Amid recent advances for multivariate time series forecasting, self-supervised learning has emerged as a promising paradigm for deriving transferable knowledge from multi-domain data. Despite its effectiveness, existing approaches exhibit two critical limitations: (1) Underestimating the significance of multivariate dependencies in learning generalizable representations and (2) Failing to reconcile the complementary strengths of autoregressive and one-shot generative paradigms. In this work, we propose TimeCAP, a novel channel-aware pre-training framework that internalizes latent causal relationships among variables inherent in multi-domain data, and effectively transfers the acquired knowledge to downstream applications. Technically, we present a flexible channel-grouping learning approach, complemented by an adaptive meta-routing mechanism, enabling TimeCAP to parallel recognize intra-group local patterns while maintaining global coherence. Intra- and inter-group multivariate dependencies are captured through the self- and cross-attention with channel-aware mask, which strictly confine interactions among time-aligned, fine-grained multivariate tokens. To seamlessly unify two advanced generative paradigms, we propose a novel dynamic dual-head decoding and optimization strategy, empowering TimeCAP to leverage critical dependencies in the output series while avoiding cumulative errors over time. In the few-shot evaluation, TimeCAP achieves average MSE and MAE reductions of 11.8% and 6% over leading baselines, while also outperforming state-of-the-art models in full-shot and zero-shot settings by large margins.

MicroRNAs (miRNAs) play a critical role in modulating drug responses, highlighting the importance of accurately identifying miRNA-drug associations. However, existing computational methods often suffer from limited representation capability and high computational complexity, frequently relying solely on either global or local feature extraction. To address these limitations, we propose MHTMG, a novel predictive framework that integrates a Multi-Head Attention mechanism with a Parameterized Transformation Matrix (MHA-PTM) and GraphSAGE. The MHA-PTM component introduces a learnable transformation matrix that acts as a proxy token to refine the query, key, and value representations, thereby replacing traditional pairwise comparisons with a two-stage attention mechanism. This design substantially reduces computational overhead while preserving the model's capacity to capture global dependencies. Concurrently, a GraphSAGE encoder is utilized to extract local topological features from the miRNA-drug bipartite graph. By iteratively aggregating information from neighboring nodes, GraphSAGE enhances graph representations using the graph's intrinsic structural information. To fully exploit both global semantic features and local structural cues, the embeddings produced by MHA-PTM and GraphSAGE are concatenated and subsequently fed into a multi-layer perceptron (MLP) for final association prediction. In addition, the superior performance of MHTMG in representative case studies highlights its practical utility and reliability for elucidating miRNA-drug interactions.

Time series forecasting is a pivotal task across diverse fields such as finance, energy, and meteorology, providing critical insights for informed decision-making and resource management. The growing complexity and scale of these tasks demand high-performance computing capabilities, which are crucial for efficiently processing vast amounts of temporal data and achieving real-time forecasting. Despite considerable progress, existing deep learning models face challenges in capturing complex temporal dependencies, lack explicit mechanisms to enhance temporal learning, and are encumbered by the computational overhead and limited adaptability of layer normalization in Transformer architectures. To address these limitations, we propose a novel GPDT model that integrates an auto-regressive pretraining framework, promoting temporal consistency and enabling the effective capture of both shortand long-term temporal dependencies. This pretraining strategy establishes a robust temporal inductive bias, which is further refined through task-specific fine-tuning to optimize forecasting performance. Additionally, the static layer normalization in the standard Transformer encoder is replaced with an adaptive dynamic tanh mechanism, which not only reduces computational costs but also improves the model's adaptability to evolving temporal patterns. Moreover, GPDT further leverages instance normalization, channel independence, and patch embedding to enhance both efficiency and accuracy. Comprehensive evaluation across a range of forecasting scenarios validates the model's superior performance and strong generalization capability.

MicroRNAs (miRNAs) have significant implications in diverse human diseases and have proven to be effectively targeted by small molecules (SMs) for therapeutic interventions. However, current SM-miRNA association prediction models do not adequately capture SM/miRNA similarity. Matrix completion is an effective method for association prediction, but existing models use nuclear norm instead of rank function, which has some drawbacks. Therefore, we proposed a new approach for predicting SM-miRNA associations by utilizing the truncated schatten p-norm (TSPN). First, the SM/miRNA similarity was preprocessed by incorporating the Gaussian interaction profile kernel similarity method. This identified more SM/miRNA similarities and significantly improved the SM-miRNA prediction accuracy. Next, we constructed a heterogeneous SM-miRNA network by combining biological information from three matrices and represented the network with its adjacency matrix. Finally, we constructed the prediction model by minimizing the truncated schatten p-norm of this adjacency matrix and we developed an efficient iterative algorithmic framework to solve the model. In this framework, we also used a weighted singular value shrinkage algorithm to avoid the problem of excessive singular value shrinkage. The truncated schatten p-norm approximates the rank function more closely than the nuclear norm, so the predictions are more accurate. We performed four different cross-validation experiments on two separate datasets, and TSPN outperformed various most advanced methods. In addition, public literature confirms a large number of predictive associations of TSPN in four case studies. Therefore, TSPN is a reliable model for SM-miRNA association prediction.

Dysregulation of microRNAs (miRNAs) is closely associated with refractory human diseases, and the identification of potential associations between small molecule (SM) drugs and miRNAs can provide valuable insights for clinical treatment. Existing computational techniques for inferring potential associations suffer from limitations in terms of accuracy and efficiency. To address these challenges, we devise a novel predictive model called RPCA$\Gamma $NR, in which we propose a new Robust principal component analysis (PCA) framework based on $\gamma $-norm and $l_{2,1}$-norm regularization and design an Augmented Lagrange Multiplier method to optimize it, thereby deriving the association scores. The Gaussian Interaction Profile Kernel Similarity is calculated to capture the similarity information of SMs and miRNAs in known associations. Through extensive evaluation, including Cross Validation Experiments, Independent Validation Experiment, Efficiency Analysis, Ablation Experiment, Matrix Sparsity Analysis, and Case Studies, RPCA$\Gamma $NR outperforms state-of-the-art models concerning accuracy, efficiency and robustness. In conclusion, RPCA$\Gamma $NR can significantly streamline the process of determining SM-miRNA associations, thus contributing to advancements in drug development and disease treatment.