Approximate message-passing for convex optimization with non-separable penalties

Open in new window