Deep Learning

Concept

Deep learning (also known as deep structured learning) is part of a broader family of machine learning methods based on artificial neural networks with representation learning. Learning can be supervised, semi-supervised or unsupervised.

Deep learning (also known as deep structured learning) is part of a broader family of machine learning methods based on artificial neural networks with representation learning. Learning can be supervised, semi-supervised or unsupervised.

Artificial neural networks (ANNs) were inspired by information processing and distributed communication nodes in biological systems. ANNs have various differences from biological brains. Specifically, neural networks tend to be static and symbolic, while the biological brain of most living organisms is dynamic (plastic) and analog.

Different from neural networks

Deep learning本身算是machine learning的一个分支，简单可以理解为neural network的发展。

Deep learning与传统的神经网络之间有相同的地方也有很多不同。 
二者的相同在于deep learning采用了神经网络相似的分层结构，系统由包括输入层、隐层（多层）、输出层组成的多层网络，只有相邻层节点之间有连接，同一层以及跨层节点之间相互无连接，每一层可以看作是一个logistic regression模型；这种分层结构，是比较接近人类大脑的结构的。

而为了克服神经网络训练中的问题，DL采用了与神经网络很不同的训练机制。传统神经网络中，采用的是back propagation的方式进行，简单来讲就是采用迭代的算法来训练整个网络，随机设定初值，计算当前网络的输出，然后根据当前输出和label之间的差去改变前面各层的参数，直到收敛（整体是一个梯度下降法）。而deep learning整体上是一个layer-wise的训练机制。这样做的原因是因为，如果采用back propagation的机制，对于一个deep network（7层以上），残差传播到最前面的层已经变得太小，出现所谓的gradient diffusion（梯度扩散）。 
传统神经网络的训练方法为什么不能用在深度神经网络。 
BP算法作为传统训练多层网络的典型算法，实际上对仅含几层网络，该训练方法就已经很不理想。深度结构（涉及多个非线性处理单元层）非凸目标代价函数中普遍存在的局部最小是训练困难的主要来源。