【UFLDL-notes of attached materials(to be continued)】

Data Preprocessing:
对基于正交化ICA的模型来说,保证输入数据尽可能地白化(即协方差矩阵为单位矩阵)非常重要。这是因为:这类模型需要对学习到的特征做正交化,以解除不同维度之间的相关性(详细内容请参考 ICA 一节)。因此在这种情况下,epsilon 要足够小(比如 epsilon = 1e − 6)。??

对于大图像,采用基于 PCA/ZCA 的白化方法是不切实际的,因为协方差矩阵太大。在这些情况下我们退而使用 1/f 白化方法(更多内容后续再讲)。??

Deriving gradients using the backpropagation idea:
example 3 还没看??
example 2 也有点迷糊

Sparse Coding:
两个To do 没有搞懂:
1.To increase the rate of convergence, you can instead run the algorithm on mini-batches instead…(TODO: explain why).

2.Very roughly and informally speaking, this initialization helps because the first step is an attempt to find a good s such that , and the second step “normalizes” s in an attempt to keep the sparsity penalty small. It turns out that initializing s using only one but not both steps results in poor performance in practice. (TODO: a better explanation for why this initialization helps?)

Exercise:Sparse Coding 还没做??

另外ICA还没看,一个新的网站还有增减的内容http://ufldl.stanford.edu/tutorial/unsupervised/ExerciseRICA/

5月中旬答辩,重心转移到毕设上去。

你可能感兴趣的:(【UFLDL-notes of attached materials(to be continued)】)