apunkt/GPy
1
0
Fork 0
mirror of https://github.com/SheffieldML/GPy.git synced 2026-05-09 03:52:39 +02:00
Code Issues Projects Releases Packages Wiki Activity Actions

t of the gradients are now working in the uncollapsed sparse GP: one

Browse source 
term still to do.
This commit is contained in:
James Hensman 2012-12-09 14:41:36 -08:00
parent ef26888b78
commit 053ebbe2d4
1 changed files with 8 additions and 8 deletions
Download patch file Download diff file Expand all files Collapse all files

16
GPy/models/uncollapsed_sparse_GP.py
View file

@ -39,7 +39,7 @@ class uncollapsed_sparse_GP(sparse_GP_regression):
self.M = Z.shape[0] self.M = Z.shape[0]
else: else:
self.M = M self.M = M
q_u = np.hstack((np.zeros(self.M*self.D),-0.5*np.eye(self.M).flatten())) q_u = np.hstack((np.random.randn(self.M*self.D),-0.5*np.eye(self.M).flatten()))
self.set_vb_param(q_u) self.set_vb_param(q_u)
sparse_GP_regression.__init__(self, X, Y, M=M,*args, **kwargs) sparse_GP_regression.__init__(self, X, Y, M=M,*args, **kwargs)
@ -57,12 +57,12 @@ class uncollapsed_sparse_GP(sparse_GP_regression):
# Compute dL_dpsi # Compute dL_dpsi
self.dL_dpsi0 = - 0.5 * self.D * self.beta * np.ones(self.N) self.dL_dpsi0 = - 0.5 * self.D * self.beta * np.ones(self.N)
self.dL_dpsi1 = np.dot(self.VmT,self.Kmmi).T self.dL_dpsi1 = np.dot(self.VmT,self.Kmmi).T*0 # This is the correct term for E I think...
self.dL_dpsi2 = - 0.5 * self.beta * (-self.D*self.Kmmi + mdot(self.Kmmi,self.q_u_expectation[1],self.Kmmi)) self.dL_dpsi2 = 0.5 * self.beta * self.D * (self.Kmmi - mdot(self.Kmmi,self.q_u_expectation[1],self.Kmmi))
# Compute dL_dKmm # Compute dL_dKmm
tmp = np.dot(self.A,self.Kmmi) tmp = -0.5*self.beta*self.D*self.psi2 -0.5*self.D*self.Kmm +0.5*self.D*(self.q_u_expectation[1]) + 0.5*self.beta*mdot(self.psi2,self.Kmmi,self.q_u_expectation[1]) + 0.5*self.beta*mdot(self.q_u_expectation[1],self.Kmmi, self.psi2)
tmp = 0.5*np.dot(np.eye(self.M) + tmp + tmp.T, self.q_u_expectation[1]) -0.5*self.Kmm - np.dot(self.psi1,self.VmT) #tmp = - np.dot(self.q_u_expectation[0],self.psi1V.T)
self.dL_dKmm = mdot(self.Kmmi,tmp,self.Kmmi) self.dL_dKmm = mdot(self.Kmmi,tmp,self.Kmmi)
def log_likelihood(self): def log_likelihood(self):
@ -71,10 +71,10 @@ class uncollapsed_sparse_GP(sparse_GP_regression):
""" """
A = -0.5*self.N*self.D*(np.log(2.*np.pi) - np.log(self.beta)) A = -0.5*self.N*self.D*(np.log(2.*np.pi) - np.log(self.beta))
B = -0.5*self.beta*self.D*self.trace_K B = -0.5*self.beta*self.D*self.trace_K
C = -0.5*self.D *(self.Kmm_logdet-self.q_u_logdet + np.sum(self.Lambda * self.q_u_expectation[1]) - self.M*self.D) C = -0.5*self.D *(self.Kmm_logdet-self.q_u_logdet + np.sum(self.Lambda * self.q_u_expectation[1]) - self.M)
D = -0.5*self.beta*self.trYYT D = -0.5*self.beta*self.trYYT
E = np.sum(np.dot(self.V.T,self.projected_mean)) E = np.sum(np.dot(self.V.T,self.projected_mean))
return A+B+C+D+E return A+B+C+D
def dL_dbeta(self): def dL_dbeta(self):
""" """
@ -87,7 +87,7 @@ class uncollapsed_sparse_GP(sparse_GP_regression):
dD_dbeta = - 0.5 * self.trYYT dD_dbeta = - 0.5 * self.trYYT
dE_dbeta = np.sum(np.dot(self.Y.T,self.projected_mean)) dE_dbeta = np.sum(np.dot(self.Y.T,self.projected_mean))
return np.squeeze(dA_dbeta + dB_dbeta + dC_dbeta + dD_dbeta) return np.squeeze(dA_dbeta + dB_dbeta + dC_dbeta + dD_dbeta + dE_dbeta)
def _raw_predict(self, Xnew, slices): def _raw_predict(self, Xnew, slices):
"""Internal helper function for making predictions, does not account for normalisation""" """Internal helper function for making predictions, does not account for normalisation"""