits_out=16, n_ones_in=3,
n_ones_out=3, n_samples=-1, algorithm="balanced"):
"""
Fills the network structure with the given parameters.
:param n_bits_in: number of input bits.
:param n_bits_out: number of output bits.
:param n_ones_in: number of ones in the input per sample.
:param n_ones_out: number of ones in the output per sample.
:param n_samples: number of samples. If negative, the number of samples
with the maximum information are chosen.
"""
# If "n_bits" is specified, set both n_bits_in and n_bits_out
if ("n_bits" in data) and (data["n_bits"] > 0):
self["n_bits_in"] = int(data["n_bits"])
self["n_bits_out"] = int(data["n_bits"])
self["n_bits"] = int(data["n_bits"])
else:
utils.init_key(self, data, "n_bits_in", n_bits_in, _type=int)
utils.init_key(self, data, "n_bits_out", n_bits_out, _type=int)
utils.init_key(self, data, "n_bits",
self["n_bits_in"] if self["n_bits_in"] == self[
"n_bits_out"]
else -1, _type=int)
utils.init_key(self, data, "n_ones_in", n_ones_in, _type=int)
utils.init_key(self, data, "n_ones_out", n_ones_out, _type=int)
utils.init_key(self, data, "n_samples", n_samples, _type=int)
utils.init_key(self, data, "algorithm", algorithm, _type=str)
if not self["algorithm"] in ["random", "balanced", "unique"]:
raise Exception("Invalid data generation algorithm \""
+ self["algorithm"]
+ "\", must be one of {\"random\", \"balanced\", \"unique\"}!")
# If n_ones_in and n_ones_out is not given, automatically calculate
# n_ones_in, n_ones_out and n_samples
if (self["n_ones_in"] <= 0) and (self["n_ones_out"] <= 0):
params = entropy.optimal_parameters(n_samples=self["n_samples"],
n_bits_in=self["n_bits_in"],
n_bits_out=self["n_bits_out"])
self["n_samples"] = params["n_samples"]
self["n_ones_in"] = params["n_ones_in"]
self["n_ones_out"] = params["n_ones_out"]
# Automatically choose the optimal number of samples
if self["n_samples"] <= 0:
self["n_samples"] = entropy.optimal_sample_count(
n_bits_out=self["n_bits_out"],
n_bits_in=self["n_bits_in"],
n_ones_out=self["n_ones_out"],
n_ones_in=self["n_ones_in"])
import binam
import data
import utils
import entropy
def cm2inch(value):
return value / 2.54
# Parameters
n_bits = 96
n_ones = 8
n_samples = entropy.optimal_sample_count(n_bits_in=n_bits,
n_bits_out=n_bits,
n_ones_in=n_ones,
n_ones_out=n_ones)
# Generate the input and output data
print("Create input data...")
X = data.generate(n_bits=n_bits, n_ones=n_ones, n_samples=n_samples)
print("Create output data...")
Y = data.generate(n_bits=n_bits, n_ones=n_ones, n_samples=n_samples)
# Train the BiNAM
print("Training BiNAM...")
M = binam.BiNAM(n_bits, n_bits)
M.train_matrix(X, Y)
print("Running experiments...")
xs = np.linspace(0.0, 1.0, 100)
# Include the PyNAM folder
import __main__
sys.path.append(os.path.join(os.path.dirname(__main__.__file__), "../pynam"))
import binam
import data
import utils
import entropy
def cm2inch(value):
return value / 2.54
# Parameters
n_bits = 96
n_ones = 8
n_samples = entropy.optimal_sample_count(n_bits_in = n_bits,
n_bits_out = n_bits, n_ones_in=n_ones, n_ones_out = n_ones)
# Generate the input and output data
print("Create input data...")
X = data.generate(n_bits=n_bits, n_ones=n_ones, n_samples=n_samples)
print("Create output data...")
Y = data.generate(n_bits=n_bits, n_ones=n_ones, n_samples=n_samples)
# Train the BiNAM
print("Training BiNAM...")
M = binam.BiNAM(n_bits, n_bits)
M.train_matrix(X, Y)
print("Running experiments...")
xs = np.linspace(0.0, 1.0, 100)
nxs = len(xs)
