Matrix：代际人工生命（GAL）框架研究人工种群的文化累积

psycho-prince/Matrix

AI 导读

Matrix 是一个名为"代际人工生命（GAL）"的实验框架，研究持久存在的人工种群——个体经历发育、学习、工作、交流、向继任者传递知识并参与多代文化演化，而非无限训练单一模型。其核心研究问题是：代际连续性能否产生独立 AI 智能体难以高效形成的累积能力。项目目前为研究原型，代码采用 Apache License 2.0，论文与文档采用 CC BY 4.0。

正文

What happens when AI becomes a population rather than a model?

Matrix: Generational Artificial Life (GAL) is an experimental framework for studying persistent artificial populations whose individuals develop, learn, work, communicate, transfer knowledge to successors, and participate in multi-generational cultural evolution.

Instead of training one model indefinitely, GAL investigates whether intelligence and knowledge can accumulate at the population level through generations of artificial individuals.

                Individual
                   │
                develops
                   │
                 learns
                   │
                 works
                   │
                teaches
                   │
               successor
                   │
                 dies
                   │
              next generation
                   │
                   ▼
             accumulated culture

Research Question

Can generational continuity produce cumulative capabilities that independent artificial agents cannot efficiently develop?

Status

🚧 Research prototype

Current focus:

persistent artificial individuals
finite lifetimes
developmental learning
knowledge inheritance
successor training
population dynamics
cultural evolution

Project Structure

gal/: Core framework for the Matrix: Generational Artificial Life simulation.
experiments/: Reproducible experimental setups.
docs/: Architecture and theoretical framework.
paper.pdf: Research manuscript.
configs/: Experiment configuration files.

License

Source Code: Apache License 2.0
Research Papers & Documentation: CC BY 4.0
Datasets: CC BY 4.0 / CC0 (as specified per dataset)

来源：GitHub · github.com