Metamaterials that learn to change shape

Yao Du, Ryan van Mastrigt, Jonas Veenstra, Corentin Coulais · arXiv · 2025

This research presents metamaterials capable of learning complex shape changes, enabling advanced adaptive behaviors.

High AI ConfidenceGood SourceConceptEarly Research

Plain English summary

This study introduces metamaterials that can learn to change their shape in response to examples provided to them. Unlike traditional materials, these metamaterials can adapt their internal properties to forget and learn new shapes over time. This ability allows them to perform complex actions like gripping and locomotion.

Why this matters

The ability of materials to learn and adapt like living organisms could revolutionize fields such as robotics, where materials that can change shape on demand can lead to more versatile and efficient machines. This research highlights the potential for creating smarter materials that can respond to their environment in real-time.

Key findings

  • Metamaterials can learn complex shape-changing responses.
  • They can update internal stiffness to adapt to new shapes.
  • These materials can forget and learn new shape changes sequentially.
  • They can learn multistable shape changes for advanced actions.
  • This establishes metamaterials as a platform for physical learning.

What's new

The introduction of a learning mechanism in metamaterials that allows for adaptive shape changes, unlike traditional materials.

Limitations

The abstract does not specify the practical applications or limitations of the learning process in real-world scenarios.

Commercial context

The research is in the conceptual stage and does not indicate any commercial applications yet.

Publication

Publisher
arXiv
Publication date
January 21, 2025
Research type
Preprint
arXiv
2501.11958
Access
open

Tags

More on Mechanical Metamaterials

See all →
Mechanical Metamaterialspaper· Aug 1, 2026

Bio-inspired re-entrant honeycomb metamaterial with programmable dual-plateau mechanical response

A bio-inspired re-entrant honeycomb mechanical metamaterial is described as having a programmable dual-plateau mechanical response.

Xihai Ni, Xiaoyu Wang +4 · Elsevier BVUnknown
Mechanical Metamaterialspaper· Jul 30, 2026

Design of binary stiffness adaptive elements with tunable linear and nonlinear response

The study presents binary stiffness adaptive elements that can switch stiffness states, enhancing reprogrammable structures for unknown environments.

Catherine Catrambone, Kai Jun Chen +2 · SAGE PublicationsLaboratory Research
Mechanical Metamaterialspaper· Jun 30, 2026

A Review of Machine Learning Applications in Mechanical Metamaterial Design

The review summarizes how machine learning models and end-to-end workflows can accelerate mechanical metamaterial design and property prediction using simulation-based validation.

Galymzhan Turysbekov, Ulanbek Auyeskhan +3 · MDPI AGSimulation
Mechanical Metamaterialspaper· Jun 22, 2026

Span-Morphing Wing Using Multistable Honeycomb Metamaterial Structures

This study proposes a span-morphing wing using multistable honeycomb structures for improved aerodynamic performance and stability.

Ruixin Wang, Bin Niu · MDPI AGWorking Prototype
Mechanical Metamaterialsnews· Jun 8, 2026

Heat breaks the rules at the nanoscale and scientists used it to their advantage

Nanoscale gold metamaterials can boost heat transfer across tiny gaps by up to 4×, enabling more efficient thermal management and precision heat engineering.

· ScienceDaily — MaterialsUnknown
Mechanical Metamaterialspaper· Apr 3, 2026

Reprogrammable metamaterial robot with embodied versatile computation and mechanical intelligence

A reprogrammable metamaterial robot is presented as combining versatile computation with mechanical intelligence.

Wu Zhou, Yi-Ze Wang · Springer Science and Business Media LLCUnknown
Method note: Summaries and ratings on this page are generated by AI from the abstract only. Read the original paper for full context. · Model: gpt-4o-mini-2024-07-18