-
Notifications
You must be signed in to change notification settings - Fork 0
/
Report_tmp.html
460 lines (394 loc) · 15 KB
/
Report_tmp.html
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
44
45
46
47
48
49
50
51
52
53
54
55
56
57
58
59
60
61
62
63
64
65
66
67
68
69
70
71
72
73
74
75
76
77
78
79
80
81
82
83
84
85
86
87
88
89
90
91
92
93
94
95
96
97
98
99
100
101
102
103
104
105
106
107
108
109
110
111
112
113
114
115
116
117
118
119
120
121
122
123
124
125
126
127
128
129
130
131
132
133
134
135
136
137
138
139
140
141
142
143
144
145
146
147
148
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
177
178
179
180
181
182
183
184
185
186
187
188
189
190
191
192
193
194
195
196
197
198
199
200
201
202
203
204
205
206
207
208
209
210
211
212
213
214
215
216
217
218
219
220
221
222
223
224
225
226
227
228
229
230
231
232
233
234
235
236
237
238
239
240
241
242
243
244
245
246
247
248
249
250
251
252
253
254
255
256
257
258
259
260
261
262
263
264
265
266
267
268
269
270
271
272
273
274
275
276
277
278
279
280
281
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
441
442
443
444
445
446
447
448
449
450
451
452
453
454
455
456
457
458
459
460
<!DOCTYPE html>
<html>
<head>
<title>Report.md</title>
<meta http-equiv="Content-type" content="text/html;charset=UTF-8">
<style>
/* https://github.com/microsoft/vscode/blob/master/extensions/markdown-language-features/media/markdown.css */
/*---------------------------------------------------------------------------------------------
* Copyright (c) Microsoft Corporation. All rights reserved.
* Licensed under the MIT License. See License.txt in the project root for license information.
*--------------------------------------------------------------------------------------------*/
body {
font-family: var(--vscode-markdown-font-family, -apple-system, BlinkMacSystemFont, "Segoe WPC", "Segoe UI", "Ubuntu", "Droid Sans", sans-serif);
font-size: var(--vscode-markdown-font-size, 14px);
padding: 0 26px;
line-height: var(--vscode-markdown-line-height, 22px);
word-wrap: break-word;
}
#code-csp-warning {
position: fixed;
top: 0;
right: 0;
color: white;
margin: 16px;
text-align: center;
font-size: 12px;
font-family: sans-serif;
background-color:#444444;
cursor: pointer;
padding: 6px;
box-shadow: 1px 1px 1px rgba(0,0,0,.25);
}
#code-csp-warning:hover {
text-decoration: none;
background-color:#007acc;
box-shadow: 2px 2px 2px rgba(0,0,0,.25);
}
body.scrollBeyondLastLine {
margin-bottom: calc(100vh - 22px);
}
body.showEditorSelection .code-line {
position: relative;
}
body.showEditorSelection .code-active-line:before,
body.showEditorSelection .code-line:hover:before {
content: "";
display: block;
position: absolute;
top: 0;
left: -12px;
height: 100%;
}
body.showEditorSelection li.code-active-line:before,
body.showEditorSelection li.code-line:hover:before {
left: -30px;
}
.vscode-light.showEditorSelection .code-active-line:before {
border-left: 3px solid rgba(0, 0, 0, 0.15);
}
.vscode-light.showEditorSelection .code-line:hover:before {
border-left: 3px solid rgba(0, 0, 0, 0.40);
}
.vscode-light.showEditorSelection .code-line .code-line:hover:before {
border-left: none;
}
.vscode-dark.showEditorSelection .code-active-line:before {
border-left: 3px solid rgba(255, 255, 255, 0.4);
}
.vscode-dark.showEditorSelection .code-line:hover:before {
border-left: 3px solid rgba(255, 255, 255, 0.60);
}
.vscode-dark.showEditorSelection .code-line .code-line:hover:before {
border-left: none;
}
.vscode-high-contrast.showEditorSelection .code-active-line:before {
border-left: 3px solid rgba(255, 160, 0, 0.7);
}
.vscode-high-contrast.showEditorSelection .code-line:hover:before {
border-left: 3px solid rgba(255, 160, 0, 1);
}
.vscode-high-contrast.showEditorSelection .code-line .code-line:hover:before {
border-left: none;
}
img {
max-width: 100%;
max-height: 100%;
}
a {
text-decoration: none;
}
a:hover {
text-decoration: underline;
}
a:focus,
input:focus,
select:focus,
textarea:focus {
outline: 1px solid -webkit-focus-ring-color;
outline-offset: -1px;
}
hr {
border: 0;
height: 2px;
border-bottom: 2px solid;
}
h1 {
padding-bottom: 0.3em;
line-height: 1.2;
border-bottom-width: 1px;
border-bottom-style: solid;
}
h1, h2, h3 {
font-weight: normal;
}
table {
border-collapse: collapse;
}
table > thead > tr > th {
text-align: left;
border-bottom: 1px solid;
}
table > thead > tr > th,
table > thead > tr > td,
table > tbody > tr > th,
table > tbody > tr > td {
padding: 5px 10px;
}
table > tbody > tr + tr > td {
border-top: 1px solid;
}
blockquote {
margin: 0 7px 0 5px;
padding: 0 16px 0 10px;
border-left-width: 5px;
border-left-style: solid;
}
code {
font-family: Menlo, Monaco, Consolas, "Droid Sans Mono", "Courier New", monospace, "Droid Sans Fallback";
font-size: 1em;
line-height: 1.357em;
}
body.wordWrap pre {
white-space: pre-wrap;
}
pre:not(.hljs),
pre.hljs code > div {
padding: 16px;
border-radius: 3px;
overflow: auto;
}
pre code {
color: var(--vscode-editor-foreground);
tab-size: 4;
}
/** Theming */
.vscode-light pre {
background-color: rgba(220, 220, 220, 0.4);
}
.vscode-dark pre {
background-color: rgba(10, 10, 10, 0.4);
}
.vscode-high-contrast pre {
background-color: rgb(0, 0, 0);
}
.vscode-high-contrast h1 {
border-color: rgb(0, 0, 0);
}
.vscode-light table > thead > tr > th {
border-color: rgba(0, 0, 0, 0.69);
}
.vscode-dark table > thead > tr > th {
border-color: rgba(255, 255, 255, 0.69);
}
.vscode-light h1,
.vscode-light hr,
.vscode-light table > tbody > tr + tr > td {
border-color: rgba(0, 0, 0, 0.18);
}
.vscode-dark h1,
.vscode-dark hr,
.vscode-dark table > tbody > tr + tr > td {
border-color: rgba(255, 255, 255, 0.18);
}
</style>
<style>
/* Tomorrow Theme */
/* http://jmblog.github.com/color-themes-for-google-code-highlightjs */
/* Original theme - https://github.com/chriskempson/tomorrow-theme */
/* Tomorrow Comment */
.hljs-comment,
.hljs-quote {
color: #8e908c;
}
/* Tomorrow Red */
.hljs-variable,
.hljs-template-variable,
.hljs-tag,
.hljs-name,
.hljs-selector-id,
.hljs-selector-class,
.hljs-regexp,
.hljs-deletion {
color: #c82829;
}
/* Tomorrow Orange */
.hljs-number,
.hljs-built_in,
.hljs-builtin-name,
.hljs-literal,
.hljs-type,
.hljs-params,
.hljs-meta,
.hljs-link {
color: #f5871f;
}
/* Tomorrow Yellow */
.hljs-attribute {
color: #eab700;
}
/* Tomorrow Green */
.hljs-string,
.hljs-symbol,
.hljs-bullet,
.hljs-addition {
color: #718c00;
}
/* Tomorrow Blue */
.hljs-title,
.hljs-section {
color: #4271ae;
}
/* Tomorrow Purple */
.hljs-keyword,
.hljs-selector-tag {
color: #8959a8;
}
.hljs {
display: block;
overflow-x: auto;
color: #4d4d4c;
padding: 0.5em;
}
.hljs-emphasis {
font-style: italic;
}
.hljs-strong {
font-weight: bold;
}
</style>
<style>
/*
* Markdown PDF CSS
*/
body {
font-family: -apple-system, BlinkMacSystemFont, "Segoe WPC", "Segoe UI", "Ubuntu", "Droid Sans", sans-serif, "Meiryo";
padding: 0 12px;
}
pre {
background-color: #f8f8f8;
border: 1px solid #cccccc;
border-radius: 3px;
overflow-x: auto;
white-space: pre-wrap;
overflow-wrap: break-word;
}
pre:not(.hljs) {
padding: 23px;
line-height: 19px;
}
blockquote {
background: rgba(127, 127, 127, 0.1);
border-color: rgba(0, 122, 204, 0.5);
}
.emoji {
height: 1.4em;
}
code {
font-size: 14px;
line-height: 19px;
}
/* for inline code */
:not(pre):not(.hljs) > code {
color: #C9AE75; /* Change the old color so it seems less like an error */
font-size: inherit;
}
/* Page Break : use <div class="page"/> to insert page break
-------------------------------------------------------- */
.page {
page-break-after: always;
}
</style>
<script src="https://unpkg.com/mermaid/dist/mermaid.min.js"></script>
</head>
<body>
<script>
mermaid.initialize({
startOnLoad: true,
theme: document.body.classList.contains('vscode-dark') || document.body.classList.contains('vscode-high-contrast')
? 'dark'
: 'default'
});
</script>
<h1 id="report">Report</h1>
<blockquote>
<p>Name: Tianzuo Zhang</p>
<p>My contact info: <a href="https://twitter.com/dvzhangtz">Twitter</a> <a href="https://www.linkedin.com/in/tianzuo-zhang/">Linkedin</a> Wechat: dvzhangtz <a href="https://www.kaggle.com/milesme">Kaggle</a></p>
<p>I also upload my homework to <a href="https://github.com/dvzhang/feedback-prize-english-language-learning">Github</a></p>
</blockquote>
<h1 id="0-background">0. Background</h1>
<h3 id="01-goal">0.1. Goal:</h3>
<p>Make an article scoring system for English Language Learners.</p>
<h3 id="02-motivation">0.2. Motivation:</h3>
<p>As a Kaggle user ( <a href="https://www.kaggle.com/milesme">my account</a> ), I found a <a href="https://www.kaggle.com/competitions/feedback-prize-english-language-learning">very interesting competition</a> . I really hope I can solve this problem in my homework.</p>
<p>The goal of this competition is to assess the language proficiency of 8th-12th grade English Language Learners (ELLs). Utilizing a dataset of essays written by ELLs will help to develop proficiency models that better support all students.</p>
<p>In the dataset given by the competition, every essays have been scored according to six analytic measures: cohesion, syntax, vocabulary, phraseology, grammar, and conventions.
Each measure represents a component of proficiency in essay writing, with greater scores corresponding to greater proficiency in that measure. The scores range from 1.0 to 5.0 in increments of 0.5.</p>
<p><strong>Our task is to predict the score of each of the six measures for the essays given in the test set</strong></p>
<h1 id="1-method-description">1. Method description</h1>
<p>With this dataset, come to our method. No doubt we must use <a href="https://arxiv.org/pdf/1810.04805.pdf&usg=ALkJrhhzxlCL6yTht2BRmH9atgvKFxHsxQ">Bert or other transformer based model</a> to solve this nlp question.</p>
<p>The Transformer models are pre-trained on the general domain corpus. But for our task, its data distribution may be different from a transformer trained on a different corpus e.g. <a href="https://arxiv.org/pdf/1907.11692.pdf%5C">RoBERTa</a> trained on BookCorpus, Wiki, CC-News, OpenWebText, Stories.</p>
<p>What is more, this competition give me a very small train set, if I use it finetune my bert model directly, It must be over fit.</p>
<p>Therefore the idea is, we can further pre-train the transformer with masked language model and next sentence prediction tasks on the domain-specific data.</p>
<p><img src="file:///home/thutsjclab/thutsjclab/kaggle/new/feedback-prize-english-language-learning/pic/WechatIMG561.png" alt="picture"></p>
<p>As a result, we need some domain specific data.</p>
<p>So here come to the other dataset. The first one is the dataset I scrape from <a href="https://lang-8.com/1">Lang8</a>, it is a multilingo language learning platform. In this platform there are lots of language learner post blogs, writing by the language they are learning.</p>
<p>The second dataset is from <a href="https://www.kaggle.com/competitions/feedback-prize-2021">another Kaggle competition</a>, which is very similar from this one.</p>
<p>Using this two dataset, I continue pretrain my bert and then finetune it with the dataset given by this competition.</p>
<h1 id="2-description-of-dataset">2. Description of Dataset</h1>
<p>I have three dataset:</p>
<p>1, <a href="https://www.kaggle.com/competitions/feedback-prize-english-language-learning/data">This competition's dataset</a> which can be downloaded from Kaggle Api.</p>
<p>2, Dataset scraped from <a href="https://lang-8.com/1">Lang-8</a>, which can be used for further pretrain.</p>
<p>3, <a href="https://www.kaggle.com/competitions/feedback-prize-2021">Dataset downloaded from Kaggle Api</a>, which can be used for further pretrain.</p>
<h2 id="21-this-competitions-dataset">2.1 <a href="https://www.kaggle.com/competitions/feedback-prize-english-language-learning/data">This competition's dataset</a></h2>
<p>Every essays in the dataset have been scored according to six analytic measures: cohesion, syntax, vocabulary, phraseology, grammar, and conventions.
Each measure represents a component of proficiency in essay writing, with greater scores corresponding to greater proficiency in that measure. The scores range from 1.0 to 5.0 in increments of 0.5.</p>
<p>Our task is to predict the score of each of the six measures for the essays given in the test set.</p>
<p>In these picture, we can see the head row of our train and test set.
<img src="file:///home/thutsjclab/thutsjclab/kaggle/new/feedback-prize-english-language-learning/pic/pic1.png" alt="pic">
<img src="file:///home/thutsjclab/thutsjclab/kaggle/new/feedback-prize-english-language-learning/pic/pic2.png" alt="pic"></p>
<p>I want to mention that the The train set only contains 3,911 texts.The test set CSV only contains 3 texts,which is very, so we must be careful about the overfiting</p>
<p><img src="file:///home/thutsjclab/thutsjclab/kaggle/new/feedback-prize-english-language-learning/pic/pic3.png" alt="pic"></p>
<p>The labels appears to be normally distributed
<img src="file:///home/thutsjclab/thutsjclab/kaggle/new/feedback-prize-english-language-learning/pic/pic4.png" alt="pic"></p>
<p>And there is a high correlation between them
<img src="file:///home/thutsjclab/thutsjclab/kaggle/new/feedback-prize-english-language-learning/pic/pic5.png" alt="pic"></p>
<h2 id="22-dataset-scraped-from-lang-8">2.2 Dataset scraped from <a href="https://lang-8.com/1">Lang-8</a></h2>
<p>This dataset can be used for further pretrain.
The logic of the scrape code can be showed in the following picture:</p>
<p><img src="file:///home/thutsjclab/thutsjclab/kaggle/new/feedback-prize-english-language-learning/pic/WechatIMG553.png" alt="pic"></p>
<h2 id="23-dataset-downloaded-from-kaggle-api">2.3 <a href="https://www.kaggle.com/competitions/feedback-prize-2021">Dataset downloaded from Kaggle Api</a></h2>
<p>This dataset can be used for further pretrain.</p>
<h1 id="3-what-the-script-does">3. What the script does</h1>
<ul>
<li>
<p>scraper.py was used to scrapy data.</p>
</li>
<li>
<p>continuePretrainDataPre.py was used to preprocess the data.</p>
</li>
<li>
<p>cotinuePretrain.py was used to further pretrain the model.</p>
</li>
<li>
<p>pretrainFtFeedback2.py was used to fine-tune the model and get the result.</p>
</li>
</ul>
<h1 id="4-results-and-conclusion">4. Results and Conclusion</h1>
<p>Using the evaluation metric given by the competition:
<img src="file:///home/thutsjclab/thutsjclab/kaggle/new/feedback-prize-english-language-learning/pic/WechatIMG563.png" alt="picture">
My score is 0.477671232111189</p>
<p>The result detail can be found in submission.csv
<img src="file:///home/thutsjclab/thutsjclab/kaggle/new/feedback-prize-english-language-learning/pic/WechatIMG569.png" alt="picture">
So, my first conclusion is I made it, I solve this question.
However, the top-1 team's score is 0.433356.
So, my second conclusion is I should do something else to improve my score, which will be mentioned in the "Extensibility" part.</p>
<h1 id="5-maintainability">5. Maintainability</h1>
<p>I use User-Agent pool to increase the maintainability of my scrape program.
I did not use IP-pool, since it is expensive.
So my scrape program is a little slow.</p>
<h1 id="6-extensibility">6. Extensibility</h1>
<p>We can arm my scrape program with IP pool to increase the maintainability.
We can learn from the top score team, we can do in the future:</p>
<ol>
<li>Layer-Wise Learning Rate Dacay</li>
<li>Fast Gradient Method</li>
<li>Adversarial Weight Perturbation</li>
<li>Re-initializing upper layer (normal, xavier_uniform, xavier_normal, kaiming_uniform, . kaiming_normal, orthogonal)</li>
<li>Initializing module (normal, xavier_uniform, xavier_normal, kaiming_uniform, kaiming_normal, . orthogonal)</li>
<li>Freeze lower layer when you use very large model (v2-xlarge, funnnel, etc.)</li>
<li>Loss function, SmoothL1 or RMSE</li>
</ol>
<p>As all of us use Pytorch, which can be easily extend to this method.</p>
</body>
</html>