990e955
11921054 5 years ago
3 changed file(s) with 191 addition(s) and 527 deletion(s). Raw diff Collapse all Expand all
+0
-188
_README.ipynb less more
0 {
1 "cells": [
2 {
3 "cell_type": "markdown",
4 "metadata": {},
5 "source": [
6 "## 1. 项目介绍\n",
7 "\n",
8 " - 项目是由模块组成、有特定功能的程序。它能够满足用户的直接使用需求，例如[古诗词生成器](https://momodel.cn/explore/5bfb634e1afd943c623dd9cf?type=app&tab=1)、[风格迁移](https://momodel.cn/explore/5bfb634e1afd943c623dd9cf?type=app&tab=1)等。\n",
9 " - 开发项目过程中你可以导入数据集，也可以通过每个 cell 上方工具栏的`<+>`直接插入[模块](https://momodel.cn/modules)和代码块。\n",
10 " - 你可以将开发好的项目进行[部署](https://momodel.cn/docs/#/zh-cn/%E5%BC%80%E5%8F%91%E5%92%8C%E9%83%A8%E7%BD%B2%E4%B8%80%E4%B8%AA%E5%BA%94%E7%94%A8%EF%BC%88app%EF%BC%89)，项目部署成功并选择正式版本发布后会展示在“项目”页面，用户可以在线使用，也可以通过 API 调用。\n",
11 "\n",
12 " - 项目目录结构:\n",
13 "\n",
14 " - ```results```*-----结果的文件存放地(如果你运行 job，务必将运行结果指定在此目录)*\n",
15 " - ```_OVERVIEW.md``` *-----项目的相关介绍*\n",
16 " - ```_README.md```*-----说明文档*\n",
17 " - ```app_spec.yml```*-----定义项目的输入输出，为部署服务*\n",
18 " - ```coding_here.ipynb```*-----输入并运行代码*"
19 ]
20 },
21 {
22 "cell_type": "code",
23 "execution_count": null,
24 "metadata": {},
25 "outputs": [],
26 "source": [
27 "def handle(conf):\n",
28 " \"\"\"\n",
29 " 该方法是部署之后，其他人调用你的服务时候的处理方法。\n",
30 " 请按规范填写参数结构，这样我们就能替你自动生成配置文件，方便其他人的调用。\n",
31 " 范例：\n",
32 " params['key'] = value # value_type: str # description: some description\n",
33 " value_type 可以选择：img, video, audio, str, int, float, [int], [str], [float]\n",
34 " 参数请放到params字典中，我们会自动解析该变量。\n",
35 " \"\"\"\n",
36 "\n",
37 " param1 = conf['param1'] # value_type: str # description: some description\n",
38 " # add your code\n",
39 " return {'ret1': 'cat'}\n",
40 " "
41 ]
42 },
43 {
44 "cell_type": "markdown",
45 "metadata": {},
46 "source": [
47 "\n",
48 "## 2. 开发环境简介\n",
49 "\n",
50 "你当前所在的页面 Notebook 是一个内嵌 JupyterLab 的在线类 IDE 编程环境，开发过程中可以使用页面右侧的 API 文档进行快速查询。Notebook 有以下主要功能：\n",
51 "\n",
52 "- [调用数据集、模块和代码块资源](https://momodel.cn/docs/#/zh-cn/%E5%A6%82%E4%BD%95%E5%AF%BC%E5%85%A5%E5%B9%B6%E4%BD%BF%E7%94%A8%E6%A8%A1%E5%9D%97%E5%92%8C%E6%95%B0%E6%8D%AE%E9%9B%86)\n",
53 "- [多人代码协作](https://momodel.cn/docs/#/zh-cn/%E5%9C%A8Mo%E8%BF%90%E8%A1%8C%E4%BD%A0%E7%9A%84%E7%AC%AC%E4%B8%80%E6%AE%B5%E4%BB%A3%E7%A0%81?id=_7-%e4%bd%a0%e5%8f%af%e4%bb%a5%e9%82%80%e8%af%b7%e5%a5%bd%e5%8f%8b%e8%bf%9b%e8%a1%8c%e5%8d%8f%e4%bd%9c)\n",
54 "- [在 GPU 资源上训练机器学习模型](https://momodel.cn/docs/#/zh-cn/%E5%9C%A8GPU%E6%88%96CPU%E8%B5%84%E6%BA%90%E4%B8%8A%E8%AE%AD%E7%BB%83%E6%9C%BA%E5%99%A8%E5%AD%A6%E4%B9%A0%E6%A8%A1%E5%9E%8B)\n",
55 "- [简单部署](https://momodel.cn/docs/#/zh-cn/%E5%BC%80%E5%8F%91%E5%92%8C%E9%83%A8%E7%BD%B2%E4%B8%80%E4%B8%AA%E5%BA%94%E7%94%A8%EF%BC%88app%EF%BC%89)\n",
56 "\n",
57 "快来动手试试吧！点击左侧工具栏的新建文件图标即可选择你需要的文件类型。\n",
58 "\n",
59 "<img src='https://imgbed.momodel.cn/006tNc79gy1g61agfcv23j31c30u0789.jpg' width=100% height=100%>\n",
60 "\n",
61 "\n",
62 "\n",
63 "左侧和右侧工具栏都可根据使用需要进行收合。\n",
64 "<img src='https://imgbed.momodel.cn/collapse_tab.2019-09-06 11_07_44.gif' width=100% height=100%>"
65 ]
66 },
67 {
68 "cell_type": "code",
69 "execution_count": null,
70 "metadata": {},
71 "outputs": [],
72 "source": [
73 "ls /home/jovyan/work/datasets/sumansid-facemask-dataset-momodel"
74 ]
75 },
76 {
77 "cell_type": "code",
78 "execution_count": null,
79 "metadata": {},
80 "outputs": [],
81 "source": []
82 },
83 {
84 "cell_type": "markdown",
85 "metadata": {},
86 "source": [
87 "## 3. 快捷键与代码补全\n",
88 "Mo Notebook 已完全采用 Jupyter Notebook 的原生快捷键，并且支持 `tab` 代码补全。\n",
89 "\n",
90 "运行代码：`shift` + `enter` 或者 `shift` + `return`"
91 ]
92 },
93 {
94 "cell_type": "markdown",
95 "metadata": {},
96 "source": [
97 "## 4. 常用指令介绍\n",
98 "\n",
99 "- 解压上传后的文件\n",
100 "\n",
101 "在 cell 中输入并运行以下命令：\n",
102 "```!7zx file_name.zip```\n",
103 "\n",
104 "- 查看所有包（package）\n",
105 "\n",
106 "`!pip list --format=columns`\n",
107 "\n",
108 "- 检查是否已有某个包\n",
109 "\n",
110 "`!pip show package_name`\n",
111 "\n",
112 "- 安装缺失的包\n",
113 "\n",
114 "`!pip install package_name`\n",
115 "\n",
116 "- 更新已有的包\n",
117 "\n",
118 "`!pip install package_name --upgrade`\n",
119 "\n",
120 "\n",
121 "- 使用包\n",
122 "\n",
123 "`import package_name`\n",
124 "\n",
125 "- 显示当前目录下的档案及目录\n",
126 "\n",
127 "`ls`\n",
128 "\n",
129 "- 使用引入的数据集\n",
130 "\n",
131 "数据集被引入后存放在 datasets 文件夹下，注意，这个文件夹是只读的，不可修改。如果需要修改，可在 Notebook 中使用\n",
132 "\n",
133 "`!cp -R ./datasets/<imported_dataset_dir> ./<your_folder>`\n",
134 "\n",
135 "指令将其复制到其他文件夹后再编辑，对于引入的数据集中的 zip 文件，可使用\n",
136 "\n",
137 "`!7zx ./datasets/<imported_dataset_dir>/<XXX.zip> ./<your_folder>`\n",
138 "\n",
139 "指令解压缩到其他文件夹后使用"
140 ]
141 },
142 {
143 "cell_type": "markdown",
144 "metadata": {},
145 "source": [
146 "## 5. 其他可参考资源\n",
147 "- [帮助文档](https://momodel.cn/docs/#/)：基本页面介绍和常见问题都可以在里面找到\n",
148 "- [平台功能教程](https://momodel.cn/classroom/class?id=5c5696cd1afd9458d456bf54&type=doc)：通过图文结合的 Notebook 详细介绍开发环境基本功能和操作\n",
149 "- [吴恩达机器学习](https://momodel.cn/classroom/class?id=5c5696191afd94720cc94533&type=video)：机器学习经典课程\n",
150 "- [李宏毅机器学习](https://s.momodel.cn/classroom/class?id=5d40fdafb5113408a8dbb4a1&type=video)：中文世界最好的机器学习课程\n",
151 "- [机器学习实战](https://momodel.cn/classroom/class?id=5c680b311afd943a9f70901b&type=practice)：通过实操指引完成独立的模型，掌握相应的机器学习知识\n",
152 "- [Python 教程](https://momodel.cn/classroom/class?id=5d1f3ab81afd940ab7d298bf&type=notebook)：简单易懂的 Python 新手教程\n",
153 "- [模块开发](https://momodel.cn/modules)：关于模型训练、开发与部署的高阶教程"
154 ]
155 }
156 ],
157 "metadata": {
158 "kernelspec": {
159 "display_name": "Python 3",
160 "language": "python",
161 "name": "python3"
162 },
163 "language_info": {
164 "codemirror_mode": {
165 "name": "ipython",
166 "version": 3
167 },
168 "file_extension": ".py",
169 "mimetype": "text/x-python",
170 "name": "python",
171 "nbconvert_exporter": "python",
172 "pygments_lexer": "ipython3",
173 "version": "3.5.2"
174 },
175 "pycharm": {
176 "stem_cell": {
177 "cell_type": "raw",
178 "metadata": {
179 "collapsed": false
180 },
181 "source": []
182 }
183 }
184 },
185 "nbformat": 4,
186 "nbformat_minor": 2
187 }
+190
-339
coding_here.ipynb less more
00 {
11 "cells": [
22 {
3 "cell_type": "code",
4 "execution_count": 1,
5 "metadata": {},
6 "outputs": [
7 {
8 "name": "stdout",
9 "output_type": "stream",
10 "text": [
11 "Hello Mo!\n"
12 ]
13 }
14 ],
15 "source": [
16 "print('Hello Mo!')"
17 ]
18 },
19 {
20 "cell_type": "code",
21 "execution_count": 5,
22 "metadata": {},
23 "outputs": [
24 {
25 "name": "stdout",
26 "output_type": "stream",
27 "text": [
28 "3\n"
29 ]
30 }
31 ],
32 "source": [
33 "b=2\n",
34 "a = b +1\n",
35 "print(a)"
36 ]
37 },
38 {
39 "cell_type": "code",
40 "execution_count": 6,
41 "metadata": {},
42 "outputs": [],
43 "source": [
44 "import torch"
45 ]
46 },
47 {
48 "cell_type": "code",
49 "execution_count": 9,
50 "metadata": {},
51 "outputs": [
52 {
53 "name": "stdout",
54 "output_type": "stream",
55 "text": [
56 "tensor([0., 0., 0.])\n"
57 ]
58 }
59 ],
60 "source": [
61 "a = torch.zeros(3)\n",
62 "print(a)"
63 ]
64 },
65 {
66 "cell_type": "code",
67 "execution_count": 1,
68 "metadata": {},
69 "outputs": [
70 {
71 "ename": "ImportError",
72 "evalue": "No module named 'mindspore'",
73 "output_type": "error",
74 "traceback": [
75 "\u001b[0;31m---------------------------------------------------------------------------\u001b[0m",
76 "\u001b[0;31mImportError\u001b[0m Traceback (most recent call last)",
77 "\u001b[0;32m<ipython-input-1-39d0ce7864e1>\u001b[0m in \u001b[0;36m<module>\u001b[0;34m\u001b[0m\n\u001b[0;32m----> 1\u001b[0;31m \u001b[0;32mimport\u001b[0m \u001b[0mmindspore\u001b[0m\u001b[0;34m.\u001b[0m\u001b[0mnn\u001b[0m \u001b[0;32mas\u001b[0m \u001b[0mnn\u001b[0m\u001b[0;34m\u001b[0m\u001b[0m\n\u001b[0m",
78 "\u001b[0;31mImportError\u001b[0m: No module named 'mindspore'"
79 ]
80 }
81 ],
82 "source": [
83 "import mindspore.nn as nn"
84 ]
85 },
86 {
87 "cell_type": "code",
88 "execution_count": 1,
89 "metadata": {},
90 "outputs": [
91 {
92 "name": "stdout",
93 "output_type": "stream",
94 "text": [
95 "Looking in indexes: https://mirrors.aliyun.com/pypi/simple\n",
96 "Collecting mindspore\n",
97 "\u001b[31m ERROR: Could not find a version that satisfies the requirement mindspore (from versions: none)\u001b[0m\n",
98 "\u001b[31mERROR: No matching distribution found for mindspore\u001b[0m\n",
99 "\u001b[33mWARNING: You are using pip version 19.1.1, however version 20.2.3 is available.\n",
100 "You should consider upgrading via the 'pip install --upgrade pip' command.\u001b[0m\n",
101 "Note: you may need to restart the kernel to use updated packages.\n"
102 ]
103 }
104 ],
105 "source": [
106 "pip install mindspore"
107 ]
108 },
109 {
110 "cell_type": "code",
111 "execution_count": 2,
112 "metadata": {},
113 "outputs": [
114 {
115 "name": "stdout",
116 "output_type": "stream",
117 "text": [
118 "Looking in indexes: https://mirrors.aliyun.com/pypi/simple\n",
119 "Collecting mindspore\n",
120 "\u001b[31m ERROR: Could not find a version that satisfies the requirement mindspore (from versions: none)\u001b[0m\n",
121 "\u001b[31mERROR: No matching distribution found for mindspore\u001b[0m\n",
122 "\u001b[33mWARNING: You are using pip version 19.1.1, however version 20.2.3 is available.\n",
123 "You should consider upgrading via the 'pip install --upgrade pip' command.\u001b[0m\n",
124 "Note: you may need to restart the kernel to use updated packages.\n"
125 ]
126 }
127 ],
128 "source": [
129 "pip install mindspore"
130 ]
131 },
132 {
133 "cell_type": "code",
134 "execution_count": 2,
135 "metadata": {},
136 "outputs": [
137 {
138 "name": "stdout",
139 "output_type": "stream",
140 "text": [
141 "Package Version \n",
142 "-------------------- ------------\n",
143 "absl-py 0.7.1 \n",
144 "alembic 1.4.2 \n",
145 "appdirs 1.4.4 \n",
146 "astor 0.7.1 \n",
147 "async-generator 1.10 \n",
148 "attrs 19.3.0 \n",
149 "Augmentor 0.2.8 \n",
150 "backcall 0.2.0 \n",
151 "baytune 0.3.10 \n",
152 "bleach 3.1.5 \n",
153 "boto3 1.14.34 \n",
154 "botocore 1.17.34 \n",
155 "certifi 2020.6.20 \n",
156 "certipy 0.1.3 \n",
157 "cffi 1.14.1 \n",
158 "chardet 3.0.4 \n",
159 "click 7.1.2 \n",
160 "configparser 4.0.2 \n",
161 "cryptography 3.0 \n",
162 "cycler 0.10.0 \n",
163 "Cython 0.29.20 \n",
164 "decorator 4.4.2 \n",
165 "defusedxml 0.6.0 \n",
166 "distlib 0.3.1 \n",
167 "docutils 0.15.2 \n",
168 "entrypoints 0.3 \n",
169 "et-xmlfile 1.0.1 \n",
170 "filelock 3.0.12 \n",
171 "func-timeout 4.3.5 \n",
172 "future 0.18.2 \n",
173 "gast 0.2.2 \n",
174 "graphviz 0.14 \n",
175 "grpcio 1.20.1 \n",
176 "h5py 2.9.0 \n",
177 "idna 2.10 \n",
178 "imageio 2.8.0 \n",
179 "imgaug 0.4.0 \n",
180 "importlib-metadata 0.23 \n",
181 "importlib-resources 3.0.0 \n",
182 "ipdb 0.13.2 \n",
183 "ipykernel 5.3.4 \n",
184 "ipython 7.9.0 \n",
185 "ipython-genutils 0.2.0 \n",
186 "ipywidgets 7.4.0 \n",
187 "jdcal 1.4.1 \n",
188 "jedi 0.17.2 \n",
189 "jieba 0.42.1 \n",
190 "Jinja2 2.11.2 \n",
191 "jmespath 0.10.0 \n",
192 "joblib 0.14.1 \n",
193 "jsonschema 3.2.0 \n",
194 "jupyter-client 6.1.6 \n",
195 "jupyter-core 4.6.3 \n",
196 "jupyterhub 1.0.0 \n",
197 "jupyterlab 1.0.0a1 \n",
198 "jupyterlab-server 0.2.0 \n",
199 "Keras 2.3.1 \n",
200 "Keras-Applications 1.0.7 \n",
201 "Keras-Preprocessing 1.0.9 \n",
202 "kiwisolver 1.1.0 \n",
203 "Mako 1.1.3 \n",
204 "Markdown 3.1 \n",
205 "MarkupSafe 1.1.1 \n",
206 "matplotlib 3.0.3 \n",
207 "minepy 1.2.4 \n",
208 "minio 5.0.10 \n",
209 "mistune 0.8.4 \n",
210 "mock 2.0.0 \n",
211 "more-itertools 7.2.0 \n",
212 "nbconvert 5.6.1 \n",
213 "nbformat 5.0.7 \n",
214 "networkx 2.4 \n",
215 "nltk 3.5 \n",
216 "notebook 6.0.3 \n",
217 "numpy 1.16.0 \n",
218 "oauthlib 3.1.0 \n",
219 "opencv-python 4.2.0.34 \n",
220 "openpyxl 2.6.4 \n",
221 "packaging 20.4 \n",
222 "pamela 1.0.0 \n",
223 "pandas 0.24.2 \n",
224 "pandocfilters 1.4.2 \n",
225 "parso 0.7.1 \n",
226 "pbr 5.2.0 \n",
227 "pexpect 4.8.0 \n",
228 "pickleshare 0.7.5 \n",
229 "Pillow 7.2.0 \n",
230 "pip 19.1.1 \n",
231 "plotly 4.8.1 \n",
232 "prometheus-client 0.8.0 \n",
233 "prompt-toolkit 2.0.10 \n",
234 "protobuf 3.12.4 \n",
235 "ptyprocess 0.6.0 \n",
236 "pycparser 2.20 \n",
237 "pydot 1.4.1 \n",
238 "pyenchant 3.1.1 \n",
239 "Pygments 2.6.1 \n",
240 "pyOpenSSL 19.1.0 \n",
241 "pyparsing 2.4.7 \n",
242 "pyrsistent 0.16.0 \n",
243 "python-dateutil 2.8.0 \n",
244 "python-editor 1.0.4 \n",
245 "pytz 2020.1 \n",
246 "PyWavelets 1.1.1 \n",
247 "PyYAML 5.3.1 \n",
248 "pyzmq 19.0.2 \n",
249 "regex 2020.7.14 \n",
250 "requests 2.24.0 \n",
251 "retrying 1.3.3 \n",
252 "s3transfer 0.3.3 \n",
253 "sacremoses 0.0.43 \n",
254 "scikit-image 0.15.0 \n",
255 "scikit-learn 0.22.2.post1\n",
256 "scipy 1.4.1 \n",
257 "seaborn 0.9.1 \n",
258 "Send2Trash 1.5.0 \n",
259 "sentencepiece 0.1.91 \n",
260 "setuptools 49.2.0 \n",
261 "Shapely 1.7.0 \n",
262 "six 1.12.0 \n",
263 "SQLAlchemy 1.3.18 \n",
264 "stevedore 1.32.0 \n",
265 "tensorboard 1.13.1 \n",
266 "tensorboardX 2.0 \n",
267 "tensorflow 1.13.1 \n",
268 "tensorflow-estimator 1.13.0 \n",
269 "termcolor 1.1.0 \n",
270 "terminado 0.8.3 \n",
271 "testpath 0.4.4 \n",
272 "tokenizers 0.5.2 \n",
273 "torch 1.4.0+cpu \n",
274 "torchtext 0.6.0 \n",
275 "torchvision 0.5.0+cpu \n",
276 "tornado 6.0.4 \n",
277 "tqdm 4.46.1 \n",
278 "traitlets 4.3.3 \n",
279 "transformers 2.5.1 \n",
280 "urllib3 1.25.10 \n",
281 "virtualenv 20.0.29 \n",
282 "virtualenv-clone 0.5.4 \n",
283 "virtualenvwrapper 4.8.4 \n",
284 "wcwidth 0.2.5 \n",
285 "webencodings 0.5.1 \n",
286 "Werkzeug 0.15.2 \n",
287 "wheel 0.34.2 \n",
288 "widgetsnbextension 3.4.2 \n",
289 "word2vec 0.10.6 \n",
290 "xlrd 1.2.0 \n",
291 "yellowbrick 1.1 \n",
292 "zipp 0.6.0 \n",
293 "\u001b[33mWARNING: You are using pip version 19.1.1, however version 20.2.3 is available.\n",
294 "You should consider upgrading via the 'pip install --upgrade pip' command.\u001b[0m\n"
295 ]
296 }
297 ],
298 "source": [
299 "!pip list --format=columns"
300 ]
301 },
302 {
303 "cell_type": "code",
304 "execution_count": 3,
305 "metadata": {},
306 "outputs": [
307 {
308 "name": "stdout",
309 "output_type": "stream",
310 "text": [
311 "/bin/sh: 1: 7zx: not found\n"
312 ]
313 }
314 ],
315 "source": [
316 "!7zx ./datasets/sumansid-facemask-dataset-momodel/facemask-dataset.zip -d"
317 ]
318 },
319 {
320 "cell_type": "code",
321 "execution_count": 4,
322 "metadata": {},
323 "outputs": [
324 {
325 "name": "stdout",
326 "output_type": "stream",
327 "text": [
328 "/bin/sh: 1: 7zx: not found\n"
329 ]
330 }
331 ],
332 "source": [
333 "!7zx ./datasets/sumansid-facemask-dataset-momodel/facemask-dataset.zip"
334 ]
335 },
336 {
337 "cell_type": "code",
338 "execution_count": null,
339 "metadata": {},
340 "outputs": [],
341 "source": []
3 "cell_type": "markdown",
4 "metadata": {
5 "inputHidden": false
6 },
7 "source": [
8 "这一节我们主要学习 - 0:06:09\n",
9 "+ 人工智能历史\n",
10 "+ 人工智能课程体系\n",
11 "+ 课程内容概要\n",
12 "+ 实训题目安排"
13 ]
14 },
15 {
16 "cell_type": "markdown",
17 "metadata": {},
18 "source": [
19 "### 1.人工智能历史 - 0:06:15\n",
20 "+ 古代哲人对智能的思考 \n",
21 "荀子在短文《荀子证明》中对智慧活动进行了如下解释 - 00:07:31 \n",
22 "<img src=\"https://imgbed.momodel.cn/1_页面_04.jpg\" width='500px'>"
23 ]
24 },
25 {
26 "cell_type": "markdown",
27 "metadata": {
28 "inputHidden": false
29 },
30 "source": [
31 "+ 计算的诞生 - 00:10:22 \n",
32 " + 原始人类结绳计数、算盘计数\n",
33 " + 可计算思想为计算机的出现提供了理论准备\n",
34 " + 大卫·希尔伯特在国际数学家大会上发表演讲，提出23个数学问题，其中第2个问题直接导致了图灵机模型的产生，如下图所示。- 00:13:50 \n",
35 " \n",
36 " <img src=\"https://imgbed.momodel.cn/1_页面_07.jpg\" width='500px'>\n",
37 " \n",
38 " + 图灵在《论数字计算在决断难题中的应用》中提出图灵机模型计算流程"
39 ]
40 },
41 {
42 "cell_type": "markdown",
43 "metadata": {
44 "inputHidden": false
45 },
46 "source": [
47 "+ 图灵测试 - 00:23:13 \n",
48 " 图灵测试是指让机器人和人类回答相同的20个问题，然后将答案交于法官判断是否能轻易识别机器人的作答，若不能，则通过图灵测试。但是图灵测试有时代的局限性，很多人类高级智能不能被体现。\n",
49 "<img src=\"https://imgbed.momodel.cn/1_页面_12.jpg\" width='500px'>"
50 ]
51 },
52 {
53 "cell_type": "markdown",
54 "metadata": {},
55 "source": [
56 "+ 摩尔定律 - 00:25:50 \n",
57 "冯·诺依曼等人提出摩尔定律，该定律支撑自动计算时代快速发展。该定律指出人类计算机速度每隔1年半要增长1倍。"
58 ]
59 },
60 {
61 "cell_type": "markdown",
62 "metadata": {},
63 "source": [
64 "+ 人工智能的诞生 - 00:26:20 \n",
65 "4位学者向达特茅斯学院提交了一个项目申请书，首次提出AI的概念。\n",
66 "<img src=\"https://imgbed.momodel.cn/1_页面_14.jpg\" width='500px'> \n",
67 "在项目申请书中提出了七个值得关注的问题，其中后面三个问题至今也是弱人工智能通向强人工智能的障碍。- 00:28:12\n",
68 "<img src=\"https://imgbed.momodel.cn/1_页面_15.jpg\" width='500px'> "
69 ]
70 },
71 {
72 "cell_type": "markdown",
73 "metadata": {},
74 "source": [
75 "+ 人工智能的分类 - 00:30:11 \n",
76 "从智能角度对人工智能的分类： \n",
77 " + 领域人工智能体：依葫芦画瓢、任务导向，如Deep Blue和AlphaGo\n",
78 " + 通用人工智能或跨领域人工智能体：举一反三、从经验中学习，如“人类”智能\n",
79 " + 混合增强人工智能：多种智能体的混合形式，如外骨骼机器人"
80 ]
81 },
82 {
83 "cell_type": "markdown",
84 "metadata": {},
85 "source": [
86 "+ 人工智能主流方法\n",
87 " + **符号主义人工智能** - 00:34:25 \n",
88 " 符号主义人工智能以逻辑推理为主。如下图所示，已知三峡大坝和葛洲坝都位于同一个流域(长江流域)，并且这两个水坝具有反调节功能。那么当已知大古力水坝和达拉斯水坝也都位于哥伦比亚河时，可以推理这两个水坝也有反调节功能。可见符号主义人工智能不断从已有知识归纳规则，然后用这些规则推导出新的知识，使得知识图谱不断膨胀。 \n",
89 " <img src=\"https://imgbed.momodel.cn/1_页面_18.jpg\" width='500px'> \n",
90 " 另一个用符号主义进行推理的例子：IBM“沃森”推理系统 - 00:36:14 \n",
91 " 下图的IBM“沃森”推理系统可以根据主持人的一段话凝练一个和句子有关的问题。“沃森”的工作流程如下，“沃森”会首先扫描主持人的这段话，由于“沃森”不认识Kathleen，所以作为一个专有名词，而Kenyon也不认识，所以也作为一个专有名词，由于两个专门名词之间没有其他单词，因此作为一个复合专有名词。另外“沃森”认识Joshua，因此也把它提出来。最终将“Kathleen”，“Kenyon”，“Kathleen Kenyon”和“Joshua”放到搜索引擎中进行检索，其中前两个单词返回结果很少，第三个单词返回了大量信息，如右图黄色图谱所示。所以“沃森”很快提出了该问题。\n",
92 " <img src=\"https://imgbed.momodel.cn/1_页面_19.jpg\" width='500px'> \n",
93 " + **联结主义人工智能(数据驱动的机器学习)** - 00:39:36 \n",
94 " 若用符号主义识别猫，需要形成很多的规则，而数据驱动的机器学习则可直接通过数据学习特征，如下图所示。利用大量的数据，网络的浅层会首先学习一些低级特征，比如图中猫的耳朵、鼻子、腿等；然后中层网络会将浅层网络学习到的低级特征进行归纳，抽象出如猫的上半身或下半身等中层语义；网络最后的几层则汇集所有中层特征，抽象出猫的整体语义信息。然而数据驱动的机器学习可解释性不强，这也导致其存在一系列安全问题。\n",
95 " <img src=\"https://imgbed.momodel.cn/补画图_页面_4.jpg\" width='500px'> \n",
96 " + **行为主义人工智能(从经验中学习)** - 00:45:51 \n",
97 " 强化学习是典型的行为主义人工智能，即一个智能体不断和环境进行交互，智能体根据环境提供的奖励不断改进自己的行为决策。 \n",
98 " <img src=\"https://imgbed.momodel.cn/1_页面_22.jpg\" width='500px'> \n",
99 " 以AlphaGo为例介绍规则、数据和行为相互协调的智能算法 - 00:47:05 \n",
100 " DeepMind研究人员首先收集了16万人类选手棋局，利用监督学习的方法训练了一个神经网络希望该网络能模拟人类水平，但实际上这种方法只能打败业余选手，然后DeepMind研究者将该网络复制了两份，利用强化学习相互博弈训练，最终战胜了人类顶尖选手。 \n",
101 " <img src=\"https://imgbed.momodel.cn/1_页面_24.jpg\" width='500px'> \n",
102 " AlphaGo的总体概述如下图所示: - 00:53:01 \n",
103 " 图中表明深度学习是数据驱动的机器学习，强化学习是行为探索的机器学习，蒙特卡罗树搜索是知识驱动的机器学习。 \n",
104 " <img src=\"https://imgbed.momodel.cn/1_页面_28.jpg\" width='500px'> \n",
105 " 随后DeepMind研究者舍弃人类棋局数据，从零开始训练AlphaGo,并命名为AlphaGoZero(元)，最终达到了比原始AlphaGo更好的效果，但这种训练流程消耗了巨量的计算资源，并且和人类的学习方法相差甚大，如何在有限的计算资源下获得更好的效果或更近似人类的学习方式是未来需要进一步思考的问题。- 00:53:40\n",
106 " <img src=\"https://imgbed.momodel.cn/1_页面_29.jpg\" width='500px'> \n",
107 " 三种人工智能的区别如下图所示：- 00:58:05 \n",
108 " <img src=\"https://imgbed.momodel.cn/1_页面_30.jpg\" width='500px'> "
109 ]
110 },
111 {
112 "cell_type": "markdown",
113 "metadata": {},
114 "source": [
115 "### 2.人工智能课程体系 - 01:00:00\n",
116 "人工智能设置在CS学科，而CS学科又来自于EE。世界上第一个EE在1882年的德国设立。随后美国、英国先后设立。1968年美国计算机协会(ACM)设计了一套培养计算机人才的课程体系，其中人工智能当时就已经出现在课程规划之中。随后ACM和IEEE每隔一段时间就会更新并发布计算机课程体系，最新的结果如下所示。 \n",
117 "<img src=\"https://imgbed.momodel.cn/1_页面_37.jpg\" width='500px'> "
118 ]
119 },
120 {
121 "cell_type": "markdown",
122 "metadata": {},
123 "source": [
124 "### 3.课程内容概要 - 01:15:43\n",
125 "本课程的主要授课对象如下图所示。 \n",
126 "<img src=\"https://imgbed.momodel.cn/1_页面_56.jpg\" width='500px'> \n",
127 "电子信息工程硕士培养方案如下：- 01:17:03 \n",
128 "<img src=\"https://imgbed.momodel.cn/1_页面_57.jpg\" width='500px'> \n",
129 "本课程主要参考的教材是《人工智能导论：模型与算法》，总体框架如下：- 01:17:48\n",
130 "+ 第一章 人工智能历史与课程内容简介 (第一周: 吴飞老师、吴超老师)\n",
131 " + 介绍人工智能发展历史、现状和未来\n",
132 " + Python介绍\n",
133 "+ 第二章 机器学习系统设计 (第二周: 吴飞老师、况琨老师)\n",
134 " + 机器学习基本概念\n",
135 " + 回归分析模型与应用\n",
136 " + 深度学习模型与应用\n",
137 "+ 第三章 人工智能编程框架 (第三周: 吴超老师，华为工程师)\n",
138 " + Tensor/Pytorch\n",
139 " + MindSpore\n",
140 " + 若干应用\n",
141 "+ 第四章 词向量模型与应用 (第四周: 汤斯亮老师)\n",
142 " + 词向量表示\n",
143 " + 上下文相关的词向量与预训练模型\n",
144 " + 词向量与预训练模型在自然语言处理中的应用\n",
145 "+ 第五章 图向量模型与应用 (第五周: 杨洋老师)\n",
146 " + 图向量模型基本概念\n",
147 " + 图向量模型算法(deepwalk, LINE, node2Vec等)\n",
148 " + 推荐系统等应用\n",
149 "+ 第六章 人工智能芯片及高性能机器学习 (第六周: 王则可老师)\n",
150 " + 深度学习加速器简介\n",
151 " + 加速算法和模块\n",
152 " + 华为Asend深度学习加速器\n",
153 " + Google TPU和寒武纪Cambridge介绍\n",
154 "+ 第七章 分布式机器学习引擎 (第七周: 吴超老师)\n",
155 " + 联邦学习基本概念\n",
156 " + 联邦学习常用框架\n",
157 " + 分布式扩展：区块链、蒸馏聚合和数据定价\n",
158 " + FedAvg和蒸馏实战\n",
159 "+ 第八章 强化学习及其应用 (第八周: 况琨老师)\n",
160 " + 强化学习基本概念\n",
161 " + Policy Learning\n",
162 " + Q Learning\n",
163 " + 深度强化学习\n",
164 " + 典型应用(AlphoGo, Dota2，星际争霸等)\n",
165 " + 强化学习实战（机器人走迷宫等）"
166 ]
167 },
168 {
169 "cell_type": "markdown",
170 "metadata": {},
171 "source": [
172 "### 4.实训题目安排 - 01:36:40\n",
173 "基于“智海--Mo平台”和华为MindSpore框架完成如下任务： \n",
174 "+ 口罩佩戴检测：训练卷积神经网络深度学习模型， 识别图像中行人对象及判断其是否佩戴口罩（20分） \n",
175 "+ 作家文风判别：用鲁迅和张爱玲等作家的作品，训练作家文风判别器。输入一个句子，基于该判断其判别其文风（20分） \n",
176 "+ 机器人走迷宫：利用基础搜索算法和强化学习算法，让机器人探索并成功走出迷宫（20分） \n",
177 "\n",
178 "期末考试以单一选择题、多选题和判断题的机考形式进行（40分）。\n"
179 ]
180 },
181 {
182 "cell_type": "markdown",
183 "metadata": {},
184 "source": [
185 "### 附: 老师联系方式\n",
186 "+ 吴飞老师 https://person.zju.edu.cn/wufei/\n",
187 "+ 吴超老师 https://person.zju.edu.cn/0017053\n",
188 "+ 杨洋老师 https://person.zju.edu.cn/yangy\n",
189 "+ 王则可老师 https://person.zju.edu.cn/wangzeke\n",
190 "+ 汤斯亮老师 https://person.zju.edu.cn/siliang\n",
191 "+ 况琨老师 https://person.zju.edu.cn/kunkuang"
192 ]
342193 }
343194 ],
344195 "metadata": {
+1
-0
untitled.md less more
0 *123*⏎