August 2025

Scientists Apply Stephen Hawking’s Theory to Propose Detectable ‘Black Hole Morsels’ in Space

A new study suggests “black hole morsels” — tiny, asteroid-sized black holes from cosmic mergers — could emit detectable bursts of Hawking radiation. Observatories like HESS, HAWC, and Fermi may already hold clues. Detecting them could unlock insights into quantum gravity, unknown particles, and even hidden dimensions beyond the Standard Model.

Read More »

China Advances Guowang Internet Constellation with Latest Satellite Launch

China has launched the eighth batch of satellites for its Guowang internet constellation, lifting off on Aug. 13 aboard a Long March 5B rocket from Wenchang Space Launch Center. Operated by state-owned China SatNet, Guowang aims to deploy about 13,000 satellites in low Earth orbit to compete with SpaceX’s Starlink. Each launch so far has carried only eight to ten relatively large satellites. The mission marks the fourth Guowang launch in less than three weeks, underscoring China’s push to accelerate global broadband coverage and strengthen its position in the satellite-internet market.

China Advances Guowang Internet Constellation with Latest Satellite Launch Read More »

ESA’s Mars Express Discovers Deep Valleys and Frozen Features Hinting at Mars’ Icy Past

In July 2025, the European Space Agency’s Mars Express orbiter captured a high-resolution image of Acheron Fossae, a region marked by deep chasms and ridges on Mars’s surface. These features, created by ancient crustal stretching, split the terrain into raised horsts and sunken grabens. Valley floors reveal smooth surfaces carved by slow-moving, ice-rich rock glaciers, forming rounded knobs and mesas. Scientists believe these structures date back 3.7 billion years, during Mars’s most geologically active era. The presence of rock glaciers hints at ancient ice ages, suggesting the Red Planet once had climatic cycles capable of supporting frozen water flow over vast periods, reshaping its landscape.

ESA’s Mars Express Discovers Deep Valleys and Frozen Features Hinting at Mars’ Icy Past Read More »

New Physics-Based Model Sheds Light on How Deep Neural Networks Learn Features

A Physical Review Letters study likens deep neural network feature learning to spring-block mechanics, linking data simplification to spring extension and nonlinearity to friction. The model reveals how noise can balance separation across layers and help predict performance, offering a powerful tool to optimise training, improve generalisation, and enhance efficiency in large AI systems.

New Physics-Based Model Sheds Light on How Deep Neural Networks Learn Features Read More »