The study not of not how life is, but how it could be (also, a lot of population genetics).
During my undergraduate years at Indiana University I got into Artificial Life. The most notable work that came out of this was my work under Larry Yaeger on the Polyworld artificial life simulator which resulted in two papers.
- Ideal Free Distribution in Agents with Evolved Neural Architectures
- Passive and Driven Trends in the Evolution of Complexity
"Passive and Driven Trends" is the better of the two. There is also a YouTube of me talking about Polyworld (obviously out of date, but captures the gist of more modern work). Looking back on it I'm embarassed by how much I stuttered---I was super-nervous.
Later studying information theory in artificial life under Chris Adami moved me towards quantitative work. From Polyworld I was into brains, but with the quanitative shift I switched to more pure genetics work. My reasoning was, "Brains are cool, but genes are pretty neat too, and genes are waaaay simpler than brains. Start with genes." After I felt like I had sufficient grip on information theory of genes I formally moved into the nitty gritty of neural complexity. There was an old joint project with Chris Adami and Henry Strickland on quantifying the suitability of different programming languages for genetic programming, but we never found results we felt were generalizable enough to publish---in short, languages are hard.
I also have an orphaned paper from this time, On the Viability of Self-Reproducing Machines that was intended to be co-authored with Doyne Farmer and Eric Smith. This project is officially orphaned and you are invited to steal it---just mention us in the acknowledgements ;)
For interested students
My most intellectually formative works in this topic came from two books: