catalogue
. 前言
. 使用的数据集
. 数据预处理
. 训练
. 测试模型运行结果: 进行实际完形填空
0. 前言
开始写这篇文章的时候是晚上12点，突然想到几点新的理解，赶紧记下来。我们用深度学习(例如tensorflow)的时候，一定要着重训练自己的建模和抽象能力，即把一个复杂的业务问题抽象为一个数学模型问题。从本质上说，阅读理解做完形填空和人机对话AI是一样的，所不同的地方在于，前者的输入一段长对话，且是带有上下文的长对话，而输出可能是一段短语，这要求神经网络需要训练出一个"长对话问题-短语回答"的最佳模型
Relevant Link:
http://blog.topspeedsnail.com/archives/11062
http://blog.topspeedsnail.com/archives/11062
http://wiki.jikexueyuan.com/project/tensorflow-zh/tutorials/mnist_pros.html
1. 使用的数据集
对于深度学习来说，训练集的覆盖度很重要，tensorflow的RGD随机递归下降会不断调整参数，直到尝试出一组最优地的参数去拟合我们的输入训练集，为了辅助tensorflow的loss函数找到最优解，我们把训练集分成2部分，一半用于RGD训练参数，另一半用于随时验证当前参数结果(数学上即loss函数获得最小值)
0x1: Children’s Book Test
Data is in the included "data" folder. Questions are separated according to whether the missing word is a named entity (NE), common noun (CN), verb (V) or preposition (P)
cbtest_NE_train.txt : questions
cbtest_NE_valid_2000ex.txt :
cbtest_NE_test_2500ex.txt : cbtest_CN_train.txt : questions
cbtest_CN_valid_2000ex.txt :
cbtest_CN_test_2500ex.txt : cbtest_V_train.txt : questions
cbtest_V_valid_2000ex.txt :
cbtest_V_test_2500ex.txt : cbtest_P_train.txt : questions
cbtest_P_valid_2000ex.txt :
cbtest_P_test_2500ex.txt :
0x2: CBT questions
Questions are built from sets of 21 consecutive sentences from the books. A sentence is defined by the Stanford Core NLP sentence splitter.
A Named Entity (NE) is any entity identified by the Stanford Core NLP NER system. A Common Noun (CN) is any word tagged as a noun by the Stanford Core NLP POS tagger that is not already a NE. Verbs and Prepositions are identified similarly.
训练语料集的组成是每20个阅读上下文，对应一个问题，然后给出一个备选答案的集合，这个集合中的任何一个都有可能成为正确的答案，它是一个开放式的问题回答
Relevant Link:
https://research.fb.com/projects/babi/
http://cs.nyu.edu/~kcho/DMQA/
2. 数据预处理
0x1: 把句子token化、把训练集转化为20条语境上下文+1条问题+一段备选答案
def preprocess_data(data_file, out_file):
# stories[x][] tories[x][] tories[x][]
stories = []
with open(data_file) as f:
story = []
for line in f:
line = line.strip()
if not line:
story = []
else:
_, line = line.split(' ', )
if line:
if '\t' in line:
q, a, _, answers = line.split('\t')
# tokenize
q = [s.strip() for s in re.split('(\W+)+', q) if s.strip()]
stories.append((story, q, a))
else:
line = [s.strip() for s in re.split('(\W+)+', line) if s.strip()]
story.append(line)
# print stories #print stories
samples = []
for story in stories:
story_tmp = []
content = []
for c in story[]:
content += c
story_tmp.append(content)
story_tmp.append(story[])
story_tmp.append(story[]) samples.append(story_tmp) #print samples # 把每一段阅读与完形填空片段顺序打乱
random.shuffle(samples)
print(len(samples)) with open(out_file, "w") as f:
for sample in samples:
f.write(str(sample))
f.write('\n')
0x2: 根据训练语料集生成词表
这里遵循的依然是word2vec词向量空间模型，我们假定训练集中的[上下文20，问题1，回答]之间都是存在关联关系的，类似马尔柯夫链中的关联预测性思想，即出现了A词汇，则A->B出现的概率是所有其他组合中最高的，对单词短语来说，词向量就相当于图像识别中的图像区域权重
# generate word vocabulary table
def read_data(data_file):
stories = []
with open(data_file) as f:
for line in f:
line = ast.literal_eval(line.strip())
stories.append(line)
return stories # generate word vocabulary table
stories = read_data(train_data_token_file) + read_data(valid_data_token_file) content_length = max([len(s) for s, _, _ in stories])
question_length = max([len(q) for _, q, _ in stories])
print(content_length, question_length) vocab = sorted(set(itertools.chain(*(story + q + [answer] for story, q, answer in stories))))
vocab_size = len(vocab) +
print(vocab_size)
word2idx = dict((w, i + ) for i, w in enumerate(vocab))
pickle.dump((word2idx, content_length, question_length, vocab_size), open(train_vocab_data_file, "wb"))
通过将词汇进行index编码，将词汇序列转化为数字序列，从而为后续计算向量最短距离作准备
0x3: 数据向量表示
将[[上下文20，问题1，回答]] list(很多段对话)纵向抽取，根据词汇表的index编号，转化为行矩阵形式，X(语境对话)、Q(问题)、A(回答)，矩阵中的每个元素都是一个index编号，代表了该字母在"向量词汇表(该词汇表中的词汇之间具备向量特征)"的索引编号
# From keras, padding
def pad_sequences(sequences, maxlen=None, dtype='int32',
padding='post', truncating='post', value=.):
lengths = [len(s) for s in sequences] nb_samples = len(sequences)
if maxlen is None:
maxlen = np.max(lengths) # take the sample shape from the first non empty sequence
# checking for consistency in the main loop below.
sample_shape = tuple()
for s in sequences:
if len(s) > :
sample_shape = np.asarray(s).shape[:]
break x = (np.ones((nb_samples, maxlen) + sample_shape) * value).astype(dtype)
for idx, s in enumerate(sequences):
if len(s) == :
continue # empty list was found
if truncating == 'pre':
trunc = s[-maxlen:]
elif truncating == 'post':
trunc = s[:maxlen]
else:
raise ValueError('Truncating type "%s" not understood' % truncating) # check `trunc` has expected shape
trunc = np.asarray(trunc, dtype=dtype)
if trunc.shape[:] != sample_shape:
raise ValueError('Shape of sample %s of sequence at position %s is different from expected shape %s' %
(trunc.shape[:], idx, sample_shape)) if padding == 'post':
x[idx, :len(trunc)] = trunc
elif padding == 'pre':
x[idx, -len(trunc):] = trunc
else:
raise ValueError('Padding type "%s" not understood' % padding)
return x # conver to vector
def to_vector(data_file, output_file):
word2idx, content_length, question_length, _ = pickle.load(open(train_vocab_data_file, "rb")) X = []
Q = []
A = []
with open(data_file) as f_i:
for line in f_i:
line = ast.literal_eval(line.strip())
x = [word2idx[w] for w in line[]]
q = [word2idx[w] for w in line[]]
a = [word2idx[line[]]] X.append(x)
Q.append(q)
A.append(a) X = pad_sequences(X, content_length)
Q = pad_sequences(Q, question_length) with open(output_file, "w") as f_o:
for i in range(len(X)):
f_o.write(str([X[i].tolist(), Q[i].tolist(), A[i]]))
f_o.write('\n')
0x4: code
# -*- coding:utf- -*- import re
import random
import ast
import itertools
import pickle
import numpy as np train_data_file = './CBTest/data/cbtest_NE_train.txt'
train_data_token_file = 'train.data' valid_data_file = './CBTest/data/cbtest_NE_valid_2000ex.txt'
valid_data_token_file = 'valid.data' train_vocab_data_file = 'vocab.data' train_vec_data_file = 'train.vec'
valid_vec_data_file = 'valid.vec' def preprocess_data(data_file, out_file):
# stories[x][] tories[x][] tories[x][]
stories = []
with open(data_file) as f:
story = []
for line in f:
line = line.strip()
if not line:
story = []
else:
_, line = line.split(' ', )
if line:
if '\t' in line:
q, a, _, answers = line.split('\t')
# tokenize
q = [s.strip() for s in re.split('(\W+)+', q) if s.strip()]
stories.append((story, q, a))
else:
line = [s.strip() for s in re.split('(\W+)+', line) if s.strip()]
story.append(line)
# print stories #print stories
samples = []
for story in stories:
story_tmp = []
content = []
for c in story[]:
content += c
story_tmp.append(content)
story_tmp.append(story[])
story_tmp.append(story[]) samples.append(story_tmp) #print samples # 把每一段阅读与完形填空片段顺序打乱
random.shuffle(samples)
print(len(samples)) with open(out_file, "w") as f:
for sample in samples:
f.write(str(sample))
f.write('\n') # generate word vocabulary table
def read_data(data_file):
stories = []
with open(data_file) as f:
for line in f:
line = ast.literal_eval(line.strip())
stories.append(line)
return stories # From keras, padding
def pad_sequences(sequences, maxlen=None, dtype='int32',
padding='post', truncating='post', value=.):
lengths = [len(s) for s in sequences] nb_samples = len(sequences)
if maxlen is None:
maxlen = np.max(lengths) # take the sample shape from the first non empty sequence
# checking for consistency in the main loop below.
sample_shape = tuple()
for s in sequences:
if len(s) > :
sample_shape = np.asarray(s).shape[:]
break x = (np.ones((nb_samples, maxlen) + sample_shape) * value).astype(dtype)
for idx, s in enumerate(sequences):
if len(s) == :
continue # empty list was found
if truncating == 'pre':
trunc = s[-maxlen:]
elif truncating == 'post':
trunc = s[:maxlen]
else:
raise ValueError('Truncating type "%s" not understood' % truncating) # check `trunc` has expected shape
trunc = np.asarray(trunc, dtype=dtype)
if trunc.shape[:] != sample_shape:
raise ValueError('Shape of sample %s of sequence at position %s is different from expected shape %s' %
(trunc.shape[:], idx, sample_shape)) if padding == 'post':
x[idx, :len(trunc)] = trunc
elif padding == 'pre':
x[idx, -len(trunc):] = trunc
else:
raise ValueError('Padding type "%s" not understood' % padding)
return x # conver to vector
def to_vector(data_file, output_file):
word2idx, content_length, question_length, _ = pickle.load(open(train_vocab_data_file, "rb")) X = []
Q = []
A = []
with open(data_file) as f_i:
for line in f_i:
line = ast.literal_eval(line.strip())
x = [word2idx[w] for w in line[]]
q = [word2idx[w] for w in line[]]
a = [word2idx[line[]]] X.append(x)
Q.append(q)
A.append(a) X = pad_sequences(X, content_length)
Q = pad_sequences(Q, question_length) with open(output_file, "w") as f_o:
for i in range(len(X)):
f_o.write(str([X[i].tolist(), Q[i].tolist(), A[i]]))
f_o.write('\n') if __name__ == "__main__":
preprocess_data(train_data_file, train_data_token_file)
preprocess_data(valid_data_file, valid_data_token_file) # generate word vocabulary table
stories = read_data(train_data_token_file) + read_data(valid_data_token_file) content_length = max([len(s) for s, _, _ in stories])
question_length = max([len(q) for _, q, _ in stories])
print(content_length, question_length) vocab = sorted(set(itertools.chain(*(story + q + [answer] for story, q, answer in stories))))
vocab_size = len(vocab) +
print(vocab_size)
word2idx = dict((w, i + ) for i, w in enumerate(vocab))
pickle.dump((word2idx, content_length, question_length, vocab_size), open(train_vocab_data_file, "wb")) to_vector(train_data_token_file, train_vec_data_file)
to_vector(valid_data_token_file, valid_vec_data_file)
Relevant Link:
3. 训练
0x1: Word Embeddings
向量空间模型 (VSMs)将词汇表达（嵌套）于一个连续的向量空间中，语义近似的词汇被映射为相邻的数据点。向量空间模型在自然语言处理领域中有着漫长且丰富的历史，不过几乎所有利用这一模型的方法都依赖于 分布式假设，其核心思想为出现于上下文情景中的词汇都有相类似的语义。采用这一假设的研究方法大致分为以下两类：基于计数的方法 (e.g. 潜在语义分析)， 和 预测方法 (e.g. 神经概率化语言模型).
0x2: 定义损失函数
loss = -tf.reduce_mean(
tf.log(tf.reduce_sum(tf.to_float(tf.equal(tf.expand_dims(A, -), X)) * X_attentions, ) + tf.constant(0.00001)))
对于训练过程来说，模型根据X上下文语境得到的answer应该和验证机中的打标结果一致的(这和图像识别只有一个正确答案的道理是一样的)
0x3: 优化器
这里使用Adam算法的Optimizer不断训练我们的输入参数
optimizer = tf.train.AdamOptimizer(learning_rate=0.001)
grads_and_vars = optimizer.compute_gradients(loss)
capped_grads_and_vars = [(tf.clip_by_norm(g, ), v) for g, v in grads_and_vars]
train_op = optimizer.apply_gradients(capped_grads_and_vars)
Tensorflow的优化器使用十分简便，已经进行了大量高层封装，我们只要实例化相应class，传入指定参数执行即可
0x4: Dropout
为了减少过拟合，我们在输出层之前加入dropout。我们用一个placeholder来代表一个神经元的输出在dropout中保持不变的概率。这样我们可以在训练过程中启用dropout，在测试过程中关闭dropout。 TensorFlow的tf.nn.dropout操作除了可以屏蔽神经元的输出外，还会自动处理神经元输出值的scale。所以用dropout的时候可以不用考虑scale。
keep_prob = tf.placeholder("float")
h_fc1_drop = tf.nn.dropout(h_fc1, keep_prob)
0x5: code
# -*- coding: utf- -*- import tensorflow as tf
import pickle
import numpy as np
import ast
from collections import defaultdict train_vec_data_file = 'train.vec'
valid_vec_data_file = 'valid.vec' train_vocab_data_file = 'vocab.data' def get_next_batch():
X = []
Q = []
A = []
for i in range(batch_size):
for line in train_file:
line = ast.literal_eval(line.strip())
X.append(line[])
Q.append(line[])
A.append(line[][])
break if len(X) == batch_size:
return X, Q, A
else:
train_file.seek()
return get_next_batch() def get_test_batch():
with open(valid_vec_data_file) as f:
X = []
Q = []
A = []
for line in f:
line = ast.literal_eval(line.strip())
X.append(line[])
Q.append(line[])
A.append(line[][])
return X, Q, A def glimpse(weights, bias, encodings, inputs):
weights = tf.nn.dropout(weights, keep_prob)
inputs = tf.nn.dropout(inputs, keep_prob)
attention = tf.transpose(tf.matmul(weights, tf.transpose(inputs)) + bias)
attention = tf.batch_matmul(encodings, tf.expand_dims(attention, -))
attention = tf.nn.softmax(tf.squeeze(attention, -))
return attention, tf.reduce_sum(tf.expand_dims(attention, -) * encodings, ) def neural_attention(embedding_dim=, encoding_dim=):
embeddings = tf.Variable(tf.random_normal([vocab_size, embedding_dim], stddev=0.22), dtype=tf.float32)
tf.contrib.layers.apply_regularization(tf.contrib.layers.l2_regularizer(1e-), [embeddings]) with tf.variable_scope('encode'):
with tf.variable_scope('X'):
X_lens = tf.reduce_sum(tf.sign(tf.abs(X)), )
embedded_X = tf.nn.embedding_lookup(embeddings, X)
encoded_X = tf.nn.dropout(embedded_X, keep_prob)
gru_cell = tf.nn.rnn_cell.GRUCell(encoding_dim)
outputs, output_states = tf.nn.bidirectional_dynamic_rnn(gru_cell, gru_cell, encoded_X,
sequence_length=X_lens, dtype=tf.float32,
swap_memory=True)
encoded_X = tf.concat(, outputs)
with tf.variable_scope('Q'):
Q_lens = tf.reduce_sum(tf.sign(tf.abs(Q)), )
embedded_Q = tf.nn.embedding_lookup(embeddings, Q)
encoded_Q = tf.nn.dropout(embedded_Q, keep_prob)
gru_cell = tf.nn.rnn_cell.GRUCell(encoding_dim)
outputs, output_states = tf.nn.bidirectional_dynamic_rnn(gru_cell, gru_cell, encoded_Q,
sequence_length=Q_lens, dtype=tf.float32,
swap_memory=True)
encoded_Q = tf.concat(, outputs) W_q = tf.Variable(tf.random_normal([ * encoding_dim, * encoding_dim], stddev=0.22), dtype=tf.float32)
b_q = tf.Variable(tf.random_normal([ * encoding_dim, ], stddev=0.22), dtype=tf.float32)
W_d = tf.Variable(tf.random_normal([ * encoding_dim, * encoding_dim], stddev=0.22), dtype=tf.float32)
b_d = tf.Variable(tf.random_normal([ * encoding_dim, ], stddev=0.22), dtype=tf.float32)
g_q = tf.Variable(tf.random_normal([ * encoding_dim, * encoding_dim], stddev=0.22), dtype=tf.float32)
g_d = tf.Variable(tf.random_normal([ * encoding_dim, * encoding_dim], stddev=0.22), dtype=tf.float32) with tf.variable_scope('attend') as scope:
infer_gru = tf.nn.rnn_cell.GRUCell( * encoding_dim)
infer_state = infer_gru.zero_state(batch_size, tf.float32)
for iter_step in range():
if iter_step > :
scope.reuse_variables() _, q_glimpse = glimpse(W_q, b_q, encoded_Q, infer_state)
d_attention, d_glimpse = glimpse(W_d, b_d, encoded_X, tf.concat_v2([infer_state, q_glimpse], )) gate_concat = tf.concat_v2([infer_state, q_glimpse, d_glimpse, q_glimpse * d_glimpse], ) r_d = tf.sigmoid(tf.matmul(gate_concat, g_d))
r_d = tf.nn.dropout(r_d, keep_prob)
r_q = tf.sigmoid(tf.matmul(gate_concat, g_q))
r_q = tf.nn.dropout(r_q, keep_prob) combined_gated_glimpse = tf.concat_v2([r_q * q_glimpse, r_d * d_glimpse], )
_, infer_state = infer_gru(combined_gated_glimpse, infer_state) return tf.to_float(tf.sign(tf.abs(X))) * d_attention def train_neural_attention():
X_attentions = neural_attention()
loss = -tf.reduce_mean(
tf.log(tf.reduce_sum(tf.to_float(tf.equal(tf.expand_dims(A, -), X)) * X_attentions, ) + tf.constant(0.00001))) optimizer = tf.train.AdamOptimizer(learning_rate=0.001)
grads_and_vars = optimizer.compute_gradients(loss)
capped_grads_and_vars = [(tf.clip_by_norm(g, ), v) for g, v in grads_and_vars]
train_op = optimizer.apply_gradients(capped_grads_and_vars) saver = tf.train.Saver()
with tf.Session() as sess:
sess.run(tf.global_variables_initializer()) # writer = tf.summary.FileWriter()
# 恢复前一次训练
ckpt = tf.train.get_checkpoint_state('.')
if ckpt != None:
print(ckpt.model_checkpoint_path)
saver.restore(sess, ckpt.model_checkpoint_path)
else:
print("checkpoint not found!!") for step in range():
train_x, train_q, train_a = get_next_batch()
loss_, _ = sess.run([loss, train_op], feed_dict={X: train_x, Q: train_q, A: train_a, keep_prob: 0.7})
print(loss_) # 保存模型并计算准确率
if step % == :
path = saver.save(sess, 'machine_reading.model', global_step=step)
print(path) test_x, test_q, test_a = get_test_batch()
test_x, test_q, test_a = np.array(test_x[:batch_size]), np.array(test_q[:batch_size]), np.array(
test_a[:batch_size])
attentions = sess.run(X_attentions, feed_dict={X: test_x, Q: test_q, keep_prob: .})
correct_count =
for x in range(test_x.shape[]):
probs = defaultdict(int)
for idx, word in enumerate(test_x[x, :]):
probs[word] += attentions[x, idx]
guess = max(probs, key=probs.get)
if guess == test_a[x]:
correct_count +=
print(correct_count / test_x.shape[]) # 读取词汇表
word2idx, content_length, question_length, vocab_size = pickle.load(open(train_vocab_data_file, "rb"))
print(content_length, question_length, vocab_size)
#print word2idx batch_size = train_file = open(train_vec_data_file) X = tf.placeholder(tf.int32, [batch_size, content_length]) # 洋文材料
Q = tf.placeholder(tf.int32, [batch_size, question_length]) # 问题
A = tf.placeholder(tf.int32, [batch_size]) # 答案 # drop out
keep_prob = tf.placeholder(tf.float32) train_neural_attention()
Relevant Link:
http://docs.pythontab.com/tensorflow/tutorials/word2vec/
http://blog.csdn.net/lenbow/article/details/52218551
http://wiki.jikexueyuan.com/project/tensorflow-zh/get_started/basic_usage.html
http://wiki.jikexueyuan.com/project/tensorflow-zh/tutorials/mnist_tf.html
http://wiki.jikexueyuan.com/project/tensorflow-zh/how_tos/variable_scope.html
http://www.jianshu.com/p/45dbfe5809d4
https://www.zhihu.com/question/51325408
http://www.jianshu.com/p/c9f66bc8f96c
4. 测试模型运行结果: 进行实际完形填空
验证模型的过程就是让cnn根据和train相同的输入格式，得到一个预测概率最大的输出结果
0x1: 测试题目
We did manage to get the taffy made but before we could sample the result satisfactorily , and just as the girls were finishing with the washing of the dishes , Felicity glanced out of the window and exclaimed in tones of dismay , `` Oh , dear me , here ' s Great - aunt Eliza coming up the lane ! Now , is n ' t that too mean ? '' We all looked out to see a tall , gray - haired lady approaching the house , looking about her with the slightly puzzled air of a stranger . We had been expecting Great - aunt Eliza ' s advent for some weeks , for she was visiting relatives in Markdale . We knew she was liable to pounce down on us any time , being one of those delightful folk who like to `` surprise '' people , but we had never thought of her coming that particular day . It must be confessed that we did not look forward to her visit with any pleasure . None of us had ever seen her , but we knew she was very deaf , and had very decided opinions as to the way in which children should behave . `` Whew ! '' whistled Dan . `` We ' re in for a jolly afternoon . She ' s deaf as a post and we ' ll have to split our throats to make her hear at all . I ' ve a notion to skin out . '' `` Oh , do n ' t talk like that , Dan , '' said Cecily reproachfully . `` She ' s old and lonely and has had a great deal of trouble . She has buried three husbands . We must be kind to her and do the best we can to make her visit pleasant . '' `` She ' s coming to the back door , '' said Felicity , with an agitated glance around the kitchen . `` I told you , Dan , that you should have shovelled the snow away from the front door this morning . Cecily , set those pots in the pantry quick -- hide those boots , Felix -- shut the cupboard door , Peter -- Sara , straighten up the lounge . She ' s awfully particular and ma says her house is always as neat as wax . ''
模型会按照同样的训练过程将测试数据输入模型，得到最大概率对应的index号，即模型预测的完形填空答案
Relevant Link:
Copyright (c) 2017 LittleHann All rights reserved