Quality Alchemist (品質煉金術師): The 2nd Nation Higher Education Teacher Training Course for "Introduction to Artificial Intelligence" (第二期全国高校《人工智能导论》师资培训班) (Day 1)

2018年12月14日星期五

The 2nd Nation Higher Education Teacher Training Course for "Introduction to Artificial Intelligence" (第二期全国高校《人工智能导论》师资培训班) (Day 1)

I am a Committee Member of Extenics Society, Chinese Association for Artificial Intelligence (CAAI) (中國人工智能學會可拓學專業委員會委員) and attended the 2^nd Nation Higher Education Teacher Training Course for “Introduction of Artificial Intelligence” (第二期全国高校《人工智能导论》师资培训班) organized by CAAI from 14^th to 16^th Dec 2018 in Fuzhou.

Before the training course, I met the executive editor Prof. Yu Jian (于剑) (北京交通大学人工智能研究院常务副院长、教授) of the book named “Introduction of Artificial Intelligence” which was published by China Science and Technology Press in Aug 2018. We took a photo for memory.

I met Prof. Liao Luchao (廖律超) (福建金學院信息科學與工程學院副教授).

I also recognized two classmates from AI education industry and they were Ms. Fan Liyang (樊麗陽) (原力AI學院董事長兼常務院長; 中國人工智能學會工委會副主任) and Mr. Hao Guo (郭浩) (原力AI學院副院長).

In the beginning, Prof. Wang Weining (王卫宁) (中國人工智能學會秘書長) gave opening speech and introduced all guest speakers.

Then Mr. Zhang Qingya (張青雅) (永泰縣副縣長) gave welcome speech and introduced Fuzhou Yongtai.

Moreover, he introduced their Yongtai AI City plan to us.

And then Prof. Ma Shaoping (马少平) (清华大学计算机系教授; 中国人工智能学会副理事长) represented CAAI to give a speech. He said the first Nation Higher Education Teacher Training course was held in Beijing successfully and more than 100 participants at that time. We needed to train more talents for China AI development.

Guest speakers were photo taken.

Then all participants were invited to take a group photo.

Day 1 (20181214):

The first topic named “Introduction of AI” was presented by Prof. Yu Jian (于剑). Firstly, he briefed China AI policy which issued on 20^th July 2017 named “A Next Generation Artificial Intelligence Development Plan” (新一代人工智能发展规划).

Then Prof. Yu Jian briefed the history of AI since 1956 and its definitions. One of famous testing named “Turing Test” (图灵测试) was introduced.

After that Prof. Yu introduced three key AI school of thought (學派) and they were symbolicism (符號主義), connectionism (聯結主義) and behavioralism (行為主義).

Finally, Prof. Yu raised the AI education in China

The second topic named “Concept Presentation” was also presented by Prof. Yu Jian (于剑). He mentioned what is the Proposition (命題) that was a description statement to have certainty true or false.

And then he introduced Set Theory.

Lastly, Prof. Yu briefed the theory of Modern Concept Presentation.

The second speaker was Prof. Wang Xuan (王轩) (中国人工智能学会青年工作委员会主任; 哈尔滨工业大学(深圳)计算机学院院长、教授) and his topic was “Analysis of Speech Processing” (淺析语音处理). He would introduce speech identification, integration, enhancement, transformation and emotional process.

Firstly, Prof. Wang briefed the foundation concept of speech and its recognition method in which needed to extract speech characteristics.

After that he introduced how to employ Deep Neutral Network (DNN) into automatic speech recognition.

Finally, Prof. Wang stated how to employ emotion speech for speech processing so as to similar human communication.

The last speaker in Day 1 was Prof. Zhang Daoqiang (张道强) (南京航空航天大学计算机与科学技术学院副院长、教授) and his topic stated “Machine Learning” (机器学习). Before his talk, he asked a question that do AI rule the world.

Then he briefed the AI history, where Geoffrey Hinton is the father of “deep learning”.

And then Prof. Zhang compared the different between traditional programming and machine learning in the following diagram.

Machine Learning history was reviewed. Tom Mitchell (1998) gave definition that Machine Learning is the study of algoritms that:

- Improve their performance P

- At some task T

- With experience E.

A well-defined learning task is given by <P, T, E>.

Prof. Zhang’s second topic named “Introduction of Computer Vision” (计算机视觉简介). He said computer vision is everywhere.

Firstly, Prof. Zhang compared human vision and computer vision. And he mentioned the basic principles of computer vision such as classification and regression.

And then he briefed the process of tradition method of computer vision that extracted the characteristic either by supervision or un-supervision.

Recently, Deep Learning was employed in computer vision. The error rate decreased from 26.25% (2012 using AlexNet) to 6.6% (2014 using GoogleNet) and down to 3.6% (2015 using ResNet).