Graph neural networks (GNNs) have garnered significant attention for their ability to effectively process graph-related data. Most existing methods assume that the input graph is noise-free; however, this assumption is frequently violated in real-world scenarios, resulting in impaired graph representations. To address this issue, we start from the essence of graph structure learning, considering edge discovery and removal, reweighting of existing edges, and differentiability of the graph structure. We introduce virtual nodes and consider connections with virtual nodes to generate optimized graph structures, and subsequently utilize Gumbel-Softmax to reweight edges and achieve differentiability of the Graph Structure Learning (VN-GSL for abbreviation). We conducted a thorough evaluation of our method on a range of benchmark datasets under both clean and adversarial circumstances. The results of our experiments demonstrate that our approach exhibits superiority in terms of both performance and efficiency. Our implementation will be made available to the public.

Temperature/pH dual-responsive hydrogels are a class of smart materials capable of undergoing reversible structural or functional changes in response to temperature and pH stimuli. Owing to their remarkable dual-stimuli-responsive characteristics, these hydrogels have demonstrated significant potential in various biomedical applications, including drug delivery, tissue engineering, and diagnostics technologies, making them a prominent research focus. Although considerable progress has been made in recent years, a systematic summary of the preparation methods, structural design strategies and complex biomedical applications of these materials remains conspicuously absent. Consequently, this review aims to comprehensively examine the latest advancements in this field. First, the primary preparation methods of temperature/pH dual-responsive hydrogels, including chemical crosslinking, physical crosslinking, and hybrid crosslinking, are introduced and compared. Subsequently, the main structural design strategies, including microsphere, core–shell and layered structures, and their corresponding fabrication processes are systematically elucidated. Finally, the recent progress of temperature/pH dual-responsive hydrogels in biomedical applications is discussed, including drug delivery, cancer therapy, biosensing and diagnosis, tissue engineering and regenerative medicine, as well as wound healing. Based on the current research progress, this review also outlines the major challenges in the development of temperature/pH dual-responsive hydrogels, and presents perspectives on future research directions.

Integrating deep learning and density functional theory to design boron–carbon compound-supported diatomic catalysts and elucidate their efficient hydrogen evolution reaction mechanism.

Owing to their small size, micron-scale accuracy, rapid response, and biocompatibility, microrobots are emerging as a promising tool for biomedical applications, especially in targeted drug delivery and minimally invasive microsurgery. The architectural configuration of microrobots governs their locomotion performance, environmental adaptability, and functional integration capacity. Although existing reviews have systematically organized this field by actuation strategies, material categories, or application scenarios, a comprehensive summary centered on the structural evolution paradigm remains conspicuously absent. This review aims to fill this gap by systematically tracing the evolutionary trajectory of microrobot structures from rigid architectures through soft configurations to rigid–soft integrated systems. First, the foundational principles underlying structural evolution were introduced, encompassing the connotation of structure, fluid dynamics constraints, and the drivers of structural innovation. Then, rigid microrobot architectures were systematically elucidated, including geometric asymmetric and surface asymmetric designs. Subsequently, soft microrobot structures, covering both predefined deformation structures and dynamically reconfigurable architectures, were discussed. Finally, the rigid–soft integrated systems reconciling compliance and performance through spatial heterogeneity or temporal stiffness modulation were systematically surveyed. This evolutionary trend reflects a transition from optimizing individual performance parameters toward achieving balanced functional synergy across distinct task phases. Based on the current research progress, this review also presents future research directions in data-driven structural optimization, reconfigurable architectures, and autonomous structural intelligence, offering strategic guidance for next-generation microrobot design.

The 4 × 4.5 m radio telescope array at Guizhou Normal University is an astronomical observation facility in operation, mainly aiming at the scientific detection of pulsars and fast radio bursts. To adequately address the observational requirements of this telescope array, we developed an observation scheduling system. This system is able to predict and plot the elevation change curves of the observed targets in relation to the telescope array during the whole day. In addition, for multiple targets, it provides intelligent scheduling and processing according to the observable time. The system also offers a comprehensive database of calibrators for the flux calibration of the targets. Moreover, it can generate pre-configured array $uv$ coverage maps, which assist in determining the optimal configuration of the array. This system has been operating during the daily observations of the $4\times 4.5$ m radio telescope array and has successfully detected two typical pulsars. It has also been tested for applicability in target-observation prediction in other radio telescope arrays.