# Towards spatio-temporal extrapolation of phase-field simulations with convolution-only neural networks

Source: [Physics : nature.com subject feeds](https://www.nature.com/articles/s41529-026-00906-5) · 2026-11-12 · Physics
Page: https://themorning.science/story/towards-spatio-temporal-extrapolation-of-phase-field-simulations-with-convolutio-aba3fcc91f6f82b0
Kind: research · Impact: low

## In brief

- A paper titled "Towards spatio-temporal extrapolation of phase-field simulations with convolution-only neural networks" appears in nature.com subject feeds dated 2026-11-12.
- The work concerns using convolution-only neural networks to extrapolate phase-field simulations in space and time.
- No summary is provided in the source, so results and methods beyond the title are not reported here.

## Why it matters

Phase-field simulations are computationally expensive, so a neural-network method that extrapolates them could matter to researchers who rely on such models, though the source does not give details of its performance.

## Coverage

- [Towards spatio-temporal extrapolation of phase-field simulations with convolution-only neural networks](https://www.nature.com/articles/s41529-026-00906-5) — Physics : nature.com subject feeds, 2026-11-12

Tags: phase-field, neural-networks, surrogate-modelling, simulation, materials-science

---

Brief generated by AI (claude-haiku-5-5) from the reporting listed under Coverage. Headlines belong to their publishers. The Morning Science – https://themorning.science