In Search of the Ingredients of Open-Endedness: Replicating Picbreeder with Large Vision-Language Models Conference: Genetic and Evolutionary Computation Conference; July 13–17, 2026; San Jose, Costa Rica Genetic and Evolutionary Computation Conference (GECCO ’26), July 13–17, 2026, San Jose, Costa Rica DOI: 10.1145/3795095.3805186 ISBN: 979-8-4007-2487-9/2026/07 CCS: Computing methodologies Multi-agent systems CCS: Computing methodologies Cognitive science
Sam Earle, Kai Arulkumaran, Andrew Dai, Akarsh Kumar, Julian Togelius, Sebastian Risi · 2026
Read In Search of the Ingredients of Open-Endedness: Replicating Picbreeder with Large Vision-Language Models Conference: Genetic and Evolutionary Computation Conference; July 13–17, 2026; San Jose, Costa Rica Genetic and Evolutionary Computation Conference (GECCO ’26), July 13–17, 2026, San Jose, Costa Rica DOI: 10.1145/3795095.3805186 ISBN: 979-8-4007-2487-9/2026/07 CCS: Computing methodologies Multi-agent systems CCS: Computing methodologies Cognitive science by Sam Earle, Kai Arulkumaran, Andrew Dai, Akarsh Kumar, Julian Togelius, Sebastian Risi (2026) in the Tower of Babel, a free online library of public-domain works preserved in clean, searchable text. No sign-up. The full text is served as a plain-text file at a predictable URL, where text is present: https://antarctic-labs.com/tower-of-babel/arxiv-2605-23908.txt.
Description
Sam Earle and colleagues replay Picbreeder, the collaborative human experiment where users bred images by evolving small neural networks, with frontier vision language models in place of people. Clear qualitative gaps separate model output from the human baseline. The team measures those gaps through lineage complexity and visual and semantic novelty. Probing causes, they test added exploratory noise, diversity between agents, and narrative memory of past choices. They hunt the ingredients that let unguided search stay open ended and fruitful.
Catalog record
| CREATOR | Sam Earle, Kai Arulkumaran, Andrew Dai, Akarsh Kumar, Julian Togelius, Sebastian Risi |
|---|---|
| YEAR | 2026 |
| COLLECTION | PAPERS |
| FORMAT | TXT |
| FILE SIZE | 62035 |
| LICENSE | http://creativecommons.org/licenses/by/4.0/ |
| RIGHTS | CC BY |
| SOURCE | arXiv |
Source: https://arxiv.org/abs/2605.23908