Search

Federated learning （FL） is a distributed machine learning technology that enables joint construction of machine learning models by transmitting intermediate results （e.g., model parameters, parameter gradients, embedding representation, etc.） applied to data distributed across various institutions.FL reduces the risk of privacy leakage, since raw data is not allowed to leave the institution.According to the difference in data distribution between institutions, FL is usually divided into horizontal federated learning （HFL）, vertical federated learning （VFL）, and federal transfer learning （TFL）.VFL is suitable for scenarios where institutions have the same sample space but different feature spaces and is widely used in fields such as medical diagnosis, financial and security of VFL.Although VFL performs well in real-world applications, it still faces many privacy and security challenges.To the best of our knowledge, no comprehensive survey has been conducted on privacy and security methods.The existing VFL was analyzed from four perspectives: the basic framework, communication mechanism, alignment mechanism, and label processing mechanism.Then the privacy and security risks faced by VFL and the related defense methods were introduced and analyzed.Additionally, the common data sets and indicators suitable for VFL and platform framework were presented.Considering the existing challenges and problems, the future direction and development trend of VFL were outlined, to provide a reference for the theoretical research of building an efficient, robust and safe VFL.

The closure section of precast concrete segmental bridges (PCSBs) usually adopts the wet joint of cast‐in‐situ concrete, the width is generally 10–20 cm, and the regular longitudinal reinforcement is discontinuous, which makes the section the vulnerable part of the bridge. To explore the mechanical property of the wet joint in the closure section of PCSB under cyclic load, six wet joint specimens are tested to investigate the influence of loading mode, interface treatment mode, setting of regular reinforcement, and prestress level. The results show that the failure modes of wet joint specimens have similar characteristics, careful roughening of joint surfaces can improve the bonding effect of the new and old concrete and delay the cracking of the specimen, the cracking load of the specimen increases by 10.3% after careful roughening. When the prestress level increases from 6 to 8 MPa and 10 MPa, the stiffness of the specimens increases obviously, and the fracture development of the specimen in the loading process is improved. The cracking load increases by 21.9% and 32.8%, the shear strength increases by 9.9% and 6.6%, respectively, but the ultimate displacement decreases by 5.6% and 11.4%, respectively. Then, the finite element model of the wet joint was established and verified, the influence of shear key, concrete strength, and wet joint width are selected for parameter analysis. It is found that stress concentration is easy to occur after shear key is set, which has adverse effects on structural flexural resistance. The shear capacity of the model decreases by 22.4% and 21.5%, respectively after the setting of single and double‐shear keys, but the stiffness of the wet joint model before cracking is improved. Compared with C50 concrete model, the ultimate bearing capacity of C60 and C70 concrete models is increased by 11.5% and 24.8%, respectively. Wet joint width has little influence on the mechanical properties of wet joints in the closure section. The research results provide references for design and construction of PCSB wet joints in practical engineering. [ABSTRACT FROM AUTHOR]

This study investigates the effects of low Reynolds number (Re) conditions on the cooling performance of fan-shaped film cooling holes in the gas turbine engines of high-altitude long-endurance Unmanned Aerial Vehicles (UAVs). The significance of film cooling technology becomes paramount due to the low Re operational environment of UAVs. A novel Very Large Eddy Simulation (VLES) method was employed to systematically analyze a range of mainstream Reynolds numbers (ReD) from 480 to 32 000 and blowing ratios (M) from 1.0 to 2.5 while keeping the density ratio (DR) constant at 1.75. Compared to other turbulence models, the VLES model showed the highest similarity in predicting thermal field distributions and the most accurate prediction of film cooling performance at low Re. The results reveal that at low ReD, significant flow concentration within the hole leads to reduced velocity uniformity at the outlet, resulting in poor film coverage downstream. Additionally, the coolant jet tends to detach from the wall, adversely affecting cooling performance. In contrast, higher ReD exhibited improved coolant jet adhesion to the wall and cooling efficiency, attributed to the reduced intensity of counter-rotating vortex pair and decreased hot gas entrainment, thereby enhancing cooling effectiveness. Notably, for the same ReD values, cases with M = 1.5 consistently showed better cooling effectiveness compared to other M conditions. This study provides new insight into the cooling performance of fan-shaped film cooling holes under low Reynolds number conditions, which is crucial for enhancing the cooling efficiency of gas turbines in high-altitude UAVs. [ABSTRACT FROM AUTHOR]