Does maximum likelihood inference have a support problem? Maximum likelihood (ML) has a problem with parameters that take values in open sets (Is that all of them? Almost!). Bayesian inference doesn't obviously have this problem. Briefly, using max...