Newbie here:

I’d like to do the following in Theano:

take a vector of inputs x of length N, and evaluate a sum of M scaled Gaussians at each value

If each Gaussian is parameterised by an M-vector of locations/mu’s and an M-vector of scales ( all gaussians having same sigma) how can I evaluate a y_est vector like:

y_est = pm.Normal(mu=T.dot(scales, <

gaussian evaluated at each y>)