Cancer is a complex and heterogeneous disease, where even patients with the same cancer type may respond differently to treatment regimens. Predicting the therapeutic effects of drugs on cancer based on cancer characteristics is a critical aspect of precision oncology. Currently, most anticancer drug response(CDR) prediction methods rely on extracting features from the cell line-drug bipartite composition. However, these methods often fail to adequately capture the features of both drugs and cell lines, ignoring the homogeneous features of cell lines and drugs and their correlation with deep heterogeneous features. To address these challenges, we propose a novel prediction framework that leverages a heterogeneous and homogeneous hybrid graph neural network named H$^{3}$CDR. H$^{3}$CDR learns the similarity features of cancer cell lines and drugs by fusing their multi-omics data. Additionally, a multi-branch network is employed to extract features from both cell lines and drugs, enabling the identification of potential features. Extensive experiments on the GDSC and CCLE databases demonstrate the superiority of our model. Evaluated by five-fold cross-validation, H$^{3}$CDR achieves an area under the ROC curve (AUC) of 0.8772 and an area under the precision-recall curve (AUPRC) of 0.8819 on the GDSC dataset.

Palm-vein has received widespread attention for reliable biometric recognition due to its robust resistance to replicate and forge. However, the rise of adversarial attacks poses a high risk of vulnerability for palm-vein recognition, leaving most existing methods vulnerable to small and human-imperceptible adversarial perturbations. In this paper, we propose a palm-vein image reconstruction network for palm-vein image protection, which mainly consists of palm-vein-specific exploration, feature refinement, and image reconstruction sub-networks. Specifically, we first specially learn the noise-insensitive palm-vein-specific feature by decoupling non-vein noise information via cascaded noise-injected and Canny-based convolution layers, and then refine palm-vein-specific features via multiple stacked basic convolution and transposed convolution pairs. Lastly, we convert the fine-grained palm-vein features into the latent sharp palm-vein images via two transposed convolution layers. Moreover, we develop both identity-aware and visual-aware loss functions to ensure the high-quality of the reconstructed palm-vein images. Experimental results on the widely used PolyU palm-vein dataset demonstrate the promising effectiveness of the proposed palm-vein image reconstruction network.