We gratefully acknowledge support from
the Simons Foundation and member institutions.
Full-text links:

Download:

Current browse context:

cs.NE

Change to browse by:

References & Citations

DBLP - CS Bibliography

Bookmark

(what is this?)
CiteULike logo BibSonomy logo Mendeley logo del.icio.us logo Digg logo Reddit logo ScienceWISE logo

Computer Science > Neural and Evolutionary Computing

Title: Growing Isotropic Neural Cellular Automata

Abstract: Modeling the ability of multicellular organisms to build and maintain their bodies through local interactions between individual cells (morphogenesis) is a long-standing challenge of developmental biology. Recently, the Neural Cellular Automata (NCA) model was proposed as a way to find local system rules that produce a desired global behaviour, such as growing and persisting a predefined pattern, by repeatedly applying the same rule over a grid starting from a single cell. In this work we argue that the original Growing NCA model has an important limitation: anisotropy of the learned update rule. This implies the presence of an external factor that orients the cells in a particular direction. In other words, 'physical' rules of the underlying system are not invariant to rotation, thus prohibiting the existence of differently oriented instances of the target pattern on the same grid. We propose a modified Isotropic NCA model that does not have this limitation. We demonstrate that cell systems can be trained to grow accurate asymmetrical patterns through either of two methods: by breaking symmetries using structured seeds; or by introducing a rotation-reflection invariant training objective and relying on symmetry breaking caused by asynchronous cell updates.
Subjects: Neural and Evolutionary Computing (cs.NE); Machine Learning (cs.LG); Cell Behavior (q-bio.CB)
Cite as: arXiv:2205.01681 [cs.NE]
  (or arXiv:2205.01681v1 [cs.NE] for this version)

Submission history

From: Craig Fouts [view email]
[v1] Tue, 3 May 2022 11:34:22 GMT (3920kb,D)
[v2] Wed, 25 May 2022 13:10:19 GMT (3924kb,D)
[v3] Wed, 1 Jun 2022 15:16:16 GMT (3924kb,D)

Link back to: arXiv, form interface, contact.