Input
x
Neural network
f
Representation
(Summary statistic)
r = f(x)
Output
Modelling cross-correlations
Implicit likelihood inference with normalising flows
No assumptions on the likelihood (likelihoods rarely Gaussian!)
No expensive MCMC chains needed to estimate posterior
Input
Output
Rotation and translation Invariant
Summarising with graph neural networks
edge embedding
node embedding
summary statistic
edge embedding
node embedding
summary statistic