19.2 Expectation Maximization的第一段最后一句

19.2 Expectation Maximization的第一段的最后一句是：Unlike most of the other algorithms we describe in this chapter, EM is not an approach to approximate inference, but rather an approach to learning with an approximate posterior.
an approach to learning with an approximate posterior的意思我觉得应该是“用一个近似后验来学习的方法”。
中文版的翻译是：一种能够学到近似后验的算法。
如果是这个中文意思的话，英文应该是an approach to learning an approximate posterior吧。
觉得这两个意思差别挺大的，请大神确认一下