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 3)

2018年12月16日星期日

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 3)

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) was organized by CAAI from 14^th to 16^th Dec 2018 in Fuzhou. In Day 3, I met Prof. Ma Shaoping (马少平) (清华大学计算机系教授、博士生导师; 中国人工智能学会副理事长) and took a photo for memory. He told me that he had visited Department of Computer Science, CityU many years ago.

Day 3 (20181216):

Prof. Ma Shaoping (马少平) was the first speaker in the Day 3 and his topic named “Search Technology” (搜索技术). Search problem included space and path.

Then Prof. Ma introduced different search algorithms. Blind search (盲目搜索) included Depth First Search (DFS) (深度优先搜索) and Width First Search (WFS) (宽度优先搜索). DFS was not guarantee to find the optimal solution if the depth limit was wrong. WFS should find the solution if it had but resource consumption was high (see following diagram).

And then Prof. Ma mentioned the Heuristic Algorithm (启发式算法) which also called A Searching Algorithm (A算法). If A Searching Algorithm fulfilled heuristic function - h(n) <= h*(n), it became A* (A Start) Searching Algorithm which usually used for the process of finding a path between multiple points, called "nodes".

After that he stated Alpha–beta pruning (α-ꞵ剪枝) which employed for chess game (except GO) that seek to decrease the number of nodes and evaluated by the minimax algorithm in its search tree.

Finally, Prof. Ma Shaoping introduced Monte Carlo Tree Search (MCTS) (蒙地卡罗树搜寻) which had four steps and they were Selection, Expansion, Simulation and Backpropagation.

The second topic presented by Prof. Ma Shaoping was “Expert System” (专业系统). Expert system is rule-based system that emulates the decision-making ability of a human expert.

Then Prof. Ma briefed the general structure of expert system including human-machine interface, translator, dynamic database, logic machine, knowledge database and knowledge engineering for collecting knowledge.

After that Prof. Ma using some examples to demonstrate how expert system operated.

(Remark: My Applied Physics Degree’s final year project was to design expert system named “Implementation of Design Cases in a Material Selection Expert System” which had finished in April 1994.)

The second speaker was Prof. Bao Jie (鲍捷) (文因互联 CEO、联合创始人) and his presentation topic was “Introduction to Knowledge Graph” (知识图谱入门). Originally, Knowledge Graph is a knowledge base used by Google and its services to enhance its search engine's results with semantic-search information gathered from a wide variety of sources in 2012.

Knowledge Graphs combine characteristics of several data management paradigms. It included knowledge extraction, knowledge performance and logic, knowledge storage and knowledge searching.

Then Prof. Bao briefed the milestone of Knowledge Graphs since 1960s.

Lastly, Prof. Bao mentioned that one of application of Knowledge Graphs was in Chabot.

After the break, Prof. Liao Luchao (廖律超) and I took a photo with Prof. Ji Jianmin (吉建民) for memory.

Prof. Ji Jianmin (吉建民) (中国科学技术大学计算机科学与技术学院副教授) was the last speaker and his topic presented was “Planning in Artificial Intelligence” (规划 - 人工智能导论). Planning is decision making about the actions to be taken.

He said that Automatic Planning started since 60s which was one of important areas in AI. It included Classical Planning, Neoclassical Planning and Planning under Uncertainty. After that he introduced the Intelligent Planning which involved status and action.

Finally, he mentioned Markov decision process (MDP) (马可夫决策过程) which provided a mathematical framework for modeling decision making in situations where outcomes are partly random and partly under the control of a decision maker. Moreover, he also stated Partially Observable Markov decision process (POMDP) (部分可观察马可夫决策过程). MDP had observation and reward between World and Agent. Where POMDP had “Belief State Estimator” inside Agent.

The last topic Prof. Liao Luchao presented was “Intelligent Robot” (智能机器人). Firstly he introduced RoboCup@Home which aimed to develop service and assistive robot technology with high relevance for future personal domestic applications.

Then he demonstrated different version of service robot.

After that he briefed the Service Robot System Framework. The Task Planning was key element in the whole system. Robot System was very complex and integrated different knowledge in AI and Mechanical Control.

Another important element was Natural Language Processing (自然语言处理) that could communicate between human and robot.

The beautiful Chinese robot named “佳佳”.

Finished 3 days training, we left Yongtai and back to Fuzhou for dinner.

Prof. Lin Chang (林暢) (Chairman, Macau Quality Management Association (MQMA); Professor, School of Social Development and Public Policy, Beijing Normal University) invited Prof. Liao Luchao (廖律超) (中国科学技术大学计算机科学与技术学院副教授) and I to have dinner in Fuzhou and we took a photo for memory.