The reverse KL divergence is defined as KL(Pg‖Pr). The Jenson-Shannon (JS) divergence: Let M be the mixture distribution M=Pr/2+Pg/2. Then.
確定! 回上一頁