BWSI Course - Serious Games and Artificial Intelligence
Course Overview
Combine modern methods in machine learning and game-like modeling to quantitatively analyze socially relevant technology and policy questions. This year’s application will be tactical routing for self-driving ambulances. We will build an analysis framework in Python to study the technical, moral, and strategic opportunities that new technologies present to that application. There will also be an emphasis on learning the practical tools and skills for working on a professional software development team.
The recent interest in gaming as a method for acquiring data on human-machine interaction, decision making and human factors has helped establish an emerging area of research called “Serious Games”. Examples of Serious Games can include:
• Training for dangerous, expensive, or rare situations
• Evaluation of critical factors in decision making
• Cognitive assessment for injuries and diseases that affect the brain
• Systems analysis
Examples, such as Foldit, a game which adds to the body of bioinformatics knowledge by challenging users to fold proteins, can actually make a significant scientific impact. Output from the game helped scientists understand the inherent structure of a key protease in a virus which causes HIV-like symptoms. As personal computing platforms become more prevalent (who, today, doesn’t own a cell phone?) the opportunity to help tackle critical challenges by harvesting the brainpower of millions while generating fun is tantalizing.
The program will consist of a one-month, intensive dive into the key aspects of serious gaming including: experimental design, game design, and application development. The course will examine and categorize different types of games, how to extract useful data, an introduction to User Interface design, rules development and play testing.
Students will be provided a basic introduction to Agile management, and coached as they follow the timeline for development. Completing the course will provide students with an understanding of software development, project management, human factors, game design, and technical collaboration as well as the emerging fields of artificial intelligence and serious games.
Prerequisites:
Python
Topics covered in this course:
Systems modeling
AI for gaming
Ethics for AI
Backend game development
Game mechanics and input interfaces
Human systems and user interfaces
Data logging and data analysis
Undead domain modelling
Agile team and software practices
Course Outline:
Week 1
Overview of serious games and their role in systems analysis and human factors
Role and ethics of AI in game design and development
Novel game extension to baseline game proposal
Public health policy overview
Week 3
Modeling undead disease propagation
Novel game extension development
Human Systems Interfaces, User Interface Design, Visualizing Information,
Data logging and analysis
Week 4
Game code finalization
Run the experiment (play the developed game) with data collection and analysis
Game debut
Final presentations and results
Expectations:
Students will focus on coding both a portion of the game back-end as well as self-designed extensions. With the assistance of instructors and Teaching Assistants, participants will learn about how Artificial Intelligence can impact the design of experiments and contrast with natural, human-centric game play. All students will participate in both back-end development, within a game-ready python framework, as well as coding of their own extensions. Introduction to supporting topics, including software development best practices for small teams, how to create user interfaces, bug and issue management, data visualization, public health and disease control, and technical presentation will be included.
Game Theme:
A single player game, which can also be played as by committee, describes the outbreak of a highly-contagious disease threatening a densely populated, urban area. Individuals who have contracted the disease have formed a zombie population which can be categorized into different architypes. These architypes may have different capabilities, propagation models, needs, and goals. The objective of the game will be to explore the efficacy of public health policies designed to deal with traditional disease outbreaks as applied to different infection models, methods and rates. Data analysis, such as the rate of infection compared against the implementation of different human or AI-enabled policy decisions, will provide an opportunity to visualize the results of different decision-making styles in remediating the humanitarian disaster.
