Yantai Nanshan University

Hypergraph Neural Networks (HGNNs) have become an important tool for processing complex structured data due to their ability to model higher-order associative relationships. However, the inherent adversarial vulnerabilities of HGNNs may raise serious security risks. The associated risks are far more pronounced in strong target attacks, which are highly targeted and demand the accurate misclassification of source-class nodes into predefined target classes. Current research on attacks against HGNNs mostly focuses on untargeted attacks or common target attacks, and lacks attacks that precisely control the attack class. Therefore, the research related to strong target attacks is still in an undeveloped state. To fill this research gap, this paper proposes a Strong Target Attack framework for HGNNs based on Label poisoning and Structure modification (STALS). The framework first uses feature similarity and hypergraph structure adaptability to select the optimal target class. Subsequently, the nodes are label-poisoned under the label change budget constraint. A gradient-guided greedy hyperedge reconstruction strategy is used to optimize the association relationship between poisoned nodes and hyperedges within the structure modification budget, maximize the propagation efficiency of mislabeled information, and achieve stable directed misclassification from source class nodes to target classes. We conducted extensive experiments on four mainstream datasets, and the experimental results show that STALS achieves excellent attack performance and significantly outperforms existing baseline methods in terms of success classification rate.

Cu-Cr-Zr alloy is a typical precipitation-strengthened alloy, and the interface stability between precipitates and the Cu matrix significantly influences the alloy’s strength and properties. However, research in this field is currently lacking. In light of this, we focus on the CuZr2 precipitate and investigate its surface and interface properties using the GGA/PBE method within density functional theory. The results indicate that the (100) and (010) surfaces of CuZr2 share the same atomic structure, with both being stoichiometric surfaces. By fitting the relationship between the total energy of surface supercells with varying numbers of atoms and the number of atomic layers, the surface energy values were accurately calculated. The (100) surface is a non-stoichiometric surface, featuring three surface terminations: Cu, Zr1, and Zr2, with the Zr2 termination being the most stable. Finally, based on experimental observations, the atomic structure of the CuZr2 (010)/Cu (110) interface was predicted. The calculated interfacial energy reveals that the lattice mismatch between the CuZr2 precipitate and the Cu matrix significantly affects interfacial stability.

The cancer-immunity cycle represents a self-amplifying cascade that is essential for generating durable antitumor immune responses. However, disruption of any individual step can render the entire cycle ineffective, thereby limiting the efficiency of cancer immunotherapy. Nucleic acid-based nanodrugs (NANDs) elicit unique interactions with both cancerous and immune cells, showing great potential in reactivating or amplifying the cancer-immunity cycle. This review first provides a brief introduction to the cancer-immunity cycle and NANDs and then outlines the major types of nonviral nanocarriers that have been developed to construct NANDs so far. Rational design strategies for smart NANDs are summarized. The recent advances in the precise modulation of each critical step within the cancer-immunity cycle by NANDs for enhanced cancer immunotherapy are highlighted. The current challenges faced by these NANDs are also briefly discussed.

The development of efficient and durable electrochemical sensor electrode materials is essential for real-time analysis. Graphitic carbon nitride (g-C3N4) nanostructures have emerged as a new generation of sensing platforms for electrochemical detection of hazardous pollutants, owing to their abundant functional amino groups, tunable nanostructures, high density of active sites and superior physicochemical properties. In this study, sulfur-doped graphitic carbon nitride (SCN) was synthesized via thermal polycondensation, followed by the fabrication of bismuth nanospheres (BiNSs) and bismuth tungstate (BWO) through a solvothermal method. BiNSs exhibit excellent electrical conductivity and strong affinity toward Pb2+ ions, while BWO possesses a unique heterojunction-regulating ability and oxygen vacancy structure. Both components synergistically interact with SCN to construct a composite system with complementary functionalities. The SCN/BiNSs/BWO nanocomposite was employed as a signal probe for Pb2+ detection. Under optimized conditions, namely pH = 5 with a deposition potential of -0.9 V and a deposition time of 180 s in (sodium acetate-acetic acid) NaAc-HAc buffer, the sensor demonstrated excellent performance, yielding a detection limit of 0.04 μM and a linear range of 0.1-4.5 μM. The SCN/BiNSs/BWO-modified glassy carbon electrode (GCE) demonstrated excellent stability. Recovery rates for Pb2+ detection in real water samples ranged from 94.7 to 109.0%, highlighting the significant practical application potential of the proposed electrode material.

This paper takes the art design major as the research object, explores how generative artificial intelligence empowers the outcome-based Education (OBE) model, constructs the theoretical framework of “AI-Enabled OBE”, and conducts empirical analysis through 586 valid questionnaires. Research findings show that the creative assistance effectiveness of AI tools, students' willingness to explore, and their recognition of compound abilities significantly promote the achievement of learning goals. Art design students pay more attention to the matching degree between tools and creative styles rather than the quantity of tools. Privacy concerns are positively correlated with the need for personalized feedback, which is reflected as a rational trade-off of “creative privacy”. Research reveals that AI in art and design education is not only a technical tool but also a creative collaborator and an expander of thinking. Finally, suggestions for teaching reform are put forward, emphasizing that attention should be paid to the cultivation of creative exploration spirit, the integration of aesthetics and technology, the construction of personalized feedback mechanisms, and data ethics education.

Controlling two-dimensional (2D) valleytronics is challenging for information technology. This study shows that a ferroelectric-assisted layer can effectively enable non-volatile control of 2D valleytronics. Using first-principles simulations, we find that different polarization states in the Sc2CO2 layer cause the RuBrF monolayer to transition from a semiconductor to a half-metal, while also changing magnetic anisotropy from in-plane to out-of-plane. In the −P state, the system behaves as a ferromagnetic semiconductor with a spontaneous valley polarization of 329 meV. In the +P state, it becomes a ferromagnetic half-metal, blocking valleytronics. This enables electro-reversible control of valley electrons in the RuBrF/Sc2CO2 heterostructure. We explain the modulation of magnetic anisotropy and valley polarization using second-order perturbation theory and the k⋅p model. Our work offers a promising approach for non-volatile valleytronic control at the nanoscale, aiding the design of new devices.

Abstract This study presents the successful synthesis of a series of stable iminium diradical(oid)s. CNR1 + OTf − , CNR2 + OTf − , and CNR3 + OTf − can be readily prepared by treating their corresponding methoxy precursors with triflate acid. Remarkably, they all display excellent stability under light and ambient conditions. CNR1 + exhibits a closed shell electronic feature, whereas CNR2 + features an open shell singlet ground state. CNR3 + OTf − possess triplet ground electronic state, as evidenced by ESR and superconducting quantum interference device (SQUID) measurements. Through meticulous fitting of the SQUID data, a singlet‐triplet energy gaps (∆ E S‐T ) of +1.04 kcal mol −1 along with intermolecular antiferromagnetic interactions ( θ ) of ‐28 K were obtained. In comparison to CNR1 + , both the open shell CNR2 + and CNR3 + display distinctively red shifted long wavelength absorption bands that extend up to approximately 2400 nm. Such long wavelength absorption characteristics are quite scarce among organic materials. This research offers a novel approach for the design and synthesis of iminium diradical(oid)s, which feature tunable ground states and exhibit intense absorption in the second near infrared (NIR‐II) region, opening up new possibilities in the field of organic materials research.