It's one thing to write about magical or science-fiction concepts like an invisibility cloak, such as the one used by Harry Potter and his wizard friends. It's another thing entirely to design one in the real world.
Creating metamaterials, or artificial materials engineered to have properties not found in naturally occurring materials (like invisibility), is hard work. That's why a team at the South Korea-based Pohang University of Science and Technology (POSTECH) want an AI to handle the challenge.
This is an exact science. It requires "designing artificial atoms that are smaller than the wavelength of light and controlling the polarization and spin of light," the team says in its press statement. Through the right polarization, the POSTECH team believes it could produce "new optical properties are made that are not found in nature."
But creating these exactly correct atoms is time consuming and expensive. That's why an AI could handle the problem better, the scientists say. An AI could be trained with a "vast amount of data and it can learn designs of various metamaterials and the correlation between photonic structures and their optical properties."
Using that data, the AI could run countless simulations on building metamaterials until something promising emerged.
Professors Junsuk Rho, Sunae So, and Jungho Mun all worked on developing an AI to show their concept could work. They built the AI to design arbitrary photonic structures, but the crucial aspect of their work came when they gave the AI an additional level of freedom of design by categorizing types of materials and adding them as a design factors. That extra freedom allowed the AI to design specifically thinking about optical features, like visibility.
"Our research was successful in bringing it to a higher degree of freedom of the design, but the new design still requires users to input certain problem settings in the beginning. It sometimes produced wrong designs and therefore make it impossible to produce desired metamaterials," Rho says in the press statement.
That's common for new AIs, which often learn through repeated action. Those actions usually have to be repeated many, many times. That's why large companies like Google tend to crowdsource their machine learning efforts. And even then, sometimes it's still not enough.
"So, I'd like to take our findings a step further by developing a complete design method of metamaterials utilizing AI," Rho says. "Also, I'd like to make innovative and practical metamaterials by training AI with reviews of the design constructed in consideration of final products."
They're ambitious steps for an ambitious project. Others have tried to create invisible material before. In 2015, a team at Penn State made a material that was undetectable, but not yet invisible.
But hey, if this brings us even one iota closer to exploring Hogwarts without being seen, we'll take it.
You Might Also Like