The Computer Science of Science & Slime Molds
What do the Scientific method and a single-celled organism called Physarum polycephalum have in common?
Last Spring I took a class on Human Behavioral Biology with Robert Sapolsky (the Lectures are online and I highly recommend them). During our lecture on Emergent Complexity & Spontaneous Order, Sapolsky gave an example of how Physarum polycephalum, a bright yellow slime mold, can solve mazes in the shortest possible path [1]. At the time, I thought it was neat but didn’t think much of it.
Earlier this month, I was exploring how biological networks dynamically balance fault tolerance and efficiency as inspiration for my Bitcoin research — and there it was again! That’s when I thought, “I better pay attention to this slime mold.” Ultimately, I read two research papers; the first was referenced by Sapolsky and the second showed how the slime can also solve the Traveling Salesman Problem [2]. As I finished the second paper, it hit me that the slime mold was using Recursive backtracking [3].
To start, researchers placed “food” at the goal of the maze. As the experiment began the slime would search and trace every path, growing and expanding as it did. If it ran into a dead end, its body shrunk due to a lack of food so, over time, wrong paths were removed and, ultimately, only the shortest path through the maze remained. Not too shabby for a slime mold, right?
One of the reasons I began to admire this unassuming creature was how much it reminded me of the Scientific method and my learning & entrepreneurial experiences. As one begins learning something new and complex (‘exploring the maze’) one can go down paths that are dead ends. Thus, one must be humble and forgiving enough to start over or forge a new path. It meant something I feel really deeply: that learning is an investment, failure is a sign of progress, and, often times, if you stick with it you can find a solution.
Einstein once said, “It’s not that I’m so smart, it’s just that I stay with problems longer.”
Special thanks to my friends Gabe Alvarez and Nathan Eidelson.
Sources:
[1] Toshiyuki Nakagaki, Hiroyasu Yamada & Ágota Tóth, ‘Intelligence: Maze-solving by an amoeboid organism,’ http://www.nature.com/nature/journal/v407/n6803/full/407470a0.html
[2] Atsushi Tero, et al, ‘Rules for Biologically Inspired Adaptive Network Design,’ http://www.uvm.edu/~pdodds/files/papers/others/2010/tero2010a.pdf
[3] Eric Roberts, ‘Chapter 7 Backtracking Algorithms,’
http://cs.stanford.edu/people/eroberts/courses/cs106b/chapters/07-backtracking-algorithms.pdf
