Skip to content

B
book

PaddlePaddle / book

通知 16
Star 4
Fork 0
B
book
切换分支/标签
查找文件 普通视图历史永久链接Permalink
index.en.html 31.6 KB
Newer Older
Y
Redo build.sh and re-generate  
Yi Wang 已提交
1 2 3 4 5 6 7 8 9 10 11 12 13 14 15 16 17 18 19 20 21 22 23 24 25 26 27 28 29 30 31 32 33 34 35 36 37 38 39 40 41 42 43 
<html>
<head>
  <script type="text/x-mathjax-config">
  MathJax.Hub.Config({
    extensions: ["tex2jax.js", "TeX/AMSsymbols.js", "TeX/AMSmath.js"],
    jax: ["input/TeX", "output/HTML-CSS"],
    tex2jax: {
      inlineMath: [ ['$','$'], ["\\(","\\)"] ],
      displayMath: [ ['$$','$$'], ["\\[","\\]"] ],
      processEscapes: true
    },
    "HTML-CSS": { availableFonts: ["TeX"] }
  });
  </script>
  <script src="https://cdnjs.cloudflare.com/ajax/libs/mathjax/2.7.0/MathJax.js" async></script>
  <script type="text/javascript" src="../.tmpl/marked.js">
  </script>
  <link href="http://cdn.bootcss.com/highlight.js/9.9.0/styles/darcula.min.css" rel="stylesheet">
  <script src="http://cdn.bootcss.com/highlight.js/9.9.0/highlight.min.js"></script>
  <link href="http://cdn.bootcss.com/bootstrap/4.0.0-alpha.6/css/bootstrap.min.css" rel="stylesheet">
  <link href="https://cdn.jsdelivr.net/perfect-scrollbar/0.6.14/css/perfect-scrollbar.min.css" rel="stylesheet">
  <link href="../.tmpl/github-markdown.css" rel='stylesheet'>
</head>
<style type="text/css" >
.markdown-body {
    box-sizing: border-box;
    min-width: 200px;
    max-width: 980px;
    margin: 0 auto;
    padding: 45px;
}
</style>


<body>

<div id="context" class="container markdown-body">
</div>

<!-- This block will be replaced by each markdown file content. Please do not change lines below.-->
<div id="markdown" style='display:none'>
# Semantic Role Labeling
L
refine mnist demo  
Luo Tao 已提交
44 
Source code of this chapter is in [book/label_semantic_roles](https://github.com/PaddlePaddle/book/tree/develop/label_semantic_roles).
Y
Redo build.sh and re-generate  
Yi Wang 已提交
45 46 47 

## Background
L
refine mnist demo  
Luo Tao 已提交
48 
Natural Language Analysis contains three components: Lexical Analysis, Syntactic Analysis, and Semantic Analysis. Semantic Role Labelling (SRL) is one way for Shallow Semantic Analysis. A predicate of a sentence is a property that a subject possesses or is characterized, such as what it does, what it is or how it is, which mostly corresponds to the core of an event. The noun associated with a predicate is called Argument. Semantic roles express the abstract roles that arguments of a predicate can take in the event, such as Agent, Patient, Theme, Experiencer, Beneficiary, Instrument, Location, Goal and Source, etc.
Y
Redo build.sh and re-generate  
Yi Wang 已提交
49
L
refine mnist demo  
Luo Tao 已提交
50 
In the following example, “遇到” (encounters) is a Predicate (“Pred”)，“小明” (Ming) is an Agent，“小红” (Hong) is a Patient，“昨天” (yesterday) indicates the Time, and “公园” (park) is the Location.
Y
Redo build.sh and re-generate  
Yi Wang 已提交
51 52 53 

$$\mbox{[小明]}_{\mbox{Agent}}\mbox{[昨天]}_{\mbox{Time}}\mbox{[晚上]}_\mbox{Time}\mbox{在[公园]}_{\mbox{Location}}\mbox{[遇到]}_{\mbox{Predicate}}\mbox{了[小红]}_{\mbox{Patient}}\mbox{。}$$
L
refine mnist demo  
Luo Tao 已提交
54 
Instead of in-depth analysis on semantic information, the goal of Semantic Role Labeling is to identify the relation of predicate and other constituents, e.g., predicate-argument structure, as specific semantic roles, which is an important intermediate step in a wide range of natural language understanding tasks (Information Extraction, Discourse Analysis, DeepQA etc). Predicates are always assumed to be given; the only thing is to identify arguments and their semantic roles.
Y
Redo build.sh and re-generate  
Yi Wang 已提交
55
L
refine mnist demo  
Luo Tao 已提交
56 
Standard SRL system mostly builds on top of Syntactic Analysis and contains five steps:
Y
Redo build.sh and re-generate  
Yi Wang 已提交
57 58 59 60 

1. Construct a syntactic parse tree, as shown in Fig. 1
2. Identity candidate arguments of given predicate from constructed syntactic parse tree.
3. Prune most unlikely candidate arguments.
L
refine mnist demo  
Luo Tao 已提交
61 
4. Identify arguments, often by a binary classifier.
Y
Redo build.sh and re-generate  
Yi Wang 已提交
62 63 64 65 66 67 68 69 70 71 72 73 74 75 76 77 78 79 
5. Multi-class semantic role labeling. Steps 2-3 usually introduce hand-designed features based on Syntactic Analysis (step 1).


<div  align="center">
<img src="image/dependency_parsing.png" width = "80%" align=center /><br>
Fig 1. Syntactic parse tree
</div>

核心关系-> HED
定中关系-> ATT
主谓关系-> SBV
状中结构-> ADV
介宾关系-> POB
右附加关系-> RAD
动宾关系-> VOB
标点-> WP
L
refine mnist demo  
Luo Tao 已提交
80 
However, complete syntactic analysis requires identifying the relation among all constitutes and the performance of SRL is sensitive to the precision of syntactic analysis, which makes SRL a very challenging task. To reduce the complexity and obtain some syntactic structure information, we often use shallow syntactic analysis. Shallow Syntactic Analysis is also called partial parsing or chunking. Unlike complete syntactic analysis which requires the construction of the complete parsing tree, Shallow Syntactic Analysis only need to identify some independent components with relatively simple structure, such as verb phrases (chunk). To avoid difficulties in constructing a syntactic tree with high accuracy, some work\[[1](#Reference)\] proposed semantic chunking based SRL methods, which convert SRL as a sequence tagging problem. Sequence tagging tasks classify syntactic chunks using BIO representation. For syntactic chunks forming a chunk of type A, the first chunk receives the B-A tag (Begin), the remaining ones receive the tag I-A (Inside), and all chunks outside receive the tag O-A.
Y
Redo build.sh and re-generate  
Yi Wang 已提交
81 82 83 84 85 86 87 88 89 90 91 92 93 

The BIO representation of above example is shown in Fig.1.

<div  align="center">
<img src="image/bio_example.png" width = "90%"  align=center /><br>
Fig 2. BIO represention
</div>

输入序列-> input sequence
语块-> chunk
标注序列-> label sequence
角色-> role
L
refine mnist demo  
Luo Tao 已提交
94 
This example illustrates the simplicity of sequence tagging because (1) shallow syntactic analysis reduces the precision requirement of syntactic analysis; (2) pruning candidate arguments is removed; 3) argument identification and tagging are finished at the same time. Such unified methods simplify the procedure, reduce the risk of accumulating errors and boost the performance further.
Y
Redo build.sh and re-generate  
Yi Wang 已提交
95
L
refine mnist demo  
Luo Tao 已提交
96 
In this tutorial, our SRL system is built as an end-to-end system via a neural network. We take only text sequences, without using any syntactic parsing results or complex hand-designed features. We give public dataset [CoNLL-2004 and CoNLL-2005 Shared Tasks](http://www.cs.upc.edu/~srlconll/) as an example to illustrate: given a sentence with predicates marked, identify the corresponding arguments and their semantic roles by sequence tagging method.
Y
Redo build.sh and re-generate  
Yi Wang 已提交
97 98 99 

## Model
L
refine mnist demo  
Luo Tao 已提交
100 
Recurrent Neural Networks are important tools for sequence modeling and have been successfully used in some natural language processing tasks. Unlike Feed-forward neural networks, RNNs can model the dependency between elements of sequences. LSTMs as variants of RNNs aim to model long-term dependency in long sequences. We have introduced this in [understand_sentiment](https://github.com/PaddlePaddle/book/tree/develop/understand_sentiment). In this chapter, we continue to use LSTMs to solve SRL problems.
Y
Redo build.sh and re-generate  
Yi Wang 已提交
101 102 103 

### Stacked Recurrent Neural Network
L
refine mnist demo  
Luo Tao 已提交
104 
Deep Neural Networks allows extracting hierarchical representations. Higher layers can form more abstract/complex representations on top of lower layers. LSTMs, when unfolded in time, is a deep feed-forward neural network, because a computational path between the input at time $k < t$ to the output at time $t$ crosses several nonlinear layers. However, the computation carried out at each time-step is only linear transformation, which makes LSTMs a shallow model. Deep LSTMs are typically constructed by stacking multiple LSTM layers on top of each other and taking the output from lower LSTM layer at time $t$ as the input of upper LSTM layer at time $t$. Deep, hierarchical neural networks can be much efficient at representing some functions and modeling varying-length dependencies\[[2](#Reference)\].
Y
Redo build.sh and re-generate  
Yi Wang 已提交
105 106
L
refine mnist demo  
Luo Tao 已提交
107 
However, deep LSTMs increases the number of nonlinear steps the gradient has to traverse when propagated back in depth. For example, four layer LSTMs can be trained properly, but the performance becomes worse as the number of layers up to 4-8. Conventional LSTMs prevent backpropagated errors from vanishing and exploding by introducing shortcut connections to skip the intermediate nonlinear layers. Therefore, deep LSTMs can consider shortcut connections in depth as well.
Y
Redo build.sh and re-generate  
Yi Wang 已提交
108 109
L
refine mnist demo  
Luo Tao 已提交
110 
The operation of a single LSTM cell contain 3 parts: (1) input-to-hidden: map input $x$ to the input of the forget gates, input gates, memory cells and output gates by linear transformation (i.e., matrix mapping); (2) hidden-to-hidden: calculate forget gates, input gates, output gates and update memory cell, this is the main part of LSTMs; (3)hidden-to-output: this part typically involves an activation operation on hidden states. Based on the stacked LSTMs, we add a shortcut connection: take the input-to-hidden from the previous layer as a new input and learn another linear transformation.
Y
Redo build.sh and re-generate  
Yi Wang 已提交
111 112 113 114 115 116 117 118 119 120 121 122 123 

Fig.3 illustrate the final stacked recurrent neural networks.

<p align="center">    
<img src="./image/stacked_lstm.png" width = "40%"  align=center><br>
Fig 3. Stacked Recurrent Neural Networks
</p>

线性变换-> linear transformation
输入层到隐层-> input-to-hidden

### Bidirectional Recurrent Neural Network
L
refine mnist demo  
Luo Tao 已提交
124 
LSTMs can summarize the history of previous inputs seen up to now, but can not see the future. In most of NLP (natural language processing) tasks, the entire sentences are ready to use. Therefore, sequential learning might be much efficient if the future can be encoded as well like histories.
Y
Redo build.sh and re-generate  
Yi Wang 已提交
125 126 127 128 129 130 131 132 133 134 135 

To address the above drawbacks, we can design bidirectional recurrent neural networks by making a minor modification. Higher LSTM layers process the sequence in reversed direction with previous lower LSTM layers, i.e., Deep LSTMs operate from left-to-right, right-to-left, left-to-right,..., in depth. Therefore, LSTM layers at time-step $t$ can see both histories and the future since the second layer. Fig. 4 illustrates the bidirectional recurrent neural networks.


<p align="center">    
<img src="./image/bidirectional_stacked_lstm.png" width = "60%" align=center><br>
Fig 4. Bidirectional LSTMs
</p>

线性变换-> linear transformation
输入层到隐层-> input-to-hidden
L
refine mnist demo  
Luo Tao 已提交
136 137 
正向处理输出序列->process sequence in the forward direction
反向处理上一层序列-> process sequence from the previous layer in backward direction
Y
Redo build.sh and re-generate  
Yi Wang 已提交
138
L
refine mnist demo  
Luo Tao 已提交
139 
Note that, this bidirectional RNNs is different with the one proposed by Bengio et al. in machine translation tasks \[[3](#Reference), [4](#Reference)\]. We will introduce another bidirectional RNNs in the following tasks[machine translation](https://github.com/PaddlePaddle/book/blob/develop/machine_translation/README.md)
Y
Redo build.sh and re-generate  
Yi Wang 已提交
140 141 142 143 144 145 146 147 

### Conditional Random Field

The basic pipeline of Neural Networks solving problems is 1) all lower layers aim to learn representations; 2) the top layer is designed for learning the final task. In SRL tasks, CRF is built on top of the network for the final tag sequence prediction. It takes the representations provided by the last LSTM layer as input.


CRF is a probabilistic graph model (undirected) with nodes denoting random variables and edges denoting dependencies between nodes. To be simplicity, CRFs learn conditional probability $P(Y|X)$, where $X = (x_1, x_2, ... , x_n)$ are sequences of input, $Y = (y_1, y_2, ... , y_n)$ are label sequences; Decoding is to search sequence $Y$ to maximize conditional probability $P(Y|X)$, i.e., $Y^* = \mbox{arg max}_{Y} P(Y | X)$。
L
refine mnist demo  
Luo Tao 已提交
148 
Sequence tagging tasks only consider input and output as linear sequences without extra dependent assumptions on graph model. Thus, the graph model of sequence tagging tasks is simple chain or line, which results in a Linear-Chain Conditional Random Field, shown in Fig.5.
Y
Redo build.sh and re-generate  
Yi Wang 已提交
149 150 151 152 153 154 155 156 157 158 159 160 161 162 163 164 165 166 167 168 169 170 171 172 173 174 175 176 

<p align="center">    
<img src="./image/linear_chain_crf.png" width = "35%" align=center><br>
Fig 5. Linear Chain Conditional Random Field used in SRL tasks
</p>

By the fundamental theorem of random fields \[[5](#Reference)\], the joint distribution over the label sequence $Y$ given $X$ has the form: 

$$p(Y | X) = \frac{1}{Z(X)} \text{exp}\left(\sum_{i=1}^{n}\left(\sum_{j}\lambda_{j}t_{j} (y_{i - 1}, y_{i}, X, i) + \sum_{k} \mu_k s_k (y_i, X, i)\right)\right)$$


where, $Z(X)$ is normalization constant, $t_j$ is feature function defined on edges, called transition feature, depending on $y_i$ and $y_{i-1}$ which represents transition probabilities from $y_{i-1}$ to $y_i$ given input sequence $X$. $s_k$ is feature function defined on nodes, called state feature, depending on $y_i$ and represents the probality of $y_i$ given input sequence $X$. $\lambda_j$ 和 $\mu_k$ are weights corresponding to $t_j$ and $s_k$. Actually, $t$ and $s$ can be wrtten in the same form, then take summation over all nodes $i$: $f_{k}(Y, X) = \sum_{i=1}^{n}f_k({y_{i - 1}, y_i, X, i})$, $f$ is defined as feature function. Thus, $P(Y|X)$ can be wrtten as:

$$p(Y|X, W) = \frac{1}{Z(X)}\text{exp}\sum_{k}\omega_{k}f_{k}(Y, X)$$

$\omega$ are weights of feature function which should be learned in CRF models. At training stage, given input sequences and label sequences $D = \left[(X_1,  Y_1), (X_2 , Y_2) , ... , (X_N, Y_N)\right]$, solve following objective function using MLE:


$$L(\lambda, D) = - \text{log}\left(\prod_{m=1}^{N}p(Y_m|X_m, W)\right) + C \frac{1}{2}\lVert W\rVert^{2}$$


This objective function can be solved via back-propagation in an end-to-end manner. At decoding stage, given input sequences $X$, search sequence $\bar{Y}$ to maximize conditional probability $\bar{P}(Y|X)$ via decoding methods (such as Viterbi, Beam Search).

### DB-LSTM SRL model

Given predicates and a sentence, SRL tasks aim to identify arguments of the given predicate and their semantic roles. If a sequence has n predicates, we will process this sequence n times. One model is as follows:

1. Construct inputs;
L
refine mnist demo  
Luo Tao 已提交
177 
 - input 1: predicate, input 2: sentence
Y
Redo build.sh and re-generate  
Yi Wang 已提交
178 
 - expand input 1 as a sequence with the same length with input 2 using one-hot representation;
L
refine mnist demo  
Luo Tao 已提交
179 180 
2. Convert one-hot sequences from step 1 to vector sequences via lookup table;
3. Learn the representation of input sequences by taking vector sequences from step 2 as inputs;
Y
Redo build.sh and re-generate  
Yi Wang 已提交
181 182 183 184 185 186 
4. Take representations from step 3 as inputs, label sequence as supervision signal, do sequence tagging tasks

We can try above method. Here, we propose some modifications by introducing two simple but effective features:

- predicate context (ctx-p): A single predicate word can not exactly describe the predicate information, especially when the same words appear more than one times in a sentence. With the expanded context, the ambiguity can be largely eliminated. Thus, we extract $n$ words before and after predicate to construct a window chunk.
L
refine mnist demo  
Luo Tao 已提交
187 
- region mark ($m_r$): $m_r = 1$ to denote word in that position locates in the predicate context region, or $m_r = 0$ if not.
Y
Redo build.sh and re-generate  
Yi Wang 已提交
188 189 190 191 

After modification, the model is as follows:

1. Construct inputs
L
refine mnist demo  
Luo Tao 已提交
192 
 - Input 1: word sequence. Input 2: predicate. Input 3: predicate context, extract $n$ words before and after predicate. Input 4: region mark sequence, element value will be 1 if word locates in the predicate context region, 0 otherwise.
Y
Redo build.sh and re-generate  
Yi Wang 已提交
193 
 - expand input 2~3 as sequences with the same length with input 1
L
refine mnist demo  
Luo Tao 已提交
194 195 
2. Convert input 1~4 to vector sequences via lookup table; input 1 and 3 shares the same lookup table, input 2 and 4 have separate lookup tables
3. Take four vector sequences from step 2 as inputs of bidirectional LSTMs; Train LSTMs to update representations
Y
Redo build.sh and re-generate  
Yi Wang 已提交
196 197 198 199 200 201 202 203 204 205 206 207 208 209 210 211 
4. Take representation from step 3 as input of CRF, label sequence as supervision signal, do sequence tagging tasks


<div  align="center">    
<img src="image/db_lstm_network.png" width = "60%"  align=center /><br>
Fig 6. DB-LSTM for SRL tasks
</div>

论元-> argu
谓词-> pred
谓词上下文-> ctx-p
谓词上下文区域标记-> $m_r$
输入-> input
原句-> sentence
反向LSTM-> LSTM Reverse
L
refine mnist demo  
Luo Tao 已提交
212 
## Data Preparation
Y
Redo build.sh and re-generate  
Yi Wang 已提交
213
L
refine mnist demo  
Luo Tao 已提交
214 
In the tutorial, we use [CoNLL 2005](http://www.cs.upc.edu/~srlconll/) SRL task open dataset as an example. It is important to note that the training set and development set of the CoNLL 2005 SRL task are not free to download after the competition. Currently, only the test set can be obtained, including 23 sections of the Wall Street Journal and three sections of the Brown corpus. In this tutorial, we use the WSJ corpus as the training dataset to explain the model. However, since the training set is small, if you want to train a usable neural network SRL system, consider paying for the full corpus.
Y
Redo build.sh and re-generate  
Yi Wang 已提交
215
L
refine mnist demo  
Luo Tao 已提交
216 
The original data includes a variety of information such as POS tagging, naming entity recognition, parsing tree, and so on. In this tutorial, we only use the data under the words folder (text sequence) and the props folder (label results) inside test.wsj parent folder. The data directory used in this tutorial is as follows:
Y
Redo build.sh and re-generate  
Yi Wang 已提交
217 218 219 220 221 222 223 224 

```text
conll05st-release/
└── test.wsj
    ├── props  # 标注结果
    └── words  # 输入文本序列
```
L
refine mnist demo  
Luo Tao 已提交
225 
The annotation information is derived from the results of Penn TreeBank\[[7](#references)\] and PropBank \[[8](# references)\]. The label of the PropBank is different from the label that we used in the example at the beginning of the article, but the principle is the same. For the description of the label, please refer to the paper \[[9](#references)\].
Y
Redo build.sh and re-generate  
Yi Wang 已提交
226
L
refine mnist demo  
Luo Tao 已提交
227 
The raw data needs to be preprocessed before used by PaddlePaddle. The preprocessing consists of the following steps:
Y
Redo build.sh and re-generate  
Yi Wang 已提交
228
L
refine mnist demo  
Luo Tao 已提交
229 230 231 232 233 
1. Merge the text sequence and the tag sequence into the same record;
2. If a sentence contains $n$ predicates, the sentence will be processed $n$ times into $n$ separate training samples, each sample with a different predicate;
3. Extract the predicate context and construct the predicate context region marker;
4. Construct the markings in BIO format;
5. Obtain the integer index corresponding to the word according to the dictionary.
Y
Redo build.sh and re-generate  
Yi Wang 已提交
234
L
refine mnist demo  
Luo Tao 已提交
235 236 237 238 239 240 
```python
# import paddle.v2.dataset.conll05 as conll05
# conll05.corpus_reader does step 1 and 2 as mentioned above.
# conll05.reader_creator does step 3 to 5.
# conll05.test gets preprocessed training instances.
```
Y
Redo build.sh and re-generate  
Yi Wang 已提交
241
L
refine mnist demo  
Luo Tao 已提交
242 
After preprocessing completes, a training sample contains nine features, namely: word sequence, predicate, predicate context (5 columns), region mark sequence, label sequence. Following table is an example of a training sample.
Y
Redo build.sh and re-generate  
Yi Wang 已提交
243
L
refine mnist demo  
Luo Tao 已提交
244 
| word sequence | predicate | predicate context（5 columns） | region mark sequence | label sequence|
Y
Redo build.sh and re-generate  
Yi Wang 已提交
245 246 247 248 249 250 251 252 253 254 
|---|---|---|---|---|
| A | set | n't been set . × | 0 | B-A1 |
| record | set | n't been set . × | 0 | I-A1 |
| date | set | n't been set . × | 0 | I-A1 |
| has | set | n't been set . × | 0 | O |
| n't | set | n't been set . × | 1 | B-AM-NEG |
| been | set | n't been set . × | 1 | O |
| set | set | n't been set . × | 1 | B-V |
| . | set | n't been set . × | 1 | O |
L
refine mnist demo  
Luo Tao 已提交
255 256 257 258 259 260 261 262 263 264 265 266 
In addition to the data, we provide following resources:

| filename | explanation |
|---|---|
| word_dict | dictionary of input sentences, total 44068 words |
| label_dict | dictionary of labels, total 106 labels |
| predicate_dict | predicate dictionary, total 3162 predicates |
| emb | a pre-trained word vector lookup table, 32-dimentional |

We trained in the English Wikipedia language model to get a word vector lookup table used to initialize the SRL model. During the SRL model training process, the word vector lookup table is no longer updated. About the language model and the word vector lookup table can refer to [word vector](https://github.com/PaddlePaddle/book/blob/develop/word2vec/README.md) tutorial. There are 995,000,000 token in training corpus, and the dictionary size is 4900,000 words. In the CoNLL 2005 training corpus, 5% of the words are not in the 4900,000 words, and we see them all as unknown words, represented by `<unk>`.

Get dictionary, print dictionary size:
Y
Redo build.sh and re-generate  
Yi Wang 已提交
267 268 

```python
L
refine mnist demo  
Luo Tao 已提交
269 270 271 272 273 274 275 276 277 278 279 
import paddle.v2 as paddle
import paddle.v2.dataset.conll05 as conll05

word_dict, verb_dict, label_dict = conll05.get_dict()
word_dict_len = len(word_dict)
label_dict_len = len(label_dict)
pred_len = len(verb_dict)

print len(word_dict_len)
print len(label_dict_len)
print len(pred_len)
Y
Redo build.sh and re-generate  
Yi Wang 已提交
280 281 
```
L
refine mnist demo  
Luo Tao 已提交
282 283 284 285 286 287 288 289 290 291 292 293 294 295 296 297 298 299 300 301 302 303 304 305 306 307 308 309 310 311 312 313 314 315 316 317 318 319 320 321 322 323 324 325 326 327 328 329 330 331 332 333 334 335 336 337 338 339 340 341 342 343 344 345 346 347 348 349 350 351 352 353 354 355 356 357 358 359 360 361 362 363 364 365 366 367 368 369 370 371 372 373 374 375 376 377 378 379 380 381 382 383 384 385 386 387 388 389 390 391 392 393 394 395 396 397 398 399 400 401 402 403 404 405 406 407 408 409 410 411 412 413 414 415 416 417 418 419 420 421 422 423 424 425 426 427 428 429 430 431 432 433 434 435 436 437 438 439 440 
## Model configuration

1. Define input data dimensions and model hyperparameters.

    ```python
    mark_dict_len = 2    # Value range of region mark. Region mark is either 0 or 1, so range is 2
    word_dim = 32        # word vector dimension
    mark_dim = 5         # adjacent dimension
    hidden_dim = 512     # the dimension of LSTM hidden layer vector is 128 (512/4)
    depth = 8            # depth of stacked LSTM
    
    # There are 9 features per sample, so we will define 9 data layers.
    # They type for each layer is integer_value_sequence.
    def d_type(value_range):
        return paddle.data_type.integer_value_sequence(value_range)

    # word sequence
    word = paddle.layer.data(name='word_data', type=d_type(word_dict_len))
    # predicate
    predicate = paddle.layer.data(name='verb_data', type=d_type(pred_len)) 

    # 5 features for predicate context
    ctx_n2 = paddle.layer.data(name='ctx_n2_data', type=d_type(word_dict_len)) 
    ctx_n1 = paddle.layer.data(name='ctx_n1_data', type=d_type(word_dict_len))
    ctx_0 = paddle.layer.data(name='ctx_0_data', type=d_type(word_dict_len))
    ctx_p1 = paddle.layer.data(name='ctx_p1_data', type=d_type(word_dict_len))
    ctx_p2 = paddle.layer.data(name='ctx_p2_data', type=d_type(word_dict_len))
    
    # region marker sequence
    mark = paddle.layer.data(name='mark_data', type=d_type(mark_dict_len))
    
    # label sequence
    target = paddle.layer.data(name='target', type=d_type(label_dict_len))
    ```
    
   Speciala note: hidden_dim = 512 means LSTM hidden vector of 128 dimension (512/4). Please refer PaddlePaddle official documentation for detail: [lstmemory](http://www.paddlepaddle.org/doc/ui/api/trainer_config_helpers/layers.html#lstmemory)。

2. The word sequence, predicate, predicate context, and region mark sequence are transformed into embedding vector sequences.

    ```python   
   
    # Since word vectorlookup table is pre-trained, we won't update it this time.
    # is_static being True prevents updating the lookup table during training.
    emb_para = paddle.attr.Param(name='emb', initial_std=0., is_static=True)
    # hyperparameter configurations
    default_std = 1 / math.sqrt(hidden_dim) / 3.0
    std_default = paddle.attr.Param(initial_std=default_std)
    std_0 = paddle.attr.Param(initial_std=0.)

    predicate_embedding = paddle.layer.embedding(
        size=word_dim,
        input=predicate,
        param_attr=paddle.attr.Param(
            name='vemb', initial_std=default_std))
    mark_embedding = paddle.layer.embedding(
        size=mark_dim, input=mark, param_attr=std_0)

    word_input = [word, ctx_n2, ctx_n1, ctx_0, ctx_p1, ctx_p2]
    emb_layers = [
        paddle.layer.embedding(
            size=word_dim, input=x, param_attr=emb_para) for x in word_input
    ]
    emb_layers.append(predicate_embedding)
    emb_layers.append(mark_embedding)
    ```

3. 8 LSTM units will be trained in "forward / backward" order.

    ```python  
    hidden_0 = paddle.layer.mixed(
        size=hidden_dim,
        bias_attr=std_default,
        input=[
            paddle.layer.full_matrix_projection(
                input=emb, param_attr=std_default) for emb in emb_layers
        ])

    mix_hidden_lr = 1e-3
    lstm_para_attr = paddle.attr.Param(initial_std=0.0, learning_rate=1.0)
    hidden_para_attr = paddle.attr.Param(
        initial_std=default_std, learning_rate=mix_hidden_lr)

    lstm_0 = paddle.layer.lstmemory(
        input=hidden_0,
        act=paddle.activation.Relu(),
        gate_act=paddle.activation.Sigmoid(),
        state_act=paddle.activation.Sigmoid(),
        bias_attr=std_0,
        param_attr=lstm_para_attr)

    # stack L-LSTM and R-LSTM with direct edges
    input_tmp = [hidden_0, lstm_0]

    for i in range(1, depth):
        mix_hidden = paddle.layer.mixed(
            size=hidden_dim,
            bias_attr=std_default,
            input=[
                paddle.layer.full_matrix_projection(
                    input=input_tmp[0], param_attr=hidden_para_attr),
                paddle.layer.full_matrix_projection(
                    input=input_tmp[1], param_attr=lstm_para_attr)
            ])

        lstm = paddle.layer.lstmemory(
            input=mix_hidden,
            act=paddle.activation.Relu(),
            gate_act=paddle.activation.Sigmoid(),
            state_act=paddle.activation.Sigmoid(),
            reverse=((i % 2) == 1),
            bias_attr=std_0,
            param_attr=lstm_para_attr)

        input_tmp = [mix_hidden, lstm]
    ```

4. We will concatenate the output of top LSTM unit with it's input, and project into a hidden layer. Then put a fully connected layer on top of it to get the final vector representation.

    ```python
    feature_out = paddle.layer.mixed(
    size=label_dict_len,
    bias_attr=std_default,
    input=[
        paddle.layer.full_matrix_projection(
            input=input_tmp[0], param_attr=hidden_para_attr),
        paddle.layer.full_matrix_projection(
            input=input_tmp[1], param_attr=lstm_para_attr)
    ], )
    ```

5.  We use CRF as cost function, the parameter of CRF cost will be named `crfw`.

    ```python
    crf_cost = paddle.layer.crf(
        size=label_dict_len,
        input=feature_out,
        label=target,
        param_attr=paddle.attr.Param(
            name='crfw',
            initial_std=default_std,
            learning_rate=mix_hidden_lr))
    ```

6.  CRF decoding layer is used for evaluation and inference. It shares parameter with CRF layer.  The sharing of parameters among multiple layers is specified by the same parameter name in these layers.

    ```python
    crf_dec = paddle.layer.crf_decoding(
       name='crf_dec_l',
       size=label_dict_len,
       input=feature_out,
       label=target,
       param_attr=paddle.attr.Param(name='crfw'))
    ```

## Train model

### Create Parameters

All necessary parameters will be traced created given output layers that we need to use.
Y
Redo build.sh and re-generate  
Yi Wang 已提交
441 442 

```python
L
refine mnist demo  
Luo Tao 已提交
443 
parameters = paddle.parameters.create([crf_cost, crf_dec])
Y
Redo build.sh and re-generate  
Yi Wang 已提交
444 445 
```
L
refine mnist demo  
Luo Tao 已提交
446 447 448 449 
We can print out parameter name. It will be generated if not specified.
   
```python
print parameters.keys()
Y
Redo build.sh and re-generate  
Yi Wang 已提交
450 451 
```
L
refine mnist demo  
Luo Tao 已提交
452 
Now we load pre-trained word lookup table.
Y
Redo build.sh and re-generate  
Yi Wang 已提交
453
L
refine mnist demo  
Luo Tao 已提交
454 455 456 457 458 459 
```python
def load_parameter(file_name, h, w):
    with open(file_name, 'rb') as f:
         f.read(16)
         return np.fromfile(f, dtype=np.float32).reshape(h, w)
parameters.set('emb', load_parameter(conll05.get_embedding(), 44068, 32))
Y
Redo build.sh and re-generate  
Yi Wang 已提交
460 461 
```
L
refine mnist demo  
Luo Tao 已提交
462 463 464 465 466 467 468 469 470 471 472 473 474 475 476 477 
### Create Trainer

We will create trainer given model topology, parameters and optimization method. We will use most basic SGD method (momentum optimizer with 0 momentum). In the meantime, we will set learning rate and regularization.

```python
optimizer = paddle.optimizer.Momentum(
    momentum=0,
    learning_rate=2e-2,
    regularization=paddle.optimizer.L2Regularization(rate=8e-4),
    model_average=paddle.optimizer.ModelAverage(
        average_window=0.5, max_average_window=10000), )

trainer = paddle.trainer.SGD(cost=crf_cost,
                             parameters=parameters,
                             update_equation=optimizer)
```
Y
Redo build.sh and re-generate  
Yi Wang 已提交
478
L
refine mnist demo  
Luo Tao 已提交
479 
### Trainer
Y
Redo build.sh and re-generate  
Yi Wang 已提交
480
L
refine mnist demo  
Luo Tao 已提交
481 
As mentioned in data preparation section, we will use CoNLL 2005 test corpus as training data set. `conll05.test()` outputs one training instance at a time. It will be shuffled, and batched into mini batches as input.
Y
Redo build.sh and re-generate  
Yi Wang 已提交
482 483 

```python
L
refine mnist formula  
Luo Tao 已提交
484 
reader = paddle.batch(
L
refine mnist demo  
Luo Tao 已提交
485 486 
    paddle.reader.shuffle(
        conll05.test(), buf_size=8192), batch_size=20)
Y
Redo build.sh and re-generate  
Yi Wang 已提交
487 488 
```
L
refine mnist formula  
Luo Tao 已提交
489 
`feeding` is used to specify relationship between data instance and layer layer. For example, according to following `feeding`, the 0th column of data instance produced by`conll05.test()` correspond to data layer named `word_data`.
L
refine mnist demo  
Luo Tao 已提交
490 491 

```python
L
refine mnist formula  
Luo Tao 已提交
492 
feeding = {
L
refine mnist demo  
Luo Tao 已提交
493 494 495 496 497 498 499 500 501 502 
    'word_data': 0,
    'ctx_n2_data': 1,
    'ctx_n1_data': 2,
    'ctx_0_data': 3,
    'ctx_p1_data': 4,
    'ctx_p2_data': 5,
    'verb_data': 6,
    'mark_data': 7,
    'target': 8
}
Y
Redo build.sh and re-generate  
Yi Wang 已提交
503 504 
```
L
refine mnist demo  
Luo Tao 已提交
505 
`event_handle` can be used as callback for training events, it will be used as an argument for `train`. Following `event_handle` prints cost during training.
Y
Redo build.sh and re-generate  
Yi Wang 已提交
506
L
refine mnist demo  
Luo Tao 已提交
507 508 509 510 511 512 513 514 515 516 517 518 519 520 521 
```python
def event_handler(event):
    if isinstance(event, paddle.event.EndIteration):
        if event.batch_id % 100 == 0:
            print "Pass %d, Batch %d, Cost %f" % (
                event.pass_id, event.batch_id, event.cost)
```

`trainer.train` will train the model.

```python
trainer.train(
    reader=reader,
    event_handler=event_handler,
    num_passes=10000,
L
refine mnist formula  
Luo Tao 已提交
522 
    feeding=feeding)
Y
Redo build.sh and re-generate  
Yi Wang 已提交
523 524 525 526 
```

## Conclusion
L
refine mnist demo  
Luo Tao 已提交
527 
Semantic Role Labeling is an important intermediate step in a wide range of natural language processing tasks. In this tutorial, we give SRL as an example to introduce how to use PaddlePaddle to do sequence tagging tasks. Proposed models are from our published paper\[[10](#Reference)\]. We only use test data as an illustration since train data on CoNLL 2005 dataset is not completely public. We hope to propose an end-to-end neural network model with fewer dependencies on natural language processing tools but is comparable, or even better than traditional models. Please check out our paper for more information and discussions.
Y
Redo build.sh and re-generate  
Yi Wang 已提交
528 529 530 531 532 533 534 535 536 537 538 539 540 541 542 543 544 545 546 547 548 549 550 551 552 553 554 555 556 557 558 559 560 561 562 563 

## Reference
1. Sun W, Sui Z, Wang M, et al. [Chinese semantic role labeling with shallow parsing](http://www.aclweb.org/anthology/D09-1#page=1513)[C]//Proceedings of the 2009 Conference on Empirical Methods in Natural Language Processing: Volume 3-Volume 3. Association for Computational Linguistics, 2009: 1475-1483.
2. Pascanu R, Gulcehre C, Cho K, et al. [How to construct deep recurrent neural networks](https://arxiv.org/abs/1312.6026)[J]. arXiv preprint arXiv:1312.6026, 2013.
3. Cho K, Van Merriënboer B, Gulcehre C, et al. [Learning phrase representations using RNN encoder-decoder for statistical machine translation](https://arxiv.org/abs/1406.1078)[J]. arXiv preprint arXiv:1406.1078, 2014.
4. Bahdanau D, Cho K, Bengio Y. [Neural machine translation by jointly learning to align and translate](https://arxiv.org/abs/1409.0473)[J]. arXiv preprint arXiv:1409.0473, 2014.
5. Lafferty J, McCallum A, Pereira F. [Conditional random fields: Probabilistic models for segmenting and labeling sequence data](http://www.jmlr.org/papers/volume15/doppa14a/source/biblio.bib.old)[C]//Proceedings of the eighteenth international conference on machine learning, ICML. 2001, 1: 282-289.
6. 李航. 统计学习方法[J]. 清华大学出版社, 北京, 2012.
7. Marcus M P, Marcinkiewicz M A, Santorini B. [Building a large annotated corpus of English: The Penn Treebank](http://repository.upenn.edu/cgi/viewcontent.cgi?article=1246&context=cis_reports)[J]. Computational linguistics, 1993, 19(2): 313-330.
8. Palmer M, Gildea D, Kingsbury P. [The proposition bank: An annotated corpus of semantic roles](http://www.mitpressjournals.org/doi/pdfplus/10.1162/0891201053630264)[J]. Computational linguistics, 2005, 31(1): 71-106.
9. Carreras X, Màrquez L. [Introduction to the CoNLL-2005 shared task: Semantic role labeling](http://www.cs.upc.edu/~srlconll/st05/papers/intro.pdf)[C]//Proceedings of the Ninth Conference on Computational Natural Language Learning. Association for Computational Linguistics, 2005: 152-164.
10. Zhou J, Xu W. [End-to-end learning of semantic role labeling using recurrent neural networks](http://www.aclweb.org/anthology/P/P15/P15-1109.pdf)[C]//Proceedings of the Annual Meeting of the Association for Computational Linguistics. 2015.

<br/>
<a rel="license" href="http://creativecommons.org/licenses/by-nc-sa/4.0/"><img alt="知识共享许可协议" style="border-width:0" src="https://i.creativecommons.org/l/by-nc-sa/4.0/88x31.png" /></a><br /><span xmlns:dct="http://purl.org/dc/terms/" href="http://purl.org/dc/dcmitype/Text" property="dct:title" rel="dct:type">本教程</span> 由 <a xmlns:cc="http://creativecommons.org/ns#" href="http://book.paddlepaddle.org" property="cc:attributionName" rel="cc:attributionURL">PaddlePaddle</a> 创作，采用 <a rel="license" href="http://creativecommons.org/licenses/by-nc-sa/4.0/">知识共享 署名-非商业性使用-相同方式共享 4.0 国际 许可协议</a>进行许可。
</div>
<!-- You can change the lines below now. -->

<script type="text/javascript">
marked.setOptions({
  renderer: new marked.Renderer(),
  gfm: true,
  breaks: false,
  smartypants: true,
  highlight: function(code, lang) {
    code = code.replace(/&amp;/g, "&")
    code = code.replace(/&gt;/g, ">")
    code = code.replace(/&lt;/g, "<")
    code = code.replace(/&nbsp;/g, " ")
    return hljs.highlightAuto(code, [lang]).value;
  }
});
document.getElementById("context").innerHTML = marked(
		document.getElementById("markdown").innerHTML)
</script>
</body>