Category Technology/Electronics

Machine learning models called convolutional neural networks (CNNs) power technologies like image recognition and language translation. A quantum counterpart—known as a quantum convolutional neural network (QCNN)—could process information more efficiently by using quantum states instead of classical bits.

Photons are fast, stable, and easy to manipulate on chips, making photonic systems a promising platform for QCNNs. However, photonic circuits typically behave linearly, limiting the flexible operations that neural networks need.

Adaptive state injection in photonic QCNNs
In a study published in Advanced Photonic...

As artificial intelligence (AI) and smart gadgets become more common, our need for reliable power sources grows. Renewable energy options like solar and wind are great, but they depend on specific conditions. A research team led by Professor Dong-Myeong Shin from the Department of Mechanical Engineering at the University of Hong Kong (HKU) has developed a novel moisture-activated electricity generator (MEG) that offers a fresh, eco-friendly way to generate electricity—just from moisture in the air.

Their findings have been published in the journal Advanced Functional Materials in an article titled “Long-Lasting Moisture Energy Scavenging in Dry Ambient Air Empowered by a Salt Concentration-Gradient Cationi...

Researchers at Oregon State University have developed a quick-setting, environmentally friendly alternative to concrete they hope can one day be used to rapidly 3D print homes and infrastructure.

Also known as additive manufacturing, 3D printing is already being used to help solve construction challenges such as the global housing crisis that’s emerged as Earth’s population approaches 8.5 billion.

But cement, the binding agent in concrete, accounts for about 8% of the planet’s carbon dioxide emissions, and concrete’s curing time—which can be multiple days—and required structural supports can inhibit progress on construction projects.

How ...

Soon, researchers may be able to create movies of their favorite protein or virus better and faster than ever before. Researchers at the Department of Energy’s SLAC National Accelerator Laboratory have pioneered a new machine learning method—called X-RAI (X-Ray single particle imaging with Amortized Inference)—that can “look” at millions of X-ray laser-generated images and create a three-dimensional reconstruction of the target particle. The team recently reported their findings in Nature Communications.

X-RAI’s ability to sort through a massive number of images and learn as it goes could unlock limits in data-gathering, allowing researchers to see molecules up close—and perhaps even on the move...