Search for: All records

With the dramatic growth of data in both amount and scale, distributed machine learning has become an important tool for the massive data to finish the tasks as prediction, classification, etc. However, due to the practical physical constraints and the potential privacy leakage of data, it is infeasible to aggregate raw data from all data owners or the learning purpose. To tackle this problem, the distributed privacy-preserving learning approaches are introduced to learn over all distributed data without exposing the real information. However, existing approaches have limits on the complicated distributed system. On the one hand, traditional privacy-preserving learning approaches rely on heavy cryptographic primitives on training data, in which the learning speed is dramatically slowed down due to the computation overheads. On the other hand, the complicated system architecture becomes a barrier in the practical distributed system. In this paper, we propose an efficient privacy-preserving machine learning scheme for hierarchical distributed systems. We modify and improve the collaborative learning algorithm. The proposed scheme not only reduces the overhead for the learning process but also provides the comprehensive protection for each layer of the hierarchical distributed system. In addition, based on the analysis of the collaborative convergency in different learning groups, we also propose an asynchronous strategy to further improve the learning efficiency of hierarchical distributed system. At the last, extensive experiments on real-world data are implemented to evaluate the privacy, efficacy, and efficiency of our proposed schemes. 

As a promising alternative to the mainstream CoFeB/MgO system with interfacial perpendicular magnetic anisotropy (PMA),L1₀‐FePd and its synthetic antiferromagnet (SAF) structure with large crystalline PMA can support spintronic devices with sufficient thermal stability at sub‐5 nm sizes. However, the compatibility requirement of preparingL1₀‐FePd thin films on Si/SiO₂wafers is still unmet. In this paper, high‐qualityL1₀‐FePd and its SAF on Si/SiO₂wafers are prepared by coating the amorphous SiO₂surface with an MgO(001) seed layer. The preparedL1₀‐FePd single layer and SAF stack are highly (001)‐textured, showing strong PMA, low damping, and sizeable interlayer exchange coupling, respectively. Systematic characterizations, including advanced X‐ray diffraction measurement and atomic resolution‐scanning transmission electron microscopy, are conducted to explain the outstanding performance ofL1₀‐FePd layers. A fully‐epitaxial growth that starts from MgO seed layer, induces the (001) texture ofL1₀‐FePd, and extends through the SAF spacer is observed. This study makes the vision of scalable spintronics more practical.