2024 Faculty Courses School of Computing Department of Computer Science Graduate major in Artificial Intelligence
Responsible Artificial Intelligence
- Academic unit or major
- Graduate major in Artificial Intelligence
- Instructor(s)
- Jun Sakuma
- Class Format
- Lecture (HyFlex)
- Media-enhanced courses
- -
- Day of week/Period
(Classrooms) - 5-6 Tue / 5-6 Fri
- Class
- -
- Course Code
- ART.T555
- Number of credits
- 200
- Course offered
- 2024
- Offered quarter
- 3Q
- Syllabus updated
- Mar 14, 2025
- Language
- English
Syllabus
Course overview and goals
This course provides an advanced look at responsible artificial intelligence. We will understand machine learning in a unified way from the framework of empirical risk minimization and analyze the generalization capabilities of machine learning. Students will then learn the operating principles of deep learning, the requirements for decisions made by AI to be trusted, and the theory and technologies for achieving these requirements. Specifically, students will learn about explainable AI, data privacy protection, AI fairness, and AI security. The aims of this course are to 1) acquire knowledge of the principles and mechanisms of machine learning and deep learning and 2) understand the mechanisms necessary for machine learning and deep learning to be trusted by humans.
Course description and aims
By the end of this course, students will learn the following:
1) machine learning technologies
2) responsible AI technologies
Student learning outcomes
実務経験と講義内容との関連 (又は実践的教育内容)
Experience working with data science/AI in a private company
Keywords
Empirical risk minimization, statistical learning theory, deep learning, explainable AI, data privacy, AI fairness, AI security
Competencies
- Specialist skills
- Intercultural skills
- Communication skills
- Critical thinking skills
- Practical and/or problem-solving skills
Class flow
Every class consists of a lecture using the slides and the exercise
Course schedule/Objectives
| Course schedule | Objectives | |
|---|---|---|
| Class 1 | Introduction to responsible AI | Lecture objectives and overview of respoisible AI |
| Class 2 | Empirical risk minimization 1 | Understanding the framework of supervised learning and empirical risk minimization |
| Class 3 | Empirical risk minimization 2 | Understanding the framework of supervised learning and empirical risk minimization |
| Class 4 | Deep Discriminative Models | Neural network for discriminative models |
| Class 5 | Fairness of discriminative models | Fairness of decisions made by discriminative models |
| Class 6 | Privacy of discriminative models | Privacy violation and protection of discriminative models |
| Class 7 | Security of discriminative models | Attacks and defenses against discriminative models |
| Class 8 | Explainablityu of discriminative models | Discussion on responsibility of discriminative models |
| Class 9 | Language model | Mechanism of language models |
| Class 10 | Alignment of language models | What is a misalignment in language models, and how can it be prevented? |
| Class 11 | Privacy of language models | Privacy breaches caused by language models and how to combat them |
| Class 12 | Security of language models | Attacks and defenses against language models |
| Class 13 | Interpretability of language models | Methodology for understanding and interpreting the computational processes of language models. |
| Class 14 | Discussion | Discussion on the responsibility of AI |
Study advice (preparation and review)
To enhance effective learning, students are encouraged to spend approximately 100 minutes preparing for class and another 100 minutes reviewing class content afterwards (including assignments) for each class.
They should do so by referring to textbooks and other course material.
Textbook(s)
Slides provided
Reference books, course materials, etc.
Foundations of Machine Learning, Mehryar Mohri, Afshin Rostamizadeh, and Ameet Talwalkar, The MIT Press
Evaluation methods and criteria
Evaluation by confirmation questions at the start of each lecture (65%) and final report (35%)
Related courses
- CSC.T254 ： Machine Learning
- ART.T458 ： Advanced Machine Learning
- CSC.T242 ： Probability Theory and Statistics
- CSC.T272 ： Artificial Intelligence
- CSC.T352 ： Pattern Recognition
Prerequisites
Experience with elementary linear algebra, analysis, probability theory, and statistics at the undergraduate level is required.