Computational Abiogenesis - BFF Simulation

Abiogenesis simulation in python using pytorch. Inspired by https://arxiv.org/pdf/2406.19108.

This simulation is built from simple instructions:
- Increment/decrement head0/1 position (3 dimensions)
- Increment/ decrement array value at head0/1 position
- Copy value from head0/1 to head 1/0
- Enter/exit loop conditioned on value at head0

The simulation begins with a random array of integers of shape (128, 256, 64). You can think of this as a 128x256 array of instruction sequences with length 64. The instructions are each encoded as an integer value, with several no-op values that can be used to store data. Each iteration, a single instruction is executed for each sequence in the array, moving down the depth of the array with each cycle.

From random noise, self-replicating programs emerge and evolve.

Differences from the original BFF simulation:
- Instruction sequences interact with neighbors through localized operations instead of concatenation
- Each instruction sequence has the same computation budget: one instruction per iteration

Code: https://github.com/Nebraskinator/BFF