score.logV: Gradient of the variational likelihood bound w.r.t. log of...
In wrbrooks/cox: Estimate Cox process regression models

Gradient of the variational likelihood bound w.r.t. log of covariance matrix

1	score.logV(logV, y, X, S, beta, wt, ltau, M)

`logV`	logarithm of vectorized upper triangle of the covariance matrix of random-effect coefficients for the variational approximation
`y`	vector of Poisson-distributed responses for DWPR
`X`	matrix of fixed-effect covariates
`S`	matrix of random-effect covariates
`beta`	fixed-effect coefficient vector
`wt`	vector of observation weights
`ltau`	log precision of the random effect coefficients
`M`	vector of means of random-effect coefficients for the variational approximation