Project Story

About the Project

"Worm Civilization: The Rise of Intelligence" is an AI-driven simulation that explores the evolution of intelligence in C. elegans (nematode worm and real-life model organism) by treating each worm as an autonomous AI agent. Inspired by the OpenWorm project and the rise of synthetic and generative biology in context of AI-driven biological tools like alphafold, ESM3, and evo2, I set out to build a simulation where the player, who is cast as a worm-obsessed mad scientist, guides the evolution of worm intelligence through genetic modifications, environmental challenges (such as mazes), and neural connectome adjustments.

What Inspired Me:

The OpenWorm project and its comprehensive mapping of the C. elegans neural connectome sparked my interest in simulating real neural networks. Since the C. elegans nervous system contains only 302 neurons, it is feasible to visualize the entire thing! While the OpenWorm Github makes the C. elegans connectome data

I was also inspired by evolution simulation games, but wanted to ground the experience in real scientific principles and AI-driven behavior. I believe that that this project occupies a unique niche, where there is enough grounding in real-world biology to spark curiosity and inquiry-driven learning, but the premise is (hopefully!) enjoyable on its own merit.

What I Learned:

• I deepened my understanding of genetic algorithms, reinforcement learning, and multi-agent systems.
• I explored how small changes in evolutionary parameters can lead to emergent intelligence over generations.
• I learned to integrate environmental challenges (mazes, shifting food patches, and obstacles) to force the evolution of adaptive, memory-driven behaviour in a simplified agar-plate ecosystem simulation.

How I Built My Project:

I used VScode and Cline to create a game using PyGame to create a simulated environment, defined the worm class, implemented an evolutionary loop, and integrated a UI for the mad scientist player character featuring sliders for memory, exploration, neural plasticity, and decision speed, a visualization of worm movement over the agar plate environment, and a visualization of the C. elegant neural connectome.

Challenges I Faced:

• Implementing a multi-agent evolutionary system and coding each of the 10 worms as its own AI agent
• Simplifying the C. elegans connectome and displaying it as an interactive visualization
• Gamifying landmarks in worm intelligence and social development
• Balancing the simulation so that environmental challenges promote intelligence without wiping out the worms' health
• Managing development time and scope within the hackathon timeframe, which required prioritizing key features and iterative improvements

Built With

• chatgpt
• claude
• cline
• numpy
• pygame
• python
• scikit-learn