We train neural nets on the same training sets for multiple iterations and expect it to come up with new/better connections. When learning a new skill we may meditate on previously solved problems and discover better solutions. Why would the immune system not have the capability to continue experimenting and learning from the same stimulus?