一、Dropout
随机失活是一种 简单但非常有效的 神经网络训练效果提升技巧,原理大概是它在一定程度上避免了某些特定特征组合对训练造成的负面影响。
在正向传播时随机挑选一部分神经元失活。在反向传播时梯度只流经没有失活的神经元。
注意:
在测试时候要对随机失活进行补益。比如说随机失活了P=0.5的神经元,那么在最后测试的结果上乘以P=0.5,测试集结果与训练集结果会有一半的误差,极大的误差,就没什么准确度可言了。
def dropout_forward(x, dropout_param):
"""
Inputs:
- x: Input data, of any shape
- dropout_param: A dictionary with the following keys:
- p: Dropout parameter. We keep each neuron output with probability p.
- mode: 'test' or 'train'. If the mode is train, then perform dropout;
if the mode is test, then just return the input.
- seed: Seed for the random number generator. Passing seed makes this
function deterministic, which is needed for gradient checking but not
in real networks.
Outputs:
- out: Array of the same shape as x.
- cache: tuple (dropout_param, mask). In training mode, mask is the dropout
mask that was used to multiply the input; in test mode, mask is None.
NOTE: Please implement **inverted** dropout, not the vanilla version of dropout.
See http://cs231n.github.io/neural-networks-2/#reg for more details.
NOTE 2: Keep in mind that p is the probability of **keep** a neuron
output; this might be contrary to some sources, where it is referred to
as the probability of dropping a neuron output.
"""
p, mode = dropout_param['p'], dropout_param['mode']
if 'seed' in dropout_param:
np.random.seed(dropout_param['seed'])
mask = None
out = None
if mode == 'train':
# 建立蒙板,取整个矩阵的1-p部分
# 为保证不影响最后结果x.W,要还原一下/(1-p)
mask = (np.random.rand(*x.shape) > p) / (1 - p)
out = mask * x
elif mode == 'test':
out = x
cache = (dropout_param, mask)
out = out.astype(x.dtype, copy=False)
return out, cache
def dropout_backward(dout, cache):
"""
Inputs:
- dout: Upstream derivatives, of any shape
- cache: (dropout_param, mask) from dropout_forward.
"""
dropout_param, mask = cache
mode = dropout_param['mode']
dx = None
if mode == 'train':
dx = dout * mask
elif mode == 'test':
dx = dout
return dx
二、Batch Normalization
机器学习中有一种思想是均值为0,方差1的非相关性数据作为输入时往往能产生更好的结果。对于神经网络而言,即使输入数据X满足条件但是经过几层之后甚至一层之后这种条件就被破坏了,于是乎就有了Batch Normalization。
BN可以很好的提升神经网络的性能,加速收敛,提高准确度,一般有了BN就不用Dropout了。
每一列为单位,前两行求整个数据集的均值方差,第三行是归一化数据集,这个公式很像高斯分布的指数部分,最后一步是重构数据集。
比照图2,写出各个部分的微分,便得到了dx。
def batchnorm_forward(x, gamma, beta, bn_param):
"""
Forward pass for batch normalization.
During training the sample mean and (uncorrected) sample variance are
computed from minibatch statistics and used to normalize the incoming data.
During training we also keep an exponentially decaying running mean of the
mean and variance of each feature, and these averages are used to normalize
data at test-time.
At each timestep we update the running averages for mean and variance using
an exponential decay based on the momentum parameter:
running_mean = momentum * running_mean + (1 - momentum) * sample_mean
running_var = momentum * running_var + (1 - momentum) * sample_var
Note that the batch normalization paper suggests a different test-time
behavior: they compute sample mean and variance for each feature using a
large number of training images rather than using a running average. For
this implementation we have chosen to use running averages instead since
they do not require an additional estimation step; the torch7
implementation of batch normalization also uses running averages.
Input:
- x: Data of shape (N, D)
- gamma: Scale parameter of shape (D,)
- beta: Shift paremeter of shape (D,)
- bn_param: Dictionary with the following keys:
- mode: 'train' or 'test'; required
- eps: Constant for numeric stability
- momentum: Constant for running mean / variance.
- running_mean: Array of shape (D,) giving running mean of features
- running_var Array of shape (D,) giving running variance of features
Returns a tuple of:
- out: of shape (N, D)
- cache: A tuple of values needed in the backward pass
"""
mode = bn_param['mode']
eps = bn_param.get('eps', 1e-5)
momentum = bn_param.get('momentum', 0.9)
N, D = x.shape
running_mean = bn_param.get('running_mean', np.zeros(D, dtype=x.dtype))
running_var = bn_param.get('running_var', np.zeros(D, dtype=x.dtype))
out, cache = None, None
if mode == 'train':
sample_mean = np.mean(x, axis=0)
sample_var = np.var(x, axis=0)
x_norm = (x - sample_mean) / np.sqrt(sample_var + eps)
out = gamma * x_norm + beta # 到此处,这轮batch norm就做完了
# 更新整个数据集的mean,var
running_mean = momentum * running_mean + (1 - momentum) * sample_mean
running_var = momentum * running_var + (1 - momentum) * sample_var
bn_param['running_mean'] = running_mean
bn_param['running_var'] = running_var
# 把中间数据存入cache,反向传播时用
cache = {
'x_minus_mean': (x - sample_mean),
'x_norm': x_norm,
'gamma': gamma,
'i_var': 1 / np.sqrt(sample_var + eps),
'sqrt_var': np.sqrt(sample_var + eps),
}
elif mode == 'test':
#######################################################################
# TODO: Implement the test-time forward pass for batch normalization. #
# Use the running mean and variance to normalize the incoming data, #
# then scale and shift the normalized data using gamma and beta. #
# Store the result in the out variable. #
#######################################################################
# pass
out = gamma * ((x - running_mean) / np.sqrt(running_var + eps)) + beta
#######################################################################
# END OF YOUR CODE #
#######################################################################
else:
raise ValueError('Invalid forward batchnorm mode "%s"' % mode)
# Store the updated running means back into bn_param
bn_param['running_mean'] = running_mean
bn_param['running_var'] = running_var
return out, cache
def batchnorm_backward(dout, cache):
"""
Backward pass for batch normalization.
For this implementation, you should write out a computation graph for
batch normalization on paper and propagate gradients backward through
intermediate nodes.
Inputs:
- dout: Upstream derivatives, of shape (N, D)
- cache: Variable of intermediates from batchnorm_forward.
Returns a tuple of:
- dx: Gradient with respect to inputs x, of shape (N, D)
- dgamma: Gradient with respect to scale parameter gamma, of shape (D,)
- dbeta: Gradient with respect to shift parameter beta, of shape (D,)
"""
dx, dgamma, dbeta = None, None, None
x_minus_mean = cache.get('x_minus_mean')
x_norm = cache.get('x_norm')
i_var = cache.get('i_var')
sqrt_var = cache.get('sqrt_var')
gamma = cache.get('gamma')
dbeta = np.sum(dout, axis=0)
dgamma = np.sum(dout * x_norm, axis=0)
N, D = dout.shape
dx_norm = dout * gamma # 1.
dxmu1 = dx_norm * i_var # 2.
di_var = np.sum(dx_norm * x_minus_mean) # 3.
dsqrtvar = di_var * (-1 / sqrt_var ** 2) # 4
dvar = dsqrtvar * 0.5 * (1 / sqrt_var) # 5
dsq = dvar * np.ones_like(dout) / N # 6
dxmu2 = dsq * 2 * x_minus_mean # 7
dx1 = dxmu1 + dxmu2 # 7
dmu = -np.sum(dxmu1 + dxmu2, axis=0) # 8
dx2 = dmu * np.ones_like(dout) / N # 9
dx = dx1 + dx2 # 10
return dx, dgamma, dbeta
def batchnorm_backward_alt(dout, cache):
"""
Alternative backward pass for batch normalization.
For this implementation you should work out the derivatives for the batch
normalizaton backward pass on paper and simplify as much as possible. You
should be able to derive a simple expression for the backward pass.
See the jupyter notebook for more hints.
Note: This implementation should expect to receive the same cache variable
as batchnorm_backward, but might not use all of the values in the cache.
Inputs / outputs: Same as batchnorm_backward
"""
dx, dgamma, dbeta = None, None, None
N, D = dout.shape
x_norm = cache.get('x_norm')
gamma = cache.get('gamma')
i_var = cache.get('i_var')
x_minus_mean = cache.get('x_minus_mean')
sqrt_var = cache.get('sqrt_var')
# Backprop dout to calculate dbeta and dgamma.
dbeta = np.sum(dout, axis=0)
dgamma = np.sum(dout * x_norm, axis=0)
# Alternative faster formula way of calculating dx.
# ref:http://cthorey.github.io./backpropagation/
dx = (1 / N) * gamma * 1 / sqrt_var * (
(N * dout) - np.sum(dout, axis=0) - (x_minus_mean) * np.square(i_var) * np.sum(dout * (x_minus_mean),
axis=0))
return dx, dgamma, dbeta