Course title
L09960003
Game Informatics

 IGARASHI Harukazu
Course description
In recent years, there has been a remarkable development in artificial intelligence, represented by deep learning. Traditionally, intelligent games such as chess and Go have been studied as research targets for artificial intelligence, including fast game tree search methods and learning methods for state evaluation functions. Recent large-scale digital games can also be regarded as one of the application cases of distributed cooperative AI systems. Furthermore, there is a growing expectation that game solving algorithms and basic techniques used in computer games can be applied to real-world problems in various fields such as information engineering and mathematical engineering.

In this course, students will learn basic techniques of artificial intelligence (search, machine learning, etc.) used in game solving algorithms, and understand how to apply them in application cases. In the future, the goal is to develop the ability and sense of algorithm design so that it can be applied to actual real-world problems.
Purpose of class
Learn the basic techniques of artificial intelligence used in solving and design algorithms for computer games, and understand their applications.
Goals and objectives
  1. Understand and briefly explain the informatics definition and classification of games, and the history of game informatics.
  2. Understand and briefly explain the basics of game tree search and learning evaluation functions. (Class 1)
  3. Understand and briefly explain the basics of game tree search and learning evaluation functions. (Class 2-3)
  4. Understand and briefly explain application cases related to game AI. (Class 4-7)
Language
Japanese
Class schedule

 Class schedule HW assignments (Including preparation and review of the class.) Amount of Time Required
1. Games and information: definition of games, informatics classification of games, games and problem solving (game trees, search volume), history of game informatics Read the handouts and chapters 1 and 2 of the reference book [1]. 100minutes
2. Basic knowledge (1): Typical search methods: A* algorithm, αβ search, Monte Carlo tree search, MC Softmax search Read the handouts. 200minutes
3. Basic knowledge (2): Gradient method, evaluation function/policy learning methods (supervised learning and reinforcement learning) and models (neural network model, linear sum of features, if-then rule, Q-table, action tree) Read the handouts and chapters 4 and 6 of the reference book [2]. 200minutes
4. Applications (1) Intelligent complete information games: History, search methods, evaluation functions and learning methods of Shogi, Go, Chess. AlphaZero Read the handouts and chapters 6 and 7 of the reference book [1]. 200minutes
5. Applications (2) Dynamic imperfect information/multi-agent games: History, rules, and research examples of RoboCup soccer (learning evaluation functions: applications of reinforcement learning, supervised learning, and imitation learning) . Iterated prisoner's dilemma games (Nash equilibrium, Folk theorem). Read the handouts and chapter 5 of the reference book [1]. 200minutes
6. Applications (3) Real-time Video Games: An example of deep Q-learning and model-based reinforcement learning applied to the Atari 2600 games. Read the handouts and chapter 11 of the reference book [2]. 200minutes
7. Application example (4) Large-scale digital games(action games and RPGs) : AI system in Final Fantasy XV(Meta AI, AIGraph, Navigation AI) Read the handouts and the reference book [3]. 100minutes
Prepare a report document. 300minutes
Total. - - 1500minutes
Relationship between 'Goals and Objectives' and 'Course Outcomes'

 Report Total.
1. 10% 10%
2. 20% 20%
3. 70% 70%
Total. 100% -
Evaluation method and criteria
Report:

[Question 1] Select a research paper on the subject of computer games, and write a brief summary of the artificial intelligence techniques and mechanisms (e.g., search, machine learning, reasoning, knowledge representation) used in the paper or book (60%). It is desirable to write in a clear and creative manner.

[Question 2] Also, please discuss the problems, improvements, and future prospects of the game and its techniques, and describe your own opinions. It can be the design of a new game or game system, or the proposal of a new solution algorithm for an existing game (40%).
Feedback on exams, assignments, etc.
ways of feedback specific contents about "Other"
Textbooks and reference materials
[1] T. Ito，K. Hoki and Y. Miyake : “Game infomatics”, 2018, Coronasha (in Japanese)
[2] M. Taki : “Introduction to deep learning”, 2017, Kodansha (in Japanese)
[3] Y. Miyake : “Game AI General Theory and its Implimentation in AAA Degital Game - A Case Study of AI System in FINAL FANTASY XY-,” Artificial Intelligence，vol.35, N.2,pp1-16, 2020 (in Japanese)

* Reference materials other than those listed above will be introduced during the lecture as needed.
Prerequisites
It is recommended that students take "Artificial Intelligence" in the first semester of their 3rd year.
Office hours and How to contact professors for questions
  • Questions are accepted in the classroom or elsewhere after class. At other times, you can ask questions or consult with me by e-mail.
Regionally-oriented
Non-regionally-oriented course
Development of social and professional independence
  • Course that cultivates a basic self-management skills
  • Course that cultivates a basic problem-solving skills
Active-learning course
N/A
Course by professor with work experience
Work experience Work experience and relevance to the course content if applicable
N/A N/A
Education related SDGs:the Sustainable Development Goals
  • 4.QUALITY EDUCATION
  • 9.INDUSTRY, INNOVATION AND INFRASTRUCTURE
Last modified : Sat Sep 09 08:12:58 JST 2023