Meta-Learning an Evolvable Developmental Encoding
Fuente:
arXiv
Gespeichert in:
| Hauptverfasser: | , , , , |
|---|---|
| Format: | Preprint |
| Veröffentlicht: |
2024
|
| Schlagworte: | |
| Online-Zugang: | |
| Tags: |
Tag hinzufügen
Keine Tags, Fügen Sie den ersten Tag hinzu!
|
| _version_ | 1866913417168683008 |
|---|---|
| author | Montero, Milton L. Plantec, Erwan Nisioti, Eleni Pedersen, Joachim W. Risi, Sebastian |
| author_facet | Montero, Milton L. Plantec, Erwan Nisioti, Eleni Pedersen, Joachim W. Risi, Sebastian |
| contents | Representations for black-box optimisation methods (such as evolutionary algorithms) are traditionally constructed using a delicate manual process. This is in contrast to the representation that maps DNAs to phenotypes in biological organisms, which is at the hear of biological complexity and evolvability. Additionally, the core of this process is fundamentally the same across nearly all forms of life, reflecting their shared evolutionary origin. Generative models have shown promise in being learnable representations for black-box optimisation but they are not per se designed to be easily searchable. Here we present a system that can meta-learn such representation by directly optimising for a representation's ability to generate quality-diversity. In more detail, we show our meta-learning approach can find one Neural Cellular Automata, in which cells can attend to different parts of a "DNA" string genome during development, enabling it to grow different solvable 2D maze structures. We show that the evolved genotype-to-phenotype mappings become more and more evolvable, not only resulting in a faster search but also increasing the quality and diversity of grown artefacts. |
| format | Preprint |
| id |
arxiv_https___arxiv_org_abs_2406_09020 |
| institution | arXiv |
| publishDate | 2024 |
| record_format | arxiv |
| spellingShingle | Meta-Learning an Evolvable Developmental Encoding Montero, Milton L. Plantec, Erwan Nisioti, Eleni Pedersen, Joachim W. Risi, Sebastian Neural and Evolutionary Computing Representations for black-box optimisation methods (such as evolutionary algorithms) are traditionally constructed using a delicate manual process. This is in contrast to the representation that maps DNAs to phenotypes in biological organisms, which is at the hear of biological complexity and evolvability. Additionally, the core of this process is fundamentally the same across nearly all forms of life, reflecting their shared evolutionary origin. Generative models have shown promise in being learnable representations for black-box optimisation but they are not per se designed to be easily searchable. Here we present a system that can meta-learn such representation by directly optimising for a representation's ability to generate quality-diversity. In more detail, we show our meta-learning approach can find one Neural Cellular Automata, in which cells can attend to different parts of a "DNA" string genome during development, enabling it to grow different solvable 2D maze structures. We show that the evolved genotype-to-phenotype mappings become more and more evolvable, not only resulting in a faster search but also increasing the quality and diversity of grown artefacts. |
| title | Meta-Learning an Evolvable Developmental Encoding |
| topic | Neural and Evolutionary Computing |
| url | https://arxiv.org/abs/2406.09020 |