def test_rff_feature_map_derivative2_d():
X = np.array([[1.]])
u = np.array([2.])
omega = np.array([[2.]])
d = 0
phi_derivative2 = rff_feature_map_grad2_d(X, omega, u, d)
phi_derivative2_manual = -rff_feature_map(X, omega, u) * (omega[:, d] ** 2)
assert_close(phi_derivative2, phi_derivative2_manual)
def test_rff_feature_map_derivative2_d():
X = np.array([[1.]])
u = np.array([2.])
omega = np.array([[2.]])
d = 0
phi_derivative2 = rff_feature_map_grad2_d(X, omega, u, d)
phi_derivative2_manual = -rff_feature_map(X, omega, u) * (omega[:, d]**2)
assert_close(phi_derivative2, phi_derivative2_manual)
def compute_b(X, omega, u):
assert len(X.shape) == 2
m = 1 if np.isscalar(u) else len(u)
D = X.shape[1]
projections_sum = np.zeros(m)
Phi2 = rff_feature_map(X, omega, u)
for d in range(D):
projections_sum += np.mean(-Phi2 * (omega[d, :]**2), axis=0)
return -projections_sum
def compute_b(X, omega, u):
assert len(X.shape) == 2
m = 1 if np.isscalar(u) else len(u)
D = X.shape[1]
projections_sum = np.zeros(m)
Phi2 = rff_feature_map(X, omega, u)
for d in range(D):
projections_sum += np.mean(-Phi2 * (omega[d, :] ** 2), axis=0)
return -projections_sum
def rff_feature_map_grad2_loop(X, omega, u):
m = 1 if np.isscalar(u) else len(u)
N = X.shape[0]
D = X.shape[1]
projections = np.zeros((D, N, m))
Phi2 = rff_feature_map(X, omega, u)
for d in range(D):
projections[d, :, :] = -Phi2
projections[d, :, :] *= omega[d, :] ** 2
return projections
def rff_feature_map_grad2_loop(X, omega, u):
m = 1 if np.isscalar(u) else len(u)
N = X.shape[0]
D = X.shape[1]
projections = np.zeros((D, N, m))
Phi2 = rff_feature_map(X, omega, u)
for d in range(D):
projections[d, :, :] = -Phi2
projections[d, :, :] *= omega[d, :] ** 2
return projections
def test_rff_feature_map_single_equals_feature_map():
N = 10
D = 20
m = 3
X = np.random.randn(N, D)
omega = np.random.randn(D, m)
u = np.random.uniform(0, 2 * np.pi, m)
phis = rff_feature_map(X, omega, u)
for i, x in enumerate(X):
phi = rff_feature_map_single(x, omega, u)
assert_allclose(phis[i], phi)
def test_rff_feature_map_single_equals_feature_map():
N = 10
D = 20
m = 3
X = np.random.randn(N, D)
omega = np.random.randn(D, m)
u = np.random.uniform(0, 2 * np.pi, m)
phis = rff_feature_map(X, omega, u)
for i, x in enumerate(X):
phi = rff_feature_map_single(x, omega, u)
assert_allclose(phis[i], phi)
def log_pdf_multiple(self, X):
assert_array_shape(X, ndim=2, dims={1: self.D})
Phi = rff_feature_map(X, self.omega, self.u)
return np.dot(Phi, self.theta)
def log_pdf_multiple(self, X):
assert_array_shape(X, ndim=2, dims={1: self.D})
Phi = rff_feature_map(X, self.omega, self.u)
return np.dot(Phi, self.theta)
