Related
I'm trying to write a function to evaluate the probability mass function for the bivariate poisson distribution.
This is easy when all of the parameters (x, y, theta1, theta2, theta0) are scalars, but tricky to scale up without loops to allow these parameters to be vectors. I need it to scale such that, for:
theta0 being a scalar - the "correlation parameter" in the equation
theta1 and theta2 having length l
x, y both having length n
the output array would have shape (l, n, n). For example, a slice [j, :, :] from the output array would look like:
The first part (the constant, before the summation) I think i've figured out:
import numpy as np
from scipy.special import factorial
def constant(theta1, theta2, theta0, x, y):
exponential_part = np.exp(-(theta1 + theta2 + theta0)).reshape(-1, 1, 1)
x = np.tile(x, (len(x), 1)).transpose()
y = np.tile(y, (len(y), 1))
double_factorial = (np.power(np.array(theta1).reshape(-1, 1, 1), x)/factorial(x)) * \
(np.power(np.array(theta2).reshape(-1, 1, 1), y)/factorial(y))
return exponential_part * double_factorial
But I'm struggling with the summation part. How can I vectorize a summation where the limits depend on variable arrays?
I think I have this figured out, based on the approach that #w-m suggests: calculate every possible summation term which could appear, based on the maximum x or y value which appears, and use a mask to get rid of the ones you don't want. Assuming you have your x and y terms go from 0 to N, in consecutive order, this is calculating up to three times more terms than are actually required, but this is offset by getting to use vectorization.
Reference implementation
I wrote this by first writing a pure-Python reference implementation, which just implements your problem using loops. With 4 nested loops, it's not exactly fast, but it's handy to have while testing the numpy version.
import numpy as np
from scipy.special import factorial, comb
import operator as op
from functools import reduce
def choose(n, r):
# https://stackoverflow.com/a/4941932/530160
r = min(r, n-r)
numer = reduce(op.mul, range(n, n-r, -1), 1)
denom = reduce(op.mul, range(1, r+1), 1)
return numer // denom # or / in Python 2
def reference_impl_constant(s_theta1, s_theta2, s_theta0, s_x, s_y):
# Cast to float to prevent overflow
s_theta1 = float(s_theta1)
s_theta2 = float(s_theta2)
s_theta0 = float(s_theta0)
s_x = float(s_x)
s_y = float(s_y)
term1 = np.exp(-(s_theta1 + s_theta2 + s_theta0))
term2 = (s_theta1 ** s_x / factorial(s_x))
term3 = (s_theta2 ** s_y / factorial(s_y))
assert term1 >= 0
assert term2 >= 0
assert term3 >= 0
return term1 * term2 * term3
def reference_impl_constant_loop(theta1, theta2, theta0, x, y):
theta_len = theta1.shape[0]
xy_len = x.shape[0]
constant_array = np.zeros((theta_len, xy_len, xy_len))
for i in range(theta_len):
for j in range(xy_len):
for k in range(xy_len):
s_theta1 = theta1[i]
s_theta2 = theta2[i]
s_theta0 = theta0
s_x = x[j]
s_y = y[k]
constant_term = reference_impl_constant(s_theta1, s_theta2, s_theta0, s_x, s_y)
assert constant_term >= 0
constant_array[i, j, k] = constant_term
return constant_array
def reference_impl_summation(s_theta1, s_theta2, s_theta0, s_x, s_y):
sum_ = 0
for i in range(min(s_x, s_y) + 1):
sum_ += choose(s_x, i) * choose(s_y, i) * factorial(i) * ((s_theta0/s_theta1/s_theta2) ** i)
assert sum_ >= 0
return sum_
def reference_impl_summation_loop(theta1, theta2, theta0, x, y):
theta_len = theta1.shape[0]
xy_len = x.shape[0]
summation_array = np.zeros((theta_len, xy_len, xy_len))
for i in range(theta_len):
for j in range(xy_len):
for k in range(xy_len):
s_theta1 = theta1[i]
s_theta2 = theta2[i]
s_theta0 = theta0
s_x = x[j]
s_y = y[k]
summation_term = reference_impl_summation(s_theta1, s_theta2, s_theta0, s_x, s_y)
assert summation_term >= 0
summation_array[i, j, k] = summation_term
return summation_array
def reference_impl(theta1, theta2, theta0, x, y):
# all array inputs must be 1D
assert len(theta1.shape) == 1
assert len(theta2.shape) == 1
assert len(x.shape) == 1
assert len(y.shape) == 1
# theta vectors must have same length
theta_len = theta1.shape[0]
assert theta2.shape[0] == theta_len
# x and y must have same length
xy_len = x.shape[0]
assert y.shape[0] == xy_len
# theta0 is scalar
assert isinstance(theta0, (int, float))
constant_array = np.zeros((theta_len, xy_len, xy_len))
output = np.zeros((theta_len, xy_len, xy_len))
constant_array = reference_impl_constant_loop(theta1, theta2, theta0, x, y)
summation_array = reference_impl_summation_loop(theta1, theta2, theta0, x, y)
output = constant_array * summation_array
return output
Numpy implementation
I split the implementation of this across two functions.
The fast_constant() function calculates everything to the left of the summation symbol. The fast_summation() function calculates everything inside the summation symbol.
import numpy as np
from scipy.special import factorial, comb
def fast_summation(theta1, theta2, theta0, x, y):
x = np.tile(x, (len(x), 1)).transpose()
y = np.tile(y, (len(y), 1))
sum_limit = np.minimum(x, y)
max_sum_limit = np.max(sum_limit)
i = np.arange(max_sum_limit + 1).reshape(-1, 1, 1)
summation_mask = (i <= sum_limit)
theta_ratio = (theta0 / (theta1 * theta2)).reshape(-1, 1, 1, 1)
theta_to_power = np.power(theta_ratio, i)
terms = comb(x, i) * comb(y, i) * factorial(i) * theta_to_power
# mask out terms which aren't part of sum
terms *= summation_mask
# axis 0 is theta
# axis 1 is i
# axis 2 & 3 are x and y
# so sum across axis 1
terms = terms.sum(axis=1)
return terms
def fast_constant(theta1, theta2, theta0, x, y):
theta1 = theta1.astype('float64')
theta2 = theta2.astype('float64')
exponential_part = np.exp(-(theta1 + theta2 + theta0)).reshape(-1, 1, 1)
# x and y must be 1D
assert len(x.shape) == 1
assert len(y.shape) == 1
# x and y must have same shape
assert x.shape == y.shape
x_len, y_len = x.shape[0], y.shape[0]
x = x.reshape((x_len, 1))
y = y.reshape((1, y_len))
double_factorial = (np.power(np.array(theta1).reshape(-1, 1, 1), x)/factorial(x)) * \
(np.power(np.array(theta2).reshape(-1, 1, 1), y)/factorial(y))
return exponential_part * double_factorial
def fast_impl(theta1, theta2, theta0, x, y):
return fast_summation(theta1, theta2, theta0, x, y) * fast_constant(theta1, theta2, theta0, x, y)
Benchmarking
Assuming that X and Y range from 0 to 20, and that theta is centered somewhere inside that range, I get the result that the numpy version is roughly 280 times faster than the pure python reference.
Numerical stability
I'm unsure how numerically stable this is. For example, when I center theta at 100, I get a floating-point overflow. Typically, when computing an expression which has lots of choose and factorial expressions inside it, you'll use some mathematical equivalent which results in smaller intermediate sums. In this case I have so little understanding of the math that I don't know how you'd do that.
This is more of a computational physics problem, and I've asked it on physics stack exchange, but no answers on there. This is, I suppose, a mix of the disciplines on here and there (and maybe even mathematics stack exchange), so finding the right place to post is a task in of itself apparently...
I'm attempting to use Crank-Nicolson scheme to solve the TDSE in 1D. The initial wave is a real Gaussian that has been normalised wrt its probability density. As the solution evolves, a depression grows in the central peak of the real part of the wave, and the imaginary part's central trough is perhaps a bit higher than I expect (image below).
Does this behaviour seem reasonable? I have searched around and not seen questions/figures that are similar. I've tested another person's code from Github and it exhibits the same behaviour, which makes me feel a bit better. But I still think the center peak should just decrease in height and increase in width. The likelihood of me getting a physics-based explanation is relatively low here I'd assume, but a computational-based explanation on errors I may have made is more likely.
I'm happy to give more information, for example my code, or the matrices used in the scheme, etc. Thanks in advance!
Here's a link to GIF of time evolution:
And the part of my code relevant to solving the 1D TDSE:
(pretty much the entire thing except the plotting)
import numpy as np
import matplotlib.pyplot as plt
from matplotlib.animation import FuncAnimation
# Define function for norm.
def normf(dxc, uc, ic):
return sum(dxc * np.square(np.abs(uc[ic, :])))
# Define function for expectation value of position.
def xexpf(dxc, xc, uc, ic):
return sum(dxc * xc * np.square(np.abs(uc[ic, :])))
# Define function for expectation value of squared position.
def xexpsf(dxc, xc, uc, ic):
return sum(dxc * np.square(xc) * np.square(np.abs(uc[ic, :])))
# Define function for standard deviation.
def sdaf(xexpc, xexpsc, ic):
return np.sqrt(xexpsc[ic] - np.square(xexpc[ic]))
# Time t: t0 =< t =< tf. Have N steps at which to evaluate the CN scheme. The
# time interval is dt. decp: variable for plotting to certain number of decimal
# places.
t0 = 0
tf = 20
N = 200
dt = tf / N
t = np.linspace(t0, tf, num = N + 1, endpoint = True)
decp = str(dt)[::-1].find('.')
# Initialise array for filling with norm values at each time step.
norm = np.zeros(len(t))
# Initialise array for expectation value of position.
xexp = np.zeros(len(t))
# Initialise array for expectation value of squared position.
xexps = np.zeros(len(t))
# Initialise array for alternate standard deviation.
sda = np.zeros(len(t))
# Position x: -a =< x =< a. M is an even number. There are M + 1 total discrete
# positions, for the points to be symmetric and centred at x = 0.
a = 100
M = 1200
dx = (2 * a) / M
x = np.linspace(-a, a, num = M + 1, endpoint = True)
# The gaussian function u diffuses over time. sd sets the width of gaussian. u0
# is the initial gaussian at t0.
sd = 1
var = np.power(sd, 2)
mu = 0
u0 = np.sqrt(1 / np.sqrt(np.pi * var)) * np.exp(-np.power(x - mu, 2) / (2 * \
var))
u = np.zeros([len(t), len(x)], dtype = 'complex_')
u[0, :] = u0
# Normalise u.
u[0, :] = u[0, :] / np.sqrt(normf(dx, u, 0))
# Set coefficients of CN scheme.
alpha = dt * -1j / (4 * np.power(dx, 2))
beta = dt * 1j / (4 * np.power(dx, 2))
# Tridiagonal matrices Al and AR. Al to be solved using Thomas algorithm.
Al = np.zeros([len(x), len(x)], dtype = 'complex_')
for i in range (0, M):
Al[i + 1, i] = alpha
Al[i, i] = 1 - (2 * alpha)
Al[i, i + 1] = alpha
# Corner elements for BC's.
Al[M, M], Al[0, 0] = 1 - alpha, 1 - alpha
Ar = np.zeros([len(x), len(x)], dtype = 'complex_')
for i in range (0, M):
Ar[i + 1, i] = beta
Ar[i, i] = 1 - (2 * beta)
Ar[i, i + 1] = beta
# Corner elements for BC's.
Ar[M, M], Ar[0, 0] = 1 - 2*beta, 1 - beta
# Thomas algorithm variables. Following similar naming as in Wiki article.
a = np.diag(Al, -1)
b = np.diag(Al)
c = np.diag(Al, 1)
NT = len(b)
cp = np.zeros(NT - 1, dtype = 'complex_')
for n in range(0, NT - 1):
if n == 0:
cp[n] = c[n] / b[n]
else:
cp[n] = c[n] / (b[n] - (a[n - 1] * cp[n - 1]))
d = np.zeros(NT, dtype = 'complex_')
dp = np.zeros(NT, dtype = 'complex_')
# Iterate over each time step to solve CN method. Maintain boundary
# conditions. Keep track of standard deviation.
for i in range(0, N):
# BC's.
u[i, 0], u[i, M] = 0, 0
# Find RHS.
d = np.dot(Ar, u[i, :])
for n in range(0, NT):
if n == 0:
dp[n] = d[n] / b[n]
else:
dp[n] = (d[n] - (a[n - 1] * dp[n - 1])) / (b[n] - (a[n - 1] * \
cp[n - 1]))
nc = NT - 1
while nc > -1:
if nc == NT - 1:
u[i + 1, nc] = dp[nc]
nc -= 1
else:
u[i + 1, nc] = dp[nc] - (cp[nc] * u[i + 1, nc + 1])
nc -= 1
norm[i] = normf(dx, u, i)
xexp[i] = xexpf(dx, x, u, i)
xexps[i] = xexpsf(dx, x, u, i)
sda[i] = sdaf(xexp, xexps, i)
# Fill in final norm value.
norm[N] = normf(dx, u, N)
# Fill in final position expectation value.
xexp[N] = xexpf(dx, x, u, N)
# Fill in final squared position expectation value.
xexps[N] = xexpsf(dx, x, u, N)
# Fill in final standard deviation value.
sda[N] = sdaf(xexp, xexps, N)
I'm following the Andrew-Ng course on Machine Learning and I'm currently doing the week 5 exercise.
I've found myself stuck on the implementation of the backpropagation algorithm, due to the fact that the relative difference, compared to numerical gradient, is very high (order of 1e-1), but I can't find any error within my implementation, so I'm gently asking if someone could take a look at it and explain what I did wrong.
Forward propagation:
def forward_propagation(thetas, X, history=False):
activation_arr = []
a = X # X is the array of the first activation values
for k in range(0, len(thetas)):
a = add_intercept(a) # add the bias unit
a = sigmoid(a # thetas[k].T)
if history:
activation_arr.append(a)
return activation_arr if history else a
Backpropagation:
def gradient_nn(thetas, X, y, num_labels, reg_lambda=None):
n_examples = X.shape[0]
Y = np.zeros(( # creates a n_examples X num_labels matrix
n_examples, # n of examples
num_labels
))
for i in range(n_examples):
Y[i, y[i, 0]] = 1 # the index corresponding to the correct label for each row has value = 1
# add intercepted X to the activation array
activation_arr = [add_intercept(X)] + forward_propagation(thetas, X, history=True)
sigma = [activation_arr[-1] - Y] # sigma^L = a^L - y
delta = [sigma[-1].T # activation_arr[-2]] # find delta for the first row
thetas_grad = []
# Calculate sigma and delta
for idx in range(1, len(thetas)): # skip last iteration
sigma = [
(sigma[0] # thetas[-idx][:, 1:]) * partial_derivative(activation_arr[-1-idx])
] + sigma
delta = [
sigma[0].T # activation_arr[-2-idx]
] + delta
return [np.sum(d) / n_examples for d in thetas_grad]
Partial derivative:
def partial_derivative(a):
return a * (1 - a) # element wise multiplication
Numerical gradient:
def compute_numerical_gradient(cost_function, thetas):
# Unroll parameters
nn_params = unroll_thetas(thetas)
num_grad = np.zeros(nn_params.shape)
perturb = np.zeros(nn_params.shape)
shapes = [theta.shape for theta in thetas]
epsilon = 1e-4 # not the one of random initialization
for p in range(nn_params.shape[0]):
# Set perturbation vector
perturb[p] = epsilon
minus_theta = nn_params - perturb
plus_theta = nn_params + perturb
# --- Roll params back in order to use the cost function ---
minus_theta = roll_thetas(minus_theta, shapes)
plus_theta = roll_thetas(plus_theta, shapes)
# calculate the loss of the cost function
minus_loss = cost_function(minus_theta)
plus_loss = cost_function(plus_theta)
# Compute Numerical Gradient
num_grad[p] = (plus_loss - minus_loss) / (2 * epsilon)
perturb[p] = 0
num_grad = roll_thetas(num_grad, shapes)
return [np.sum(num_g) for num_g in num_grad]
Cost function:
def J_nn(num_labels, reg_lambda=None):
def non_reg_func(thetas, X, y):
n_examples = X.shape[0]
Y = np.zeros(( # creates a n_examples X num_labels matrix
n_examples, # n of examples
num_labels
))
for i in range(n_examples):
Y[i, y[i, 0]] = 1 # the index corresponding to the correct label for each row has value = 1
prediction = forward_propagation(thetas, X)
return np.sum(np.sum(-Y * np.log(prediction) - (1 - Y) * np.log(1 - prediction))) / n_examples
if reg_lambda is None:
func = non_reg_func
else: # regularization
def func(thetas, X, y):
cost = non_reg_func(thetas, X, y)
for theta in thetas: # regularize for every layer
theta = theta[1:] # remove bias unit
cost = cost + (reg_lambda / (2 * y.shape[0])) * np.sum(np.sum(theta[:, ] ** 2))
return cost
return func
Checking backpropagation with numerical gradient:
def check_nn_gradients(reg_lambda=None):
"""
Creates a small neural network (max 8 x 8 x 7 x 8) and checks that
the implementation of the backpropagation algorithm is good
"""
#n_examples, sizes = random.randint(5, 10), [random.randint(2, 8), random.randint(2, 8), random.randint(1, 8)]
n_examples, sizes = 5, [8, 8, 5, 4]
n_labels = sizes[-1] # Last size is equal to the number of labels
init_epsilon = 0.0001
thetas = random_init_thetas(sizes, init_epsilon)
X = np.array(
random_init_thetas([sizes[0]-1, n_examples], init_epsilon)
).squeeze() # We squeeze it because random_init_thetas returns a 3D array, but we want X to be 2D
y = np.array([random.randint(0, n_labels-1) for _ in X])
y = y[:, np.newaxis]
inner_cost = lambda _thetas: J_nn(n_labels, reg_lambda)(_thetas, X, y)
gradients = gradient_nn(thetas, X, y, n_labels, 0)
unrolled_gradients = unroll_thetas(gradients)
print(unrolled_gradients)
# finite difference method
grad_checking_epsilon = 1e-4
num_grad = compute_numerical_gradient(inner_cost, thetas)
unrolled_num_grad = unroll_thetas(num_grad)
print(unrolled_num_grad)
return diff = np.linalg.norm(unrolled_num_grad - unrolled_gradients) / np.linalg.norm(unrolled_num_grad + unrolled_gradients)
I have a torch tensor with shape (batch_size, number_maps, x_val, y_val). The tensor is normalized with a sigmoid function, so within range [0, 1]. I want to find the covariance for each map, so I want to have a tensor with shape (batch_size, number_maps, 2, 2). As far as I know, there is no torch.cov() function as in numpy. How can I efficiently calculate the covariance without converting it to numpy?
Edit:
def get_covariance(tensor):
bn, nk, w, h = tensor.shape
tensor_reshape = tensor.reshape(bn, nk, 2, -1)
x = tensor_reshape[:, :, 0, :]
y = tensor_reshape[:, :, 1, :]
mean_x = torch.mean(x, dim=2).unsqueeze(-1)
mean_y = torch.mean(y, dim=2).unsqueeze(-1)
xx = torch.sum((x - mean_x) * (x - mean_x), dim=2).unsqueeze(-1) / (h*w - 1)
xy = torch.sum((x - mean_x) * (y - mean_y), dim=2).unsqueeze(-1) / (h*w - 1)
yx = xy
yy = torch.sum((y - mean_y) * (y - mean_y), dim=2).unsqueeze(-1) / (h*w - 1)
cov = torch.cat((xx, xy, yx, yy), dim=2)
cov = cov.reshape(bn, nk, 2, 2)
return cov
I tried the following now, but I m pretty sure it's not correct.
You could try the function suggested on Github:
def cov(x, rowvar=False, bias=False, ddof=None, aweights=None):
"""Estimates covariance matrix like numpy.cov"""
# ensure at least 2D
if x.dim() == 1:
x = x.view(-1, 1)
# treat each column as a data point, each row as a variable
if rowvar and x.shape[0] != 1:
x = x.t()
if ddof is None:
if bias == 0:
ddof = 1
else:
ddof = 0
w = aweights
if w is not None:
if not torch.is_tensor(w):
w = torch.tensor(w, dtype=torch.float)
w_sum = torch.sum(w)
avg = torch.sum(x * (w/w_sum)[:,None], 0)
else:
avg = torch.mean(x, 0)
# Determine the normalization
if w is None:
fact = x.shape[0] - ddof
elif ddof == 0:
fact = w_sum
elif aweights is None:
fact = w_sum - ddof
else:
fact = w_sum - ddof * torch.sum(w * w) / w_sum
xm = x.sub(avg.expand_as(x))
if w is None:
X_T = xm.t()
else:
X_T = torch.mm(torch.diag(w), xm).t()
c = torch.mm(X_T, xm)
c = c / fact
return c.squeeze()
https://github.com/pytorch/pytorch/issues/19037
I try to create a function for performing a convolution between a matrix and a filter. I managed to do the basic operations, but I stumbled on calculating the norm of the sliced matrix (the submatrix of the main matrix), corresponding to each position in the output.
The code is this:
def convol2d(matrix, kernel):
# matrix - input matrix indexed (v, w)
# kernel - filtre indexed (s, t),
# h -output indexed (x, y),
# The output size is calculated by adding smid, tmid to each side of the dimensions of the input image.
norm_filter = np.linalg.norm(kernel) # The norm of the filter
vmax = matrix.shape[0]
wmax = matrix.shape[1]
smax = kernel.shape[0]
tmax = kernel.shape[1]
smid = smax // 2
tmid = tmax // 2
xmax = vmax + 2 * smid
ymax = wmax + 2 * tmid
window_list = [] # Initialized an empty list for storing the submatrix
print vmax
print xmax
h = np.zeros([xmax, ymax], dtype=np.float)
for x in range(xmax):
for y in range(ymax):
s_from = max(smid - x, -smid)
s_to = min((xmax - x) - smid, smid + 1)
t_from = max(tmid - y, -tmid)
t_to = min((ymax - y) - tmid, tmid + 1)
value = 0
for s in range(s_from, s_to):
for t in range(t_from, t_to):
v = x - smid + s
w = y - tmid + t
print matrix[v, w]
value += kernel[smid - s, tmid - t] * matrix[v, w]
# This does not work
window_list.append(matrix[v,w])
norm_window = np.linalg.norm(window_list)
h[x, y] = value / norm_filter * norm_window
return h
For example, my input matrix is A(v, w), I want that my output values in the output matrix h (x,y), be calculated as:
h(x,y) = value/ (norm_of_filer * norm_of_sumbatrix)
Thanks for any help!
Edit: Following the suggestions, I modified like this:
I modified like this, but I only get the first row appended, and used in calculation and not the entire submatrix.
`for s in range(s_from, s_to):
for t in range(t_from, t_to):
v = x - smid + s
w = y - tmid + t
value += kernel[smid - s, tmid - t] * matrix[v, w]
window_list.append(matrix[v,w])
window_array = np.asarray(window_list, dtype=float)
window_list = []
norm_window = np.linalg.norm(window_array)
h[x, y] = value / norm_filter * norm_window`
The input of np.linalg.norm is supposed to be an "Input array." Try converting the list of matrices to an array. (python: list of matrices to numpy array?)
Also, maybe move the norm_window line out of the loop, since you only later use it as evaluated at the last step, with everything in it. In fact, wait 'til the loop is done, convert the finished list to an array (so it's only done once) and evaluate norm_window on that.