Intelligent Information Systems Exercise 1

In the field of artificial intelligence, research in neural computing, distributed artificial intelligence, and complex system has been active since around 1980. What is common about these research subjects is their targeting of large-scale and difficult problems in the natural science and social science fields, and their algorithm design ideology to solve the problems of control and learning for the overall system by modeling systems as groups of simple information processing units such as neuro elements and agents. In the backdrop of such research, research methods from fields other than conventional information science have played a large role. They include improvements in computer performance enabling large-scale calculations for large-scale simulations and machine learning, information processing models in the human brain, the evolutionary process of organisms, and technical methods for handling many-body systems in statistical physics.

The following research will be conducted from the aforementioned perspective.

1. Basic research in multi-agent systems
Agent modeling, agent learning through reinforcement learning, emergence and learning of coordinated behavior
2. Applied research in multi-agent systems
Pursuit problem, soccer game, navigation of mobile robots, path planning, search in optimization problems
3. Basic theoretical research in reinforcement learning
Reinforcement learning method in the non-Markovian process
4. Basic research in natural language processing
Semantic analysis of natural language sentence, application and acquirement of language knowledge, statistical natural language processing
5. Applied research in natural language processing
Interactive user interface, information search and integration on the Web, application in education and daily living
6. Basic research in programming languages and its applications to program development systems
Completing variable names for functional languages, code clone detection

1. Understand basic theories and techniques of machine learning such as neurocomputing and reinforcement learning
2. Understand basic concetps and structures of multiagent systems and make a simple simulation system
3. Understand basic theories and skills of natural language processing and apply those skills to simple examples
4. Understand basic techniques for processing programs and apply them to a simple language. Design and implement tools for programming.

Japanese(English accepted)

Apply new technologies and theories in this field to research themes. That will develop student's ablities of problem finding and resolution. Challenging to basic and theoretical research topis is more desirable.

Students will be evaluated based on progress reports (more than 3 seminars/week) (50%), academic society presentations (10%), and writing of master's theses (40%).

Will be specified, if necessary.

Probability theory, linear algebra, knowledge on partial differential, and discrete mathematics

• H. Igarashi: 17:00-18:00, Tuesday

Non-environment-related course

Non-regionally-oriented course

• Course that cultivates a basic problem-solving skills
• Course that cultivates an ability for utilizing knowledge

Most classes are interactive