On Jan. 1, 2024, a 7.5-magnitude earthquake struck the Noto Peninsula in Japan, resulting in extensive damage in the region caused by uplift, when the land rises due to shifting tectonic plates. The observed uplift, however, varied significantly, with some areas experiencing as much as a 5-meter rise of the ground surface. To better understand how the characteristics of the affected fault lines impact earthquake dynamics, researchers in Japan used recently developed simulations to make a detailed model of the fault. The findings could help develop models to simulate scenarios of different earthquakes and mitigate disasters in the future.

Annealing processors, used for solving combinatorial optimization problems based on fully coupled Ising models, are limited in capacity and precision. While previous studies have developed scalable systems that allow scaling of the number of spins or capacity, their precision or interaction bit width remains fixed. In a new study, researchers have developed a novel dual scalable annealing processing system that allows both expansion in the number of spins and interaction bit width.

Multipath interference disrupts wireless signals, causing issues like TV ghosting and fading. Now, researchers from Japan have developed a passive metasurface that overcomes traditional filtering limits. Using a time-varying interlocking mechanism with field-effect transistors, it transmits the first signal while blocking delayed ones from other angles—without power or processing. This innovation enables low-cost, reliable wireless communication, which is ideal for IoT applications and environments prone to interference.

Sleep is essential for memory consolidation, but could it also prepare the brain for future learning? Researchers from Japan investigated this dual role, using advanced imaging to track neuronal activity in mice. They identified a distinct population of brain cells that became active during post-learning sleep and later encoded new experiences. Their findings, supported by neural network modeling, reveal that sleep not only preserves past memories but also primes the brain for forming future memories.

Researchers from AIMR developed an AI-driven approach to model carbon growth on metal surfaces with high accuracy. By integrating molecular dynamics, time-stamped force-biased Monte Carlo simulations, and machine learning, their method replicates key processes like carbon diffusion and graphene nucleation, enabling scalable, efficient carbon nanomaterial production for electronics and energy storage.

A Japanese research team has, for the first time, captured real-time images of how supramolecular gels form using high-speed atomic force microscopy, revealing a previously unseen mechanism of fiber growth at the nanoscale.