def batch_predict(SA):
sa = np.mean(SA, 0)
idx = kdt.query(sa.reshape(1, -1), k=K, return_distance=False)
X_nn = Xtrain[idx, :].reshape(K, state_action_dim)
Y_nn = Ytrain[idx, :].reshape(K, state_dim)
if useDiffusionMaps:
X_nn, Y_nn = reduction(sa, X_nn, Y_nn)
m = np.zeros((SA.shape[0], state_dim))
s = np.zeros((SA.shape[0], state_dim))
for i in range(state_dim):
if i == 0:
gp_est = GaussianProcess(X_nn[:, :4],
Y_nn[:, i],
optimize=False,
theta=None)
theta = gp_est.cov.theta
else:
gp_est = GaussianProcess(X_nn[:, :4],
Y_nn[:, i],
optimize=False,
theta=theta)
mm, ss = gp_est.batch_predict(SA[:, :4])
m[:, i] = mm
s[:, i] = np.diag(ss)
return m, s
def batch_predict(self, SA):
sa = np.mean(SA, 0)
# Theta, K = self.get_theta(sa) # Get hyper-parameters for this query point
K = 1
idx = self.kdt.query(sa.reshape(1, -1), k=K, return_distance=False)
X_nn = self.Xtrain[idx, :].reshape(K, self.state_action_dim)
Y_nn = self.Ytrain[idx, :].reshape(K, self.state_dim)
if useDiffusionMaps:
X_nn, Y_nn = self.reduction(sa, X_nn, Y_nn)
dS_next = np.zeros((SA.shape[0], self.state_dim))
std_next = np.zeros((SA.shape[0], self.state_dim))
for i in range(self.state_dim):
gp_est = GaussianProcess(X_nn[:, :self.state_action_dim],
Y_nn[:, i],
optimize=True,
theta=None,
algorithm='Matlab')
mm, vv = gp_est.batch_predict(SA[:, :self.state_action_dim])
dS_next[:, i] = mm
std_next[:, i] = np.sqrt(np.diag(vv))
S_next = SA[:, :self.
state_dim] + dS_next #np.random.normal(dS_next, std_next)
return S_next
def batch_predict_iterative(self, SA):
S_next = []
while SA.shape[0]:
sa = np.copy(SA[np.random.randint(SA.shape[0]), :])
Theta, K = self.get_theta(
sa) # Get hyper-parameters for this query point
D, idx = self.kdt.query(sa.reshape(1, -1),
k=K,
return_distance=True)
r = np.max(D) * 1.1
X_nn = self.Xtrain[idx, :].reshape(K, self.state_action_dim)
Y_nn = self.Ytrain[idx, :].reshape(K, self.state_dim)
neigh = NearestNeighbors(radius=r)
neigh.fit(SA)
idx_local = neigh.radius_neighbors(sa.reshape(1, -1),
return_distance=False)[0]
SA_local = np.copy(SA[idx_local, :])
SA = np.delete(SA, idx_local, axis=0)
if useDiffusionMaps:
X_nn, Y_nn = self.reduction(sa, X_nn, Y_nn)
dS_next = np.zeros((SA_local.shape[0], self.state_dim))
std_next = np.zeros((SA_local.shape[0], self.state_dim))
for i in range(self.state_dim):
gp_est = GaussianProcess(X_nn[:, :self.state_action_dim],
Y_nn[:, i],
optimize=False,
theta=Theta[i],
algorithm='Matlab')
mm, vv = gp_est.batch_predict(
SA_local[:, :self.state_action_dim])
dS_next[:, i] = mm
std_next[:, i] = np.sqrt(np.diag(vv))
S_next_local = SA_local[:, :self.state_dim] + np.random.normal(
dS_next, std_next)
for s in S_next_local:
S_next.append(s)
return np.array(S_next)
m, s = gpup_est.predict(x_n, var_n)
# We sample N particles in the initial distribution to validate the GPUP computation
N = int(1e4)
X_belief = np.array([np.random.normal(x_n, np.sqrt(var_n))
for _ in range(N)]).reshape(N, 1) #
ax4 = plt.subplot2grid((3, 5), (2, 2), colspan=3, rowspan=1)
plt.plot(X_belief, np.tile(0., N), '.k', label='input particles')
x = np.linspace(0, 6, 1000).reshape(-1, 1)
plt.plot(x, scipy.stats.norm.pdf(x, x_n, np.sqrt(var_n)), label='input dist.')
plt.xlabel('x')
plt.legend()
# Propagate all particles through the GP to get an approximated distribution
ax2 = plt.subplot2grid((3, 5), (0, 0), colspan=1, rowspan=2)
means_b, variances_b = gp_est.batch_predict(
X_belief) # Use the GP batch prediction
variances_b = np.diag(variances_b)
Y_belief = np.array([
np.random.normal(means_b[i], np.sqrt(variances_b[i])) for i in range(N)
]).reshape(N, 1) #
plt.plot(np.tile(0., N), Y_belief, '.k', label='output particles')
plt.ylabel('p(y)')
ylim = ax1.get_ylim()
mu_Y = np.mean(Y_belief)
sigma2_Y = np.std(Y_belief)
y = np.linspace(ylim[0], ylim[1], 100000).reshape(-1, 1)
plt.plot(scipy.stats.norm.pdf(y, mu_Y, sigma2_Y),
y,
'-b',
label='particles output dist.')
