model based optimize? discriminative learning?有何区别？

如题，究竟有何区别呢？在看图像复原（超分）的论文IRCNN(Image Restoration by Convolution Neural Network)的时候，有所领悟哈，觉得写的很明白，只可意会，不可言传，翻译如下：

首先，明确两个英文单词：

matrix : 矩阵

matrices: matrix的复数形式，多个矩阵

The model　based optimization methods aim to directly solve Eqn. (2) with some optimization algorithms which usually involve a time-consuming iterative inference. On the contrary, discriminative learning methods try to learn the prior parameters Θ and a compact inference through an optimization of a loss function on a training set containing degraded-clean image pairs 

翻译：model-based优化方法，直接针对公式（２）（损失函数）进行反复跌倒优化（耗时训练）．相反，discriminative learning尝试优化公式（３）中的先验项参数Θ

It can be seen that one obvious difference between model-based optimization method and discriminative learning method is that, the former is flexible to handle various IR tasks by specifying degradation matrix H, whereas the later needs to use the training data with certain degradation matrices to learn the model. As a consequence, different from model-based optimization methods which have flexibility to handle different IR tasks, discriminative learning methods are usually restricted by specialized tasks. 

容易看出，model-based optimization method通过（同一个）降质矩阵Ｈ处理不同的IR任务（一个人能干多件事，flexibility），

discriminative learning则需要certain degradation matrices（多个降质矩阵）来学习不同的任务（每个任务受限于自己的降质矩阵H）

As a result, those two kinds of methods have their respective merits and drawbacks, and thus it would be attractive to investigate their integration which leverages their respective merits.

这两种方法各有优缺点，通常结合使用．