Social Behavior Modeling & Social Stimulation

This course introduces graduate students to the interdisciplinary field of computational social science, with a focus on modeling social behavior using network science and large language model (LLM)-based social simulation. Students will learn how computational approaches can be used to analyze and simulate collective human behavior, including opinion formation, cooperation dynamics, and information diffusion in social networks. The course is delivered in a fully on-demand format using an LLM-driven active learning system, where students engage with academic papers in the field through AI-facilitated dialogue, fostering deep comprehension and critical thinking skills.

This course covers two major pillars of computational social science: (1) network science for understanding collective behavior, and (2) LLM-based social simulation for modeling social phenomena. In the first half, students study foundational concepts of network science, including graph theory, network structure, and models of social influence and collective dynamics. In the second half, the focus shifts to how large language models can be used as autonomous agents to simulate social interactions and test hypotheses about human behavior. Throughout the course, students read and critically analyze seminal and recent academic papers in computational social science using an AI-powered teaching system (Cat’s Paw BotTA). This system acts as a ”teacher agent” that guides students through each paper via interactive dialogue, assesses their understanding in real time, and provides targeted feedback. The on-demand format allows students to progress at their own pace while ensuring deep engagement with the material.

1. Understand the fundamental concepts and analytical frameworks of computational social science, including network science and agent-based modeling.
2. Demonstrate the ability to engage in structured academic dialogue, articulating key findings, methodological approaches, and limitations of research papers.
3. Synthesize knowledge from multiple papers to propose potential research directions in computational social science.

This course emphasizes the use of generative AI.
The final grade is calculated based on 12 reading assignments using generative AI (total 80 points) + attendance (20 points). Detailed evaluation criteria will be explained in the first session guidance materials.
Based on the above, a comprehensive evaluation will be conducted, and a score of 60 or above out of 100 points is required to pass.
※To achieve 60 points, students need to reach at least 30 points in the reading assignments for every week and finish the feedback survey in the first and the last week.

No specific textbook is required. All course materials consist of academic papers assigned each week, which will be made available through the AI teaching system (Cat’s Paw BotTA).

Basic knowledge of programming (preferably Python) and statistics is recommended. Familiarity with linear algebra and probability theory is helpful but not required. No prior knowledge of network science or social simulation is assumed.

• By appointment. Students are encouraged to contact the instructor via email. During on-demand sessions, the AI teaching system can also address questions about weekly paper content.

Non-regionally-oriented course

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

Most classes are interactive

• 1.NO POVERTY
• 4.QUALITY EDUCATION
• 5.GENDER EQUALITY
• 9.INDUSTRY, INNOVATION AND INFRASTRUCTURE
• 11.SUSTAINABLE CITIES AND COMMUNITIES
• 13.CLIMATE ACTION
• 16.PEACE, JUSTICE AND STRONG INSTITUTIONS