Reproduce “Automating the Search for Artificial Life with Foundation Models” on OpenResearch

Turn your favorite coding agents into research agents that review literature, develop hypotheses, and run experiments.

Universal app for Apple silicon and Intel. Requires macOS 11 or later.

About this paper

Akarsh Kumar, Chris Lu, Louis Kirsch, Yujin Tang, Kenneth O. Stanley, Phillip Isola, David Ha

A framework leverages vision-language foundation models to automate the discovery and analysis of artificial life simulations, successfully finding targeted, open-ended, and diverse emergent behaviors. This approach also quantifies previously qualitative aspects of complex systems using human-aligned representations.

Key insights

  • Vision-language models provide human-aligned representations, allowing quantitative evaluation of previously qualitative ALife phenomena.
  • The ASAL framework enables automated, systematic exploration of ALife simulation spaces, reducing reliance on manual design and human intuition.
  • This approach offers a generalizable paradigm for ALife research, adaptable to various simulation substrates and future foundation models.

Summary by alphaXiv.

Tip: change arxiv.org to autoarxiv.org in any arXiv abstract link to open the paper here.