Abstract:This paper presents Coralai, a framework for exploring diverse ecosystems of Neural Cellular Automata (NCA). Organisms in Coralai utilize modular, GPU-accelerated Taichi kernels to interact, enact environmental changes, and evolve through local survival, merging, and mutation operations implemented with HyperNEAT and PyTorch. We provide an exploratory experiment implementing physics inspired by slime mold behavior showcasing the emergence of competition between sessile and mobile organisms, cycles of resource depletion and recovery, and symbiosis between diverse organisms. We conclude by outlining future work to discover simulation parameters through measures of multi-scale complexity and diversity. Code for Coralai is available at <a class="link-external link-https" href="https://github.com/aidanbx/coralai" rel="external noopener nofollow">this https URL</a> , video demos are available at <a class="link-external link-https" href="https://www.youtube.com/watch?v=NL8IZQY02-8" rel="external noopener nofollow">this https URL</a> .
What problem does this paper attempt to address?
The problem that this paper attempts to solve is: How to study the self - organizing behavior and evolutionary characteristics of Neural Cellular Automata (NCA) by simulating complex ecosystems. Specifically, the author proposes a framework named Coralai, aiming to explore the diverse ecosystems composed of NCA.
### Problem Background
Most life forms exist in complex biological networks. These organisms develop adaptability and resilience through interaction, physical environment, and co - evolution of the ecosystem. However, manually simulating such complexity is both impractical and limited by our understanding of the system. On the other hand, although cell automata based on simple rules (such as Conway's Game of Life) can produce complex behaviors, these behaviors are difficult to explain, control, and calculate.
### Coralai's Solution
The Coralai framework utilizes a modular, GPU - accelerated Taichi kernel, enabling NCA to interact in the simulation environment, change the environment, and achieve evolution through local survival, merging, and mutation operations. Specifically, Coralai solves the following key problems:
1. **Simulation of Diverse Ecosystems**:
- Coralai allows researchers to define various organisms, environments, and physical configurations, thereby exploring ecosystem behaviors under different conditions.
- Each experiment consists of Substrate, Sensors, Actuators, Kernel, and Physics. These components work together to simulate complex ecosystems.
2. **Efficient Evolution Mechanism**:
- Neural network parameters are generated through HyperNEAT (Hypercube - based Topology - enhancing NeuroEvolution), and mutation and crossover operations are implemented using PyTorch.
- This enables the system to simulate more complex and more life - like diverse behaviors.
3. **Flexibility of Physical Rules**:
- The use of a GPU - accelerated Taichi kernel enables the definition of arbitrary physical rules, which provides great flexibility for simulation.
- Experiments can be run with real - time visualization and can be scaled to larger system sizes.
4. **Exploration of the Emergence of Complex Behaviors**:
- By simulating slime - mold - like behaviors, it shows the competition between stationary and mobile organisms, resource consumption and recovery cycles, and symbiotic relationships between different organisms.
- Phenomena such as large - scale extinction events caused by greedy organisms, followed by the recovery of diversity and resilience, and the spread of spore - like cells in network - like organisms have been observed.
### Future Work
To further study these complex ecosystems, the author proposes the following directions:
- **Multi - scale Complexity (MSC)**: Detect self - organizing criticality through the renormalization group method, which is an important feature of life systems.
- **Measurement of Ecological Diversity**: Use NEAT's genetic similarity measure and lineage tracking to evaluate the diversity of the ecosystem.
- **Intelligent Search Strategy**: Develop meta - evolution algorithms to optimize parameters and physical rules, thereby better simulating life characteristics.
In conclusion, Coralai provides a powerful experimental platform for studying the self - organizing and evolutionary characteristics of artificial life systems.