apunkt/GPy
1
0
Fork 0
mirror of https://github.com/SheffieldML/GPy.git synced 2026-05-18 13:55:14 +02:00
Code Issues Projects Releases Packages Wiki Activity Actions

Some changes to speed up... just a little

Browse source
This commit is contained in:
Ricardo 2013-05-15 16:26:33 +01:00
parent 2052463ba7
commit 9431cb2d55
1 changed files with 22 additions and 15 deletions
Download patch file Download diff file Expand all files Collapse all files

37
GPy/likelihoods/EP.py
View file

@ -1,6 +1,6 @@
import numpy as np
from scipy import stats, linalg
from ..util.linalg import pdinv,mdot,jitchol,DSYR
from ..util.linalg import pdinv,mdot,jitchol,DSYR,tdot
from likelihood import likelihood
class EP(likelihood):
@ -117,8 +117,6 @@ class EP(likelihood):
self.v_tilde[i] += Delta_v
#Posterior distribution parameters update
DSYR(Sigma,Sigma[:,i].copy(), -float(Delta_tau/(1.+ Delta_tau*Sigma[i,i])))
#si=Sigma[:,i:i+1]
#Sigma -= Delta_tau/(1.+ Delta_tau*Sigma[i,i])*np.dot(si,si.T)#DSYR
mu = np.dot(Sigma,self.v_tilde)
self.iterations += 1
#Sigma recomptutation with Cholesky decompositon
@ -135,12 +133,13 @@ class EP(likelihood):
return self._compute_GP_variables()
#def fit_DTC(self, Knn_diag, Kmn, Kmm):
#@profile
def fit_DTC(self, Kmm, Kmn):
"""
The expectation-propagation algorithm with sparse pseudo-input.
For nomenclature see ... 2013.
"""
M = Kmm.shape[0]
#TODO: this doesn't work with uncertain inputs!
@ -149,12 +148,20 @@ class EP(likelihood):
q(f|X) = int_{df}{N(f|KfuKuu_invu,diag(Kff-Qff)*N(u|0,Kuu)} = N(f|0,Sigma0)
Sigma0 = Qnn = Knm*Kmmi*Kmn
"""
Kmmi, Lm, Lmi, Kmm_logdet = pdinv(Kmm)
KmnKnm = np.dot(Kmn, Kmn.T)
KmnKnm = np.dot(Kmn,Kmn.T)
Lm = jitchol(Kmm)
Lmi,info = linalg.lapack.flapack.dtrtrs(Lm,np.asfortranarray(np.eye(M)),lower=1,trans=1)
Kmmi = np.dot(Lmi.T,Lmi)
KmmiKmn = np.dot(Kmmi,Kmn)
Qnn_diag = np.sum(Kmn*KmmiKmn,-2)
LLT0 = Kmm.copy()
#Kmmi, Lm, Lmi, Kmm_logdet = pdinv(Kmm)
#KmnKnm = np.dot(Kmn, Kmn.T)
#KmmiKmn = np.dot(Kmmi,Kmn)
#Qnn_diag = np.sum(Kmn*KmmiKmn,-2)
#LLT0 = Kmm.copy()
"""
Posterior approximation: q(f|y) = N(f| mu, Sigma)
Sigma = Diag + P*R.T*R*P.T + K
@ -197,19 +204,19 @@ class EP(likelihood):
#Site parameters update
Delta_tau = self.delta/self.eta*(1./sigma2_hat[i] - 1./Sigma_diag[i])
Delta_v = self.delta/self.eta*(mu_hat[i]/sigma2_hat[i] - mu[i]/Sigma_diag[i])
self.tau_tilde[i] = self.tau_tilde[i] + Delta_tau
self.v_tilde[i] = self.v_tilde[i] + Delta_v
self.tau_tilde[i] += Delta_tau
self.v_tilde[i] += Delta_v
#Posterior distribution parameters update
LLT = LLT + np.outer(Kmn[:,i],Kmn[:,i])*Delta_tau
DSYR(LLT,Kmn[:,i].copy(),Delta_tau) #LLT = LLT + np.outer(Kmn[:,i],Kmn[:,i])*Delta_tau
L = jitchol(LLT)
#cholUpdate(L,Kmn[:,i]*np.sqrt(Delta_tau))
V,info = linalg.lapack.flapack.dtrtrs(L,Kmn,lower=1)
Sigma_diag = np.sum(V*V,-2)
si = np.sum(V.T*V[:,i],-1)
mu = mu + (Delta_v-Delta_tau*mu[i])*si
mu += (Delta_v-Delta_tau*mu[i])*si
self.iterations += 1
#Sigma recomputation with Cholesky decompositon
LLT0 = LLT0 + np.dot(Kmn*self.tau_tilde[None,:],Kmn.T)
LLT = LLT0 + np.dot(Kmn*self.tau_tilde[None,:],Kmn.T)
L = jitchol(LLT)
V,info = linalg.lapack.flapack.dtrtrs(L,Kmn,lower=1)
V2,info = linalg.lapack.flapack.dtrtrs(L.T,V,lower=0)
@ -290,8 +297,8 @@ class EP(likelihood):
#Site parameters update
Delta_tau = self.delta/self.eta*(1./sigma2_hat[i] - 1./Sigma_diag[i])
Delta_v = self.delta/self.eta*(mu_hat[i]/sigma2_hat[i] - mu[i]/Sigma_diag[i])
self.tau_tilde[i] = self.tau_tilde[i] + Delta_tau
self.v_tilde[i] = self.v_tilde[i] + Delta_v
self.tau_tilde[i] += Delta_tau
self.v_tilde[i] += Delta_v
#Posterior distribution parameters update
dtd1 = Delta_tau*Diag[i] + 1.
dii = Diag[i]
@ -301,8 +308,8 @@ class EP(likelihood):
Rp_i = np.dot(R,pi_.T)
RTR = np.dot(R.T,np.dot(np.eye(M) - Delta_tau/(1.+Delta_tau*Sigma_diag[i]) * np.dot(Rp_i,Rp_i.T),R))
R = jitchol(RTR).T
self.w[i] = self.w[i] + (Delta_v - Delta_tau*self.w[i])*dii/dtd1
self.gamma = self.gamma + (Delta_v - Delta_tau*mu[i])*np.dot(RTR,P[i,:].T)
self.w[i] += (Delta_v - Delta_tau*self.w[i])*dii/dtd1
self.gamma += (Delta_v - Delta_tau*mu[i])*np.dot(RTR,P[i,:].T)
RPT = np.dot(R,P.T)
Sigma_diag = Diag + np.sum(RPT.T*RPT.T,-1)
mu = self.w + np.dot(P,self.gamma)