The beam focusing characteristic of frequency diverse array (FDA) in direction-distance dimension has been widely applied in covert communication. To reduce the covert performance loss caused by time-varying FDA channels, the instantaneous optimization of antenna beamforming vector (ABV) and antenna frequency vector (AFV) is required for FDA. However, due to the difficulty in acquiring instantaneous channel state information (CSI) and the limitations of transmit device, instantaneous optimization scheme (IOS) may not be feasible in practical communication scenarios. Moreover, the existing robust schemes only adopt some simple and fixed AFVs without optimizing AFV. This paper investigates robust optimization scheme (ROS) in FDA-aided covert system, where the optimization frequency of the ABV and AFV at the transmitter can be adaptively adjusted based on the acquired CSIs. We maximize the average covert rate during the transmission by jointly optimizing ABV and AFV, and modify the conventional block successive upper-bound minimization (BSUM) method to solve the optimization problem. Furthermore, both perfect CSI and partial CSI at the warden are considered. Numerical results demonstrate that our proposed ROS outperforms the existing robust schemes, especially in the cases of highly correlated channels and partial CSI. Although ROS is indeed inferior to IOS, enhancing the channel decoupling capability of FDA and deteriorating the detection capability of the warden can effectively narrow the covert performance gap between ROS and IOS. Moreover, the beampattern generated by the optimal ABV and AFV at each time slot ensures that the beamfoming gain is maximized at the legitimate receiver and minimized at the warden simultaneously.

MOTIVATION: Understanding the genomic foundation of human diversity and disease requires models that effectively capture sequence variation, such as single nucleotide polymorphisms (SNPs). While recent genomic foundation models have scaled to larger datasets and multi-species inputs, they often fail to account for the sparsity and redundancy inherent in human population data, such as those in the 1000 Genomes Project. SNPs are rare in humans, and current masked language models (MLMs) trained directly on whole-genome sequences may struggle to efficiently learn these variations. Additionally, training on the entire dataset without prioritizing regions of genetic variation results in inefficiencies and negligible gains in performance. RESULTS: We present MutBERT, a probabilistic genome-based masked language model that efficiently utilizes SNP information from population-scale genomic data. By representing the entire genome as a probabilistic distribution over observed allele frequencies, MutBERT focuses on informative genomic variations while maintaining computational efficiency. We evaluated MutBERT against DNABERT-2, various versions of Nucleotide Transformer, and modified versions of MutBERT across multiple downstream prediction tasks. MutBERT consistently ranked as one of the top-performing models, demonstrating that this novel representation strategy enables better utilization of biobank-scale genomic data in building pretrained genomic foundation models. AVAILABILITY AND IMPLEMENTATION: https://github.com/ai4nucleome/mutBERT.

BACKGROUND: Inferring Gene Regulatory Networks (GRNs) from gene expression data is a pivotal challenge in systems biology. Most existing methods fail to consider the skewed degree distribution of genes, complicating the application of directed graph embedding methods. RESULTS: The Cross-Attention Complex Dual Graph Embedding Model (XATGRN) was proposed to address this issue. It employs a cross-attention mechanism and a dual complex graph embedding approach to manage the skewed degree distribution, ensuring precise prediction of regulatory relationships and their directionality. The model consistently outperforms existing state-of-the-art methods across various datasets. CONCLUSIONS: XATGRN provides an effective solution for inferring GRNs with skewed degree distribution, enhancing the understanding of complex gene regulatory mechanisms. The codes and detailed requirements have been released on Github: ( https://github.com/kikixiong/XATGRN ).

Decompilation is a widely used process for reverse engineers to significantly enhance code readability by lifting assembly code to a higher-level C-like language, pseudo-code. Nevertheless, the process of compilation and stripping irreversibly discards high-level semantic information that is crucial to code comprehension, such as comments, identifier names, and types. Existing approaches typically recover only one type of information, making them suboptimal for semantic inference. In this paper, we treat pseudo-code as a special programming language, then present a unified pre-trained model, HexT5, that is trained on vast amounts of natural language comments, source identifiers, and pseudo-code using novel pseudo-code-based pre-training objectives. We fine-tune HexT5 on various downstream tasks, including code summarization, variable name recovery, function name recovery, and similarity detection. Comprehensive experiments show that HexT5 achieves state-of-the-art performance on four downstream tasks, and it demonstrates the robust effectiveness and generalizability of HexT5 for binary-related tasks.