what should be the parametric form of the l2 regularization in a Bayesian setting?

Question

In a Bayesian setting for parameter estimation, what should be the parametric form of the prior distribution in order to perform l2 regularization?

Answer 1

L2 regularization is equivalent to a Gaussian prior.

See, for example, Jason Rennie's On L2-norm Regularization and the Gaussian Prior (2003).

(For future reference, this kind of question should probably go to http://stats.stackexchange.com . . . where I see you also posted this. Don't do that.)