Bio-inspired growth: introducing emergence into computational design

In today’s age of neural networks and brain study, creativity is being introduced into lifeless systems by modelling the concept of learning. Many believe the artificial intelligence that is leading technology will eventually do most of a designer’s work. However, this artificial intelligence only results after long hours of training and is limited to the area within which it is trained. In nature, many systems can produce unpredictable solutions without the retention of information - such as trees. Although computers cannot accurately model nature’s growth mechanisms, it can be approximated with the concept of predictive non-determinism – where what is not understood is treated as random - and the rest of the system built around this. This paper lays out a four-tiered structure, inspired by growth principles seen in nature, for introducing emergence into the design system. The models presented are grown by random functions, controlled by a restriction of misfit and guided by the concept of fitness. It gives a bottom up approach to the design, with the user providing the desired functionality and asking what the possible designs are. The resulting models grown by these stochastic rules are emergent, providing the computer with the chance of creating unexpected and innovative solutions.