Levin, Risi et al. preprint: rank-one LoRA tweaks give neural cellular automata reusable morphogenetic handles
drmichaellevin · x · 2026-09-26
A new preprint by Benedikt Hartl, Michael Levin, Sebastian Risi and colleagues, On Growth and Form, and Function, revisits D'Arcy Thompson's 1917 classic to ask whether large-scale morphological transformations can be encoded as low-dimensional modulations of a self-organizing developmental system.
- Neural cellular automata (NCAs) serve as bio-inspired models of distributed development, growing target morphologies from a single cell via a shared local regulatory network
- LoRA is applied to pretrained NCAs, so each adapted developmental program becomes a low-rank modulation of a fixed regulatory scaffold
- Horizontal and vertical scaling of a grown 2D emoji phenotype each require only rank-one adaptations; linear combinations parametrically control size and generalize beyond the training distribution
- Strikingly, adaptations learned on one phenotype transfer zero-shot to structurally and semantically diverse phenotypes sharing the same scaffold, largely preserving internal features
- This suggests reusable system-level "hyper-directions" of scale rather than morphology-specific transformations; the study spans roughly 25,000 independently trained phenotypes
More from Research
- Distillation criticized: gradient descent can't even copy the teacher model — deliprao · 2026-09-26
- ResolVI models measurement errors to fix single-cell RNA data, cutting false signals from 14.8% to under 0.01% — bravo_abad · 2026-09-26
- Define Dimensions First: Hamel Husain's Method for Synthetic Eval Data — HamelHusain · 2026-09-26
- LeCun's lab borrows a brain trick to double AI goal-reaching success to 94% — alex_verem · 2026-09-26
- AgentOdyssey, a Text-Game Engine for Test-Time Continual Learning Agents, Accepted at NeurIPS — DanielKhashabi · 2026-09-26
- Boris Cherny's viral tweet puts TLA+ and formal methods in the agent-coding spotlight — fhuszar · 2026-09-26