Neural Cellular Automata for ARC-AGI

Fuente: arXiv
Salvato in:
Dettagli Bibliografici
Autori principali: Xu, Kevin, Miikkulainen, Risto
Natura: Preprint
Pubblicazione: 2025
Soggetti:
Accesso online:
Tags: Aggiungi Tag
Nessun Tag, puoi essere il primo ad aggiungerne!!
_version_ 1866915647547506688
author Xu, Kevin
Miikkulainen, Risto
author_facet Xu, Kevin
Miikkulainen, Risto
contents Cellular automata and their differentiable counterparts, Neural Cellular Automata (NCA), are highly expressive and capable of surprisingly complex behaviors. This paper explores how NCAs perform when applied to tasks requiring precise transformations and few-shot generalization, using the Abstraction and Reasoning Corpus for Artificial General Intelligence (ARC-AGI) as a domain that challenges their capabilities in ways not previously explored. Specifically, this paper uses gradient-based training to learn iterative update rules that transform input grids into their outputs from the training examples and apply them to the test inputs. Results suggest that gradient-trained NCA models are a promising and efficient approach to a range of abstract grid-based tasks from ARC. Along with discussing the impacts of various design modifications and training constraints, this work examines the behavior and properties of NCAs applied to ARC to give insights for broader applications of self-organizing systems.
format Preprint
id arxiv_https___arxiv_org_abs_2506_15746
institution arXiv
publishDate 2025
record_format arxiv
spellingShingle Neural Cellular Automata for ARC-AGI
Xu, Kevin
Miikkulainen, Risto
Neural and Evolutionary Computing
Artificial Intelligence
Cellular automata and their differentiable counterparts, Neural Cellular Automata (NCA), are highly expressive and capable of surprisingly complex behaviors. This paper explores how NCAs perform when applied to tasks requiring precise transformations and few-shot generalization, using the Abstraction and Reasoning Corpus for Artificial General Intelligence (ARC-AGI) as a domain that challenges their capabilities in ways not previously explored. Specifically, this paper uses gradient-based training to learn iterative update rules that transform input grids into their outputs from the training examples and apply them to the test inputs. Results suggest that gradient-trained NCA models are a promising and efficient approach to a range of abstract grid-based tasks from ARC. Along with discussing the impacts of various design modifications and training constraints, this work examines the behavior and properties of NCAs applied to ARC to give insights for broader applications of self-organizing systems.
title Neural Cellular Automata for ARC-AGI
topic Neural and Evolutionary Computing
Artificial Intelligence
url https://arxiv.org/abs/2506.15746