Researchers have developed a noninvasive method to detect anemia using grayscale photos of the eye’s conjunctiva, taken with standard smartphones. By applying machine learning to spatial and textural features extracted from over 12,000 photos of 565 children aged 5 to 15, the study found strong associations between these features and anemia status. Unlike other approaches, this method does not rely on color analysis or specialized equipment, making it practical for use in low-resource settings. The findings suggest a scalable, affordable tool for anemia screening in children, especially in areas with limited access to laboratory testing.

A groundbreaking integrated encryption and communication (IEAC) framework, enabled by end-to-end deep learning, has achieved a record-breaking single-channel secure transmission rate of 1 Terabit per second (Tb/s) over a 1,200-km optical fibre link. Published in National Science Review, this innovation optimizes mutual information for legal users while minimizing leakage to eavesdroppers, offering a scalable solution for high-speed, secure data transmission in the era of big data and AI.

Recently, Professor Peng Xue's team from the Beijing Computational Science Research Center published an article titled "Quantum cooling engine fueled by quantum measurements" in Science Bulletin. The research team employed a linear optical platform to simulate a two-stroke, two-qubit engine. In the experiment, they demonstrated various quantum thermodynamic processes by tuning the energy level spacing of the working substance and adjusting the temperature parameters of the bath. They successfully realized a quantum cooling engine driven by quantum measurements and discovered the influence of entanglement on the energy exchange between the working substance and the measurement apparatus.

An Osaka Metropolitan University-led research team conducted a meta-analysis of the diagnostic capabilities of generative AI in the field of medicine using 83 research papers.