Researchers contend that with the rapid development of machine learning, new materials can be engineered for various applications, from electric motors to carbon capture technologies. AI-generated paints could help mitigate the effects of urban heat islands and lower air conditioning costs.
Experts in materials science are harnessing artificial intelligence to create innovative coatings that can maintain building temperatures 5°C to 20°C cooler than conventional paint after exposure to direct sunlight. This technology is also applicable to vehicles, trains, electrical devices, and other entities that need enhanced cooling in a warming world.
Through machine learning, interdisciplinary teams from institutions in the US, China, Singapore, and Sweden have developed a new paint formulation optimized for reflecting sunlight and radiating heat, as evidenced by peer-reviewed research. Published in Science Journal Nature.
This represents the latest instance of AI circumventing traditional trial-and-error methods in the pursuit of scientific innovation. Last year, UK-based Matnex employed AI to design a new type of permanent magnet for electric vehicle motors, aiming to reduce reliance on carbon-heavy rare earth elements.
Microsoft has also released AI tools tailored for researchers to swiftly create novel inorganic materials—such as crystal structures commonly utilized in solar panels and medical implants. There’s optimism surrounding the potential for new materials to enhance carbon capture capabilities and improve battery efficiency.
Investigations into paint were conducted by scholars at the University of Texas at Austin, Shanghai Jiao Tong University, National University of Singapore, and Umeå University in Sweden. In scorching locations like Rio de Janeiro and Bangkok, researchers determined that applying one of the newly developed AI-enhanced paints to the roof of a four-story apartment building could conserve 15,800 kilowatt-hours of electricity annually. When this paint is used on 1,000 buildings, it saves enough energy to power over 10,000 air conditioning units each year.
“As a scientist at the University of Texas and a co-leader of this research,” said Yuebbing Zeng, “our machine learning framework signifies a significant advancement in thermal meta-emitter design. By automating processes and broadening the design landscape, we can generate materials with exceptional properties that were previously unfeasible.”
He mentioned that what previously took a month can now be achieved in days using AI to innovate new materials, including those that might not have been uncovered through traditional exploration methods.
“Now we follow the machine learning outputs; its directives can be executed without numerous design and manufacturing test cycles,” he added.
Dr. Alex Ganoce, a lecturer at Imperial College London, emphasized: “We are also leveraging machine learning to innovate new materials. Developments in this field are occurring rapidly. Over the last year, numerous startups have emerged aiming to utilize generative AI for materials creation.”
He noted that the journey to design new materials can involve assessing millions of potential combinations. AI empowers material scientists to overcome limitations associated with computing resources and allows them to specify desired characteristics to the AI upfront, thereby reversing the conventional method of material creation and trait testing.
Source: www.theguardian.com

