1
2
3
4
5
6
7
8
9
10
11
12
13
14
15
16
17
18
19
20
21
22
23
24
25
26
27
28
29
30
31
32
33
34
35
36
37
38
39
40
41
42
43
44
45
46
47
48
49
50
51
52
53
54
55
56
57
58
59
60
61
62
63
64
65
66
67
68
69
70
71
72
73
74
75
76
77
78
79
80
81
82
83
84
85
86
87
88
89
90
|
import numpy as np
class MatrixEncoder:
def ident(dim, value=1):
return np.identity(dim) * value
class Vector:
def __init__(self, arr):
self.arr = arr
self.len = len(arr)
def add(self, other):
return self.__class__(self.arr + other.arr)
def __str__(self):
return str(self.arr)
def print(self, f="{}"):
print(f.format(self))
return self
def gaussian(size, mu, sigma):
return Vector(sigma * np.random.randn(size) + mu)
class BitVector(Vector):
def gen_uniform(size):
arr = np.random.randint(0, 2, size)
return BitVector(arr)
def to_int(self):
reverse = np.flip(self.arr)
res = 0
for (i, b) in enumerate(reverse):
res += b * 2**i
return res
def encode(self, matrix):
return SymbolVector(matrix[self.to_int()])
def from_int(index, size=None):
bits = []
while index > 0:
remainder = index % 2
index = index // 2
bits.insert(0, remainder)
# Padding if they want
if size is not None:
missing = max(0, size - len(bits))
bits = [0] * missing + bits
return BitVector(bits)
def check_if_error(self, other):
return np.array_equal(self.arr, other.arr)
class SymbolVector(Vector):
def with_noise(self, c):
noise = Vector.gaussian(self.len, 0, c)
return self.add(noise)
def decide_and_decode(self, matrix):
dim = len(matrix)
# Collect all distances in vector
dists = np.arange(dim)
for (i, vec) in enumerate(matrix):
dists[i] = np.dot(self.arr, vec)
# Find index with smallest
index = np.argmax(dists)
# Convert integer back to bit vector
return BitVector.from_int(index, size=int(np.log2(dim)))
encoder = MatrixEncoder.ident(4, np.sqrt(1))
faults = 0
total = 10000
for i in range(total):
original = BitVector.gen_uniform(2)
after = original.encode(encoder) \
.with_noise(1) \
.decide_and_decode(encoder)
if original.check_if_error(after):
faults += 1
print(f"Fault percentage {(faults / total) * 100}%")
|
