Researchers at the University of Amsterdam’s Institute of Physics and AMOLF have developed metamaterials capable of learning complex shape-changing responses. Using a contrastive learning scheme, a technique borrowed from machine learning, the materials adapt their mechanical behaviour based on applied deformations. The research was led by physicist Corentin Coulais and co-authored by Ryan van Mastrigt, who holds affiliations with both institutes.
Living organisms change shape as a fundamental strategy for adaptation — from individual cells contracting to animals moving through their environment. Human-made materials can mimic shape morphing, but until now lacked any ability to learn. The research team has closed that gap. By embedding a contrastive learning mechanism into a metamaterial’s elastic skeleton, the material can evolve toward new mechanical equilibria in response to external inputs. In practical terms: the material can be taught a desired shape-change response and retains that learned behaviour.
Metamaterials that can learn are relevant far beyond the physics lab. Potential applications include soft robotics, adaptive prosthetics, reconfigurable structures, and smart mechanical interfaces — domains where materials need to respond intelligently to their environment without external computing. The crossover between mechanical engineering and machine learning principles demonstrated here reflects a broader trend in deep-tech research: fundamental science at the frontier of multiple disciplines, with long-term impact across industries.
For more information, read the full article on the website of the Institute of Physics
If you are interested in Amsterdam Science Park, exploring opportunities, or simply have a question, feel free to get in touch. We’ll be happy to help.
Subscribe to our LinkedIn-newsletter and we will keep you updated on all that our park has to offer. It will arrive to your inbox five times a year and you can unsubscribe easily at any time.
Subscribe