Multimodal intent detection leverages complementary information from diverse sensors to achieve precise semantic understanding; however, existing methodologies predominantly operate under the ideal assumption of modality completeness. In practical deployments, missing modalities, stemming from sensor failure, background noise, or privacy constraints, are inevitable and uncertain, leading to severe performance degradation. To address this challenge, we propose SADF, a Structure Aware Distillation Framework designed to facilitate robust cross-level knowledge transfer from a full-modality teacher to a partial-modality student. At the semantic level, semantic topology distillation aligns prototype similarity distributions between the teacher and the student, capturing class topology and hierarchical relations. At the discriminative level, boundary aware distillation decouples predictions into target and nontarget classes, enforcing alignment with the teacher to enhance discriminability and suppress noise. By integrating these strategies, SADF enables the student to learn robust representations and decisions when uncertain modalities are missing. The results of experiments based on two benchmarks demonstrate that SADF consistently outperforms strong baselines.

Compared with monotherapy, combination drug therapy plays a crucial role in clinical treatment. However, the exponential expansion of the drug combination space has rendered traditional exploration methods for synergistic drug combinations inadequate. Recently, numerous efficient and accurate computational approaches have been developed to predict anticancer drug synergy, particularly those leveraging hypergraphs to model the multifaceted relationships between drug combinations and cell lines, which have demonstrated remarkable potential. Nevertheless, existing hypergraph-based methods fail to account for the heterogeneity of anticancer synergy hypergraphs and overlook the underlying similarities among drugs and cell lines, thereby limiting their ability to fully capture the complex interactions between drug combinations and cell lines. To address these limitations, we propose an Adaptive Heterogeneous Hypergraph Representation Learning Method (HHGSynergy) for predicting anticancer drug synergy, enabling more precise identification of synergistic drug combinations. Specifically, our framework first constructs drug/cell line similarity-based synergy hypergraphs based on the foundational anticancer synergy hypergraph, thereby establishing a comprehensive heterogeneous hypergraph. Next, a node importance calculation module is employed to learn both local and global importance weights of nodes, effectively capturing the structural characteristics of the hypergraph. Finally, a type-specific multi-head attention mechanism is utilized to iteratively update node embeddings, adaptively learning the significance of heterogeneous hyperedges. Experimental results demonstrate that HHGSynergy achieves state-of-the-art performance in both classification and regression tasks across diverse experimental scenarios, outperforming existing leading models. Case studies further underscore its potential for discovering novel synergistic drug combinations.

Circular RNA (circRNA) is a widely distributed class of non-coding RNA molecules that have been shown to play a significant role in cancer development and drug resistance, significantly influencing cellular sensitivity to therapeutic drugs and treatment outcomes. However, traditional biomedical experimental methods are limited by low efficiency and high costs when verifying the association between circular RNA and drug sensitivity. Therefore, developing an efficient and accurate computational method to predict new associations between circRNA and drug sensitivity has become an urgent need in current research. To address this, this study proposes HECLCDA, a novel method based on heterogeneous cross-scale contrastive learning. To construct a comprehensive initial information base for drugs and circRNAs, circRNA gene sequence similarity, drug structural inclusion similarity (SIS), and Gaussian kernel similarity were integrated.Based on the integrated and complete known information of circRNAs and drugs, a heterogeneous graph was built. The model used the Heterogeneous Graph Transformer to extract heterogeneous network topological information, effectively distinguishing the heterogeneity of nodes and edges. The model broke through the information relationship between node attributes and network topology at two scales, and innovatively introduced a cross-scale contrastive learning mechanism in a sparse labeling scenario. Using self-supervised signals, we aimed to enhance the discriminative power of node embeddings and maximize the mutual information between paired nodes at different scales. Cross-validation experiments demonstrated that HECLCDA performs excellently on real data and can efficiently predict drug sensitivity. Additionally, case studies further validate the model's effectiveness in predicting potential circRNA-drug sensitivity associations.

Accurate medical image segmentation is a pivotal task in medical image analysis, serving as the foundation for extracting critical clinical information and directly influencing the precision of disease identification, diagnosis, and treatment decision-making. Although Convolutional Neural Networks (CNNs) and Transformers have demonstrated remarkable performance in medical image segmentation, their sensitivity to variations in target size and morphology remains insufficient, resulting in limited capability of single-scale features to comprehensively capture multi-target details. Current approaches also face bottlenecks, including restricted global context modeling capabilities, high computational complexity, and inefficient multilevel feature fusion. To address these challenges in medical image segmentation, this study proposes an innovative architecture named GroupTransUNet. The model implements dual enhancements based on the UNet framework: (i) A novel Next-Generation Transformer Block (NGTB) is designed for the bottleneck layer, which organically integrates Efficient Multi-Head Self-Attention (E-MHSA) with Convolution-Enhanced Multi-Layer Perceptron (MLP) through a complementary mechanism to simultaneously enhance global semantic understanding and local detail characterization. (ii) The Grouped Feature Fusion Module (GFFM) is introduced in skip connections, employing grouped convolutions and multi-dilation rate strategies to construct multi-scale contextual receptive fields while preserving detailed integrity, thereby significantly improving feature fusion efficiency. Experimental results on the Synapse and ACDC datasets validate the effectiveness of our proposed GroupTransUNet. The source code can be obtained at https://anonymous.4open.science/r/GroupTransUNetA3D6.

Accurate prediction of drug side effect frequencies is critical for drug safety evaluation and clinical decision-making. Current methods primarily emphasize the associations between drugs and side effects, yet they often neglect the underlying structural and semantic features of both, which limits further advancements in prediction accuracy. In this study, we propose a novel multi-view fusion framework, MVFDSP, which integrates pre-trained molecular representation of 1D and 2D views with graph-based side effect information for side effect frequency prediction. Firstly, we obtain both 1D and 2D molecular representations from the pretrained molecular language model, and combine them using an adaptive fusion strategy. Subsequently, we construct a similarity network based on the side effect frequency matrix using K-Nearest Neighbors (KNN), and incorporate semantic embeddings derived from the terminology system of MedDRA to construct a side effect information graph. A multi-head graph attention network is then employed to capture the multi-dimensional information within this graph, allowing the model to attend to diverse aspects of the semantic and structural relationships among side effects. The final frequency prediction matrix is derived from the inner product between the learned drug and side effect embeddings. Experimental results on the SIDER 4.1 dataset demonstrate that MVFDSP outperforms existing methods, highlighting its effectiveness in capturing complex relationships of drugs and side effects. The code and data are available at https://github.com/Sonder-Echo/MVFDSP.

Survival prediction serves as a pivotal component in precision oncology, enabling the optimization of treatment strategies through mortality risk assessment. While the integration of histopathological images and genomic profiles offers enhanced potential for patient stratification, existing methodologies are constrained by two fundamental limitations: (i) insufficient attention to fine-grained local features in favor of global representations, and (ii) suboptimal cross-modal fusion strategies that either neglect intrinsic correlations or discard modality-specific information. To address these challenges, we propose DSCASurv, a novel cross-modal fusion alignment framework designed to explore and integrate intrinsic correlations across multimodal data, thereby improving the accuracy of survival prediction. Specifically, DSCASurv leverages the local feature extraction capabilities of convolutional layers and the long-range dependency modeling of scanning state space models to extract intra-modal representations, while generating cross-modal representations through dual parallel mixer architectures. A cross-modal attention module functions as a bridge for inter-modal information exchange and complementary information transfer. The framework ultimately integrates all intra-modal representations to generate survival predictions by enhancing and recalibrating complementary information. Extensive experiments on five benchmark cancer datasets demonstrate the superior performance of our approach compared to existing methods.

Predicting potential drug-disease associations (RDAs) plays a pivotal role in elucidating therapeutic strategies for diseases and facilitating drug repositioning, making it of paramount importance. However, existing methods are constrained and rely heavily on limited domain-specific knowledge, impeding their ability to effectively predict candidate associations between drugs and diseases. Moreover, the simplistic definition of unknown information pertaining to drug-disease relationships as negative samples presents inherent limitations. To overcome these challenges, we introduce a novel hierarchical negative sampling-based graph contrastive model, termed HSGCLRDA, which aims to forecast latent associations between drugs and diseases. In this study, HSGCLRDA integrates the association information as well as similarity between drugs, diseases and proteins. Meanwhile, the model constructs a drug-disease-protein heterogeneous network. Subsequently, employing a hierarchical structural sampling technique, we establish reliable negative drug-disease samples utilizing PageRank algorithms. Utilizing meta-path aggregation within the heterogeneous network, we derive low-dimensional representations for drugs and diseases, thereby constructing global and local feature graphs that capture their interactions comprehensively. To obtain representation information, we adopt a self-supervised graph contrastive approach that leverages graph convolutional networks (GCNs) and second-order GCNs to extract feature graph information. Furthermore, we integrate a contrastive cost function derived from the cross-entropy cost function, facilitating holistic model optimization. Experimental results obtained from benchmark datasets not only showcase the superior performance of HSGCLRDA compared to various baseline methods in predicting RDAs but also emphasize its practical utility in identifying novel potential diseases associated with existing drugs through meticulous case studies.

Studies have shown that IncRNA-miRNA interactions can affect cellular expression at the level of gene molecules through a variety of regulatory mechanisms and have important effects on the biological activities of living organisms. Several biomolecular network-based approaches have been proposed to accelerate the identification of lncRNA-miRNA interactions. However, most of the methods cannot fully utilize the structural and topological information of the lncRNA-miRNA interaction network. In this article, we proposed a new method, ISLMI, a prediction model based on information injection and second order graph convolution network(SOGCN). The model calculated the sequence similarity and Gaussian interaction profile kernel similarity between lncRNA and miRNA, fused them to enhance the intrinsic interaction between the nodes, using SOGCN to learn second-order representations of similarity matrix information. At the same time, multiple feature representations obtain using different graph embedding methods were also injected into the second-order graph representation. Finally, matrix complementation was used to increase the model accuracy. The model combined the advantages of different methods and achieved reliable performance in 5-fold cross-validation, significantly improved the performance of predicting lncRNA-miRNA interactions. In addition, our model successfully confirmed the superiority of ISLMI by comparing it with several other model algorithm.

MOTIVATION: Accumulating evidences have indicated that microRNA (miRNA) plays a crucial role in the pathogenesis and progression of various complex diseases. Inferring disease-associated miRNAs is significant to explore the etiology, diagnosis and treatment of human diseases. As the biological experiments are time-consuming and labor-intensive, developing effective computational methods has become indispensable to identify associations between miRNAs and diseases. RESULTS: We present an Ensemble learning framework with Resampling method for MiRNA-Disease Association (ERMDA) prediction to discover potential disease-related miRNAs. Firstly, the resampling strategy is proposed for building multiple different balanced training subsets to address the challenge of sample imbalance within the database. Then, ERMDA extracts miRNA and disease feature representations by integrating miRNA-miRNA similarities, disease-disease similarities and experimentally verified miRNA-disease association information. Next, the feature selection approach is applied to reduce the redundant information and increase the diversity among these subsets. Lastly, ERMDA constructs an individual learner on each subset to yield primitive outcomes, and the soft voting method is introduced for making the final decision based on the prediction results of individual learners. A series of experimental results demonstrates that ERMDA outperforms other state-of-the-art methods on both balanced and unbalanced testing sets. Besides, case studies conducted on the three human diseases further confirm the ERMDA's prediction capability for identifying potential disease-related miRNAs. In conclusion, these experimental results demonstrate that our method can serve as an effective and reliable tool for researchers to explore the regulatory role of miRNAs in complex diseases.

Long non-coding RNA(lncRNA) can interact with microRNA(miRNA) and play an important role in inhibiting or activating the expression of target genes and the occurrence and development of tumors. Accumulating studies focus on the prediction of miRNA-lncRNA interaction, and mostly are concerned with biological experiments and machine learning methods. These methods are found with long cycles, high costs, and requiring over much human intervention. In this paper, a data-driven hierarchical deep learning framework was proposed, which was composed of a capsule network, an independent recurrent neural network with attention mechanism and bi-directional long short-term memory network. This framework combines the advantages of different networks, uses multiple sequence-derived features of the original sequence and features of secondary structure to mine the dependency between features, and devotes to obtain better results. In the experiment, five-fold cross-validation was used to evaluate the performance of the model, and the zea mays data set was compared with the different model to obtain better classification effect. In addition, sorghum, brachypodium distachyon and bryophyte data sets were used to test the model, and the accuracy reached 0.9850, 0.9859 and 0.9777, respectively, which verified the model's good generalization ability.

RNA binding protein (RBP) is extensively involved in various cellular regulatory processes through the interaction with RNAs. Capturing the RBP binding preferences is fundamental for revealing the pathogenesis of complex diseases. Many experimental detection techniques are still time-consuming and labor-intensive, therefore, it is indispensable to develop a computational method with convincing accuracy. In this study, we proposed a CNN-BLSTM hybrid deep learning framework, named DeepDW, for predicting the RBP binding sites on RNAs with high-order encoding features of RNA sequence and secondary structure. The high-order encoding strategy was used to characterize the dependencies among adjacency nucleotides. For CNN-BLSTM hybrid model, DeepDW first employed two 1-D convolutional neural networks (CNNs) for learning the local features from high-order encoded matrices of RNA sequence and structure separately, and then applied two bidirectional long short-term memory networks (BLSTMs) to capture the global information in a higher level. Moreover, a series of experiments were carried out on 31 public datasets to evaluate our proposed framework, and DeepDW achieved superior performance than the state-of-the-art methods. The results indicated that the combination of high-order encoding method and CNN-BLSTM hybrid model had advantages in identifying RBP-RNA binding sites.

Recent studies have found that lncRNA (long non-coding RNA) in ncRNA (non-coding RNA) is not only involved in many biological processes, but also abnormally expressed in many complex diseases. Identification of lncRNA-disease associations accurately is of great significance for understanding the function of lncRNA and disease mechanism. In this paper, a deep learning framework consisting of stacked autoencoder(SAE), multi-scale ResNet and stacked ensemble module, named DHNLDA, was constructed to predict lncRNA-disease associations, which integrates multiple biological data sources and constructing feature matrices. Among them, the biological data including the similarity and the interaction of lncRNAs, diseases and miRNAs are integrated. The feature matrices are obtained by node2vec embedding and feature extraction respectively. Then, the SAE and the multi-scale ResNet are used to learn the complementary information between nodes, and the high-level features of node attributes are obtained. Finally, the fusion of high-level feature is input into the stacked ensemble module to obtain the prediction results of lncRNA-disease associations. The experimental results of five-fold cross-validation show that the AUC of DHNLDA reaches 0.975 better than the existing methods. Case studies of stomach cancer, breast cancer and lung cancer have shown the great ability of DHNLDA to discover the potential lncRNA-disease associations.

Most current metaphor detection methods use restricted context, such as modeling the context of a single sentence. Considering the language environment of metaphors, we argue that combining broader discourse features has a greater impact on the improvement of metaphor detection performance. We propose a metaphor detection method based on fine-grained discourse, which embeds the current sentence and surrounding context in a weighted manner. With the help of fine-grained discourse, our model learns local and remote information as a reference for decision-making, and provides an efficient and natural method for metaphor detection tasks. Experimental results on VU Amsterdam Metaphor Corpus show that our technique surpasses the state-of-the-art models.

The interaction between ncRNA and protein is a kind of crucial molecular activities in a cell. Developing computational methods to predict ncRNA-protein interactions has attracted increasing attentions in recent years. In this work, a novel stacked ensemble learning framework is presented for predicting ncRNA-protein interaction based on heterogeneous feature combinations, named HFC-RPI. Firstly, the compositional features of k-mer with different orders were extracted from the primary sequence and secondary structure of RNA and protein respectively. Secondly, we trained a set of base learners using a variety of heterogeneous combinations of the extracted features respectively. Thirdly, the prediction results of these base learners were employed to train the stacked learner, which output the final prediction result at the higher layer in HFC-RPI. Moreover, in order to improve the generalization of HFC-RPI, when training the base learners, a cross-validation based method was applied. Extensive experimental results showed that the proposed learning framework HFC-RPI was effective and feasible for predicting the interaction of ncRNA and protein. By comparing with state-of-the-art methods, HFC-RPI was superior to them on most performance evaluation metrics.

Deep neural network models for Uyghur personal pronoun resolution learn semantic information for personal pronoun and antecedents, but tend to be short-sighted—they ignore the importance of each feature. In this article, we propose a Uyghur personal pronoun resolution model based on Attention mechanism, Convolutional neural networks and Gated recurrent unit (ATCG). Our model studies the grammatical structure and semantic features of Uyghur, and extracts 11 key features for Uyghur resolution task. Attention mechanism can focus on the importance of words in sentences. Gated Recurrent Unit (GRU) is applied in this model to achieve the interdependent features with long distance. The ATCG model effectively makes up for the shortcomings of relying only on the features of the content level and achieves better classification performance. Experimental results on Uyghur resolution dataset show that our model surpasses the state-of-the-art models.