Estimates show that roughly 70% of energy generated around the globe is wasted. Experts believe that more efficient systems to generate power could be designed if researchers were able to forecast how heat passes through insulators and semiconductors. This is hard to do at the moment, however, because it is challenging to model the thermal properties of different materials.
The issue stems from subatomic particles that carry heat, known as phonons. A group of MIT researchers have designed a new framework based on machine-learning, which can forecast phonon dispersion relation about 1,000 times faster than other methods based on artificial intelligence (“AI”).
This technique may assist engineers in designing systems to generate energy that efficiently generate more power. It could also aid in the development of efficient microelectronics.
Mingda Li, an associate professor of engineering and nuclear science, states that obtaining the properties of phonons is challenging, either experimentally or computationally. The researchers’ approach calculates the phonon dispersion relation for a material, with the researchers using graph neural networks to convert the atomic structure of a material into a graph made up of multiple nodes linked by edges.
While graph neural networks can help calculate quantities such as electrical polarization or magnetization, they aren’t flexible in a way that allows them to forecast phonon dispersion relation. To acquire flexibility, the researchers came up with a virtual node graph neural network that allows the neural network’s output to vary. This allowed them to estimate phonon dispersion relations, with the researchers observing that the virtual node graph neural network provided more accuracy when forecasting the heat capacity of a material.
The researchers believe that this efficiency may enable the creation of a bigger space when looking for materials with specific thermal properties such as superconductivity, or thermal storage. Additionally, they posit that their technique could also be utilized in forecasting challenging magnetic and optical properties.
For the future, the scientists are focused on refining this method so virtual nodes can record any small change that may impact the structure of phonons with additional sensitivity.
Li, the senior author of this study, was joined in doing this research by MIT’s Thomas Siebel; professor of electrical engineering and computer science Tommi Jaakkola; chemistry grad student Ryotaro Okabe; and MIT electrical engineering and computer science graduate student Abhijatmedhi Chotrattanapituk, among others.
Their findings were reported in “Nature Computational Science.”
This study was supported by the Oak Ridge National Laboratory, the Harvard Quantum Initiative, a Sow-Hsin Chen Fellowship, a Mathworks Fellowship, the National Science Foundation and the U.S. Department of Energy.
It is anticipated that during the coming years, tech companies such as Intel Corp. (NASDAQ: INTC) are likely to bring many innovative AI solutions onto the market that could revolutionize the different industries in which they are used.
About AINewsWire
AINewsWire (“AINW”) is a specialized communications platform with a focus on the latest advancements in artificial intelligence (“AI”), including the technologies, trends and trailblazers driving innovation forward. It is one of 60+ brands within the Dynamic Brand Portfolio @ IBN that delivers: (1) access to a vast network of wire solutions via InvestorWire to efficiently and effectively reach a myriad of target markets, demographics and diverse industries; (2) article and editorial syndication to 5,000+ outlets; (3) enhanced press release enhancement to ensure maximum impact; (4) social media distribution via IBN to millions of social media followers; and (5) a full array of tailored corporate communications solutions. With broad reach and a seasoned team of contributing journalists and writers, AINW is uniquely positioned to best serve private and public companies that want to reach a wide audience of investors, influencers, consumers, journalists, and the general public. By cutting through the overload of information in today’s market, AINW brings its clients unparalleled recognition and brand awareness.
AINW is where breaking news, insightful content and actionable information converge.
To receive SMS alerts from AINewsWire, text “AI” to 888-902-4192 (U.S. Mobile Phones Only)
For more information, please visit www.AINewsWire.com
Please see full terms of use and disclaimers on the AINewsWire website applicable to all content provided by AINW, wherever published or re-published: https://www.AINewsWire.com/Disclaimer
AINewsWire
Los Angeles, CA
www.AINewsWire.com
310.299.1717 Office
Editor@AINewsWire.com
AINewsWire is powered by IBN