Search

Vertical federated learning is designed to protect user privacy by building local models over disparate datasets and transferring intermediate parameters without directly revealing the underlying data. However, the intermediate parameters uploaded by participants may memorize information about the training data. With the recent legislation on the"right to be forgotten", it is crucial for vertical federated learning systems to have the ability to forget or remove previous training information of any client. For the first time, this work fills in this research gap by proposing a vertical federated unlearning method on logistic regression model. The proposed method is achieved by imposing constraints on intermediate parameters during the training process and then subtracting target client updates from the global model. The proposed method boasts the advantages that it does not need any new clients for training and requires only one extra round of updates to recover the performance of the previous model. Moreover, data-poisoning attacks are introduced to evaluate the effectiveness of the unlearning process. The effectiveness of the method is demonstrated through experiments conducted on four benchmark datasets. Compared to the conventional unlearning by retraining from scratch, the proposed unlearning method has a negligible decrease in accuracy but can improve training efficiency by over 400%. [ABSTRACT FROM AUTHOR]

Introducing the exploration of stimulated CD4+ cells adenosine triphosphate (sATPCD4) levels for immune monitoring post non-small cell lung cancer (NSCLC) chemotherapy, the present study aimed to investigate its efficacy in gauging the potential risk of disease progression (PD) in patients with NSCLC. Therefore, a total of 89 patients with advanced NSCLC, who underwent chemotherapy between August 15 2022 and August 30 2023 at the Fifth Affiliated Hospital of Guangzhou Medical University (Guangzhou, China), were retrospectively studied. Patients were divided into the PD (n=21) and disease stability (non-PD; n=68) groups and their clinical data were compared. The thresholds for predicting PD were identified using receiver operating characteristics (ROC) curves. Multivariate logistic regression analysis was carried out to assess the association between peripheral blood markers and the incidence of PD. Therefore, post-chemotherapy, significant differences in white blood cell count, non-stimulated CD4+ cells ATP and sATPCD4 levels were obtained between patients in the PD and non-PD groups (P<0.05). In addition, sATPCD4 levels were notably decreased in the PD group compared with the non-PD group. Furthermore, ROC analysis revealed that the predictive threshold for PD was 224.5 ng/ml [area under the curve=0.887; 95% confidence interval, 0.811–0.963]. Additionally, patients with low immunity (ATP <224.5 ng/ml) exhibited a higher risk of PD compared with the high-immunity group (ATP >224.5 ng/ml; P<0.0001). Finally, multivariate logistic regression analysis suggested that sATPCD4 could serve as an independent factor for predicting NSCLC progression. Overall, the current study predicted that immune function could be possibly associated with the risk of PD in patients with NSCLC. [ABSTRACT FROM AUTHOR]

One of the key materials of interest for p-type oxide semiconductor thin film electronics is cuprous oxide (Cu2O), due to its relatively high hole mobility. In this work, we use experiments, analytical models, and density functional theory calculations to study the scattering mechanisms that determine Hall mobility in two Cu2O samples. First, we examine a polycrystalline Cu2O thin film deposited by RF magnetron sputtering, and second, a single-crystalline Cu2O bulk substrate. Temperature-dependent Hall measurements indicate that neutral impurity and grain boundary scattering are dominant for the polycrystalline Cu2O thin film, while phonon scattering is dominant for single-crystalline Cu2O. Our first-principles calculations show that the room-temperature intrinsic hole mobility of Cu2O is 106 cm2 V−1 s−1, indicating the great promise of the material for p-type electronic devices. This intrinsic mobility is limited by phonon scattering, with the most dominant scattering modes having phonon energies of 88.4 and 17.1 meV. These results indicate that the key pathways to increase the hole mobility in Cu2O thin films are by reducing the impurity concentration and by increasing grain size. Our work thus sets the stage for the future development of high performance Cu2O-based p-type thin film transistors. [ABSTRACT FROM AUTHOR]

Motivation Cryo-Electron Tomography (cryo-ET) is a 3D imaging technology that enables the visualization of subcellular structures in situ at near-atomic resolution. Cellular cryo-ET images help in resolving the structures of macromolecules and determining their spatial relationship in a single cell, which has broad significance in cell and structural biology. Subtomogram classification and recognition constitute a primary step in the systematic recovery of these macromolecular structures. Supervised deep learning methods have been proven to be highly accurate and efficient for subtomogram classification, but suffer from limited applicability due to scarcity of annotated data. While generating simulated data for training supervised models is a potential solution, a sizeable difference in the image intensity distribution in generated data as compared with real experimental data will cause the trained models to perform poorly in predicting classes on real subtomograms. Results In this work, we present Cryo-Shift , a fully unsupervised domain adaptation and randomization framework for deep learning-based cross-domain subtomogram classification. We use unsupervised multi-adversarial domain adaption to reduce the domain shift between features of simulated and experimental data. We develop a network-driven domain randomization procedure with 'warp' modules to alter the simulated data and help the classifier generalize better on experimental data. We do not use any labeled experimental data to train our model, whereas some of the existing alternative approaches require labeled experimental samples for cross-domain classification. Nevertheless, Cryo-Shift outperforms the existing alternative approaches in cross-domain subtomogram classification in extensive evaluation studies demonstrated herein using both simulated and experimental data. Availabilityand implementation https://github.com/xulabs/aitom. Supplementary information Supplementary data are available at Bioinformatics online. [ABSTRACT FROM AUTHOR]