Ultrawide bandgap semiconductor optoelectronic synapses can perform high-parallel computing with low false alarm rate, making them ideal for building deep-ultraviolet (DUV) neuromorphic visual system (NVS). Scientists in China developed a breakthrough DUV optoelectronic synapse using Ga₂O₃-based cascade heterojunctions. The device mimics biological vision with 4,096 conductance states and ultrahigh linearity, enabling high-accuracy fingerprint recognition and fault-tolerant logic operations. This innovation paves the way for energy-efficient neuromorphic vision systems in military, environmental monitoring, and secure communications.

Flexible electronics face critical challenges in achieving monolithic three-dimensional (3D) integration, including material compatibility, structural stability, and scalable fabrication methods. Inspired by the tactile sensing mechanism of the human skin, we have developed a flexible monolithic 3D-integrated tactile sensing system based on a holey MXene paste, where each vertical one-body unit simultaneously functions as a microsupercapacitor and pressure sensor. The in-plane mesopores of MXene significantly improve ion accessibility, mitigate the self-stacking of nanosheets, and allow the holey MXene to multifunctionally act as a sensing material, an active electrode, and a conductive interconnect, thus drastically reducing the interface mismatch and enhancing the mechanical robustness. Furthermore, we fabricate a large-scale device using a blade-coating and stamping method, which demonstrates excellent mechanical flexibility, low-power consumption, rapid response, and stable long-term operation. As a proof-of-concept application, we integrate our sensing array into a smart access control system, leveraging deep learning to accurately identify users based on their unique pressing behaviors. This study provides a promising approach for designing highly integrated, intelligent, and flexible electronic systems for advanced human–computer interactions and personalized electronics.

As the number of wireless applications and devices grows, higher standards for the quality of service and navigation performance of mobile networks are required. Numerous critical applications, including unmanned aerial vehicles, internet of things, digital twin, and military systems, require reliable communication and accurate navigation services. To meet these requirements, the development of the sixth generation (6G) network is necessary. 6G networks provide seamless 3-dimensional coverage in space–air–ground–sea area, as well as deep coupling of communication, sensing, and computation. 6G networks will be combined with low Earth orbit (LEO) satellites to construct a universal and intelligent integrated system of communication, sensing, and computing by utilizing the benefits of LEO satellites, such as miniaturization, modularity, large bandwidth, low latency, and wide area coverage. One of the critical construction tasks in this system is the integration of communication and navigation (ICAN), which can break the limitations of the global navigation satellite system, provide high-precision, robust navigation capability, and enable high- quality communication services.

Space manipulators play an important role in the on-orbit services and planetary surface operation whose reliability is a key issue. In the extreme environment of space, space manipulators are susceptible to a variety of unknown disturbances. Since it is difficult for the manipulator to be repaired immediately, once it fails, it will mean that it cannot complete the mission as expected, which may cause serious losses and dangers. How to have a resilient guarantee, i.e., the manipulator’s ability to resiliently recover and continue to complete tasks, in failure or disturbance is the core capability of its future development. The motion planning unit is used as the computing terminal of the manipulator’s joint motion trajectory. Compared with traditional motion planning, learning-based motion planning has gradually become a hot spot in current research. However, no matter what kind of research ideas, the single robotic manipulator is studied as an independent agent, making it unable to provide sufficient flexibility under conditions such as external force disturbance, observation noise, and mechanical failure.

This study created a highly accurate machine learning model to predict acute exacerbations of COPD (AECOPD) using data from 878 patients. By integrating biochemical, demographic, and pulmonary function parameters with stepwise Cox regression and random survival forest algorithms, the model outperformed traditional methods, demonstrating excellent predictive performance and stability.

University of Utah atmospheric scientists combine machine learning and manually gathered snowfall data from across the West to create forecast platform that outperforms existing weather prediction models.