Through the looking-glass with ALiCE: artificial life, culture and evolution:
UCLA Human Complex Systems Program
Last modified: April 25, 2006
In 15 classes over the last 7 years, we have developed a successful course in which participants explore the simulated worlds of artificial life, artificial culture and evolutionary computation. Philosophically, we contrast discursive, mathematical and computational modeling in their effectiveness in describing, understanding and explaining complex phenomena. We show them examples of cutting-edge research in the field and introduce them to its practice in a critically informed manner by having them create and run their own software using Borland’s C++ Builder for PCs. Students of both gender range across status from freshmen to graduates, across major from the humanities to the sciences, and across experience. Nintety percent of participants no programming experience, though we have had a few computer science majors. We meet for two 3-hour sessions per week for 10 weeks in a lab with 24 work-stations. We build a simulation from scratch on the very first day, and then advance, step by step, through several open-ended simulations, each one introducing new concepts and building on those that came before. We discuss and sample different representations and granularities of space (cellular and vector) and time (sequential, random and cinematic polling) and how they introduce biases into the simulation. We progress from single-run simulations, through those that gather statistics on several runs, to those that automatically explore a wide range of parameters. We move from simple agents and environments to more compelx evolutionary models with both fixed and dynamic fitness allocation. We emphasize a quick feedback cycle of writing-compiling-and-running code to build confidence. We stress the importance of good visualization and sonification practices as both debugging tools and as rich analytical channels for understanding the complexities that are taking place. The course project is a simulation of each student’s own creation. Our goal is to empower students with first-hand experience in the promises and limitations of simulations in order that they may be informed advocates and critics of the field. We hope to give them the confidence and resources to explore this expansive field on their own as well as to realistically evaluate the assumptions behind the simulations they may encounter in their careers. The Honors Collegium version of this course has won the UCLA Brian C. Copenhaver “Teaching with Technology Award” for introducing students to how to REASON with technology.