In recent years, multilabel feature selection (MFS) has gained considerable attention as a key technique in the field of data mining. Embedded methods have been widely adopted due to their simplicity and efficiency. However, most embedded MFS methods assume that all labels share the same feature space. Although previous works have noted this issue and proposed solutions, they still suffer from the inability to accurately identify a specific subset of features for each label. Furthermore, most embedded MFS methods often only consider feature similarity through manifold learning concepts, neglecting the impact of feature redundancy, leading to suboptimal performance. In addition, their weight matrix is usually derived from a single perspective of the loss function. To address these limitations, we propose a novel embedded label-specific MFS method based on fuzzy implication granularity information called LSFSFI. This method uses the partial fuzzy mutual implication granularity information to capture the relationships between labels and features, and introduces an auxiliary matrix to achieve label-specific feature selection, which provides a new perspective for the calculation of the weight matrix. In addition, the normalized fuzzy mutual implication granularity information is used to describe the redundancy between features, and a new redundancy constraint regularization term is proposed to ensure a more reasonable assignment of feature weights. Experimental results on several multilabel datasets show that LSFSFI outperforms existing methods in terms of both performance and practicality.

Multi-label feature selection (MFS) plays a vital role in enhancing model performance by identifying the most relevant features. However, most existing methods assume complete feature information and overlook the potential issue of missing features in real-world applications. Furthermore, existing methods mainly focus on feature redundancy while paying insufficient attention to the potential positive interaction among features, which adversely affects the effectiveness of feature selection. To overcome these limitations, this paper proposes a novel incomplete MFS method based on dynamic dual-graph optimization and fuzzy feature interaction (D2GOFI). The method addresses two core challenges systematically: feature-missing processing and feature interaction modeling. Specifically, D2GOFI introduces a fuzzy tolerance relation to mitigate information loss caused by missing features, and proposes fuzzy tolerance implication granularity information to reveal the data's intrinsic structure and guide the feature selection process effectively. To address the potential issue of over-inclusivity introduced by the fuzzy tolerance relation, a non-negative adjustment matrix has been designed to improve the stability and reliability of feature selection. Meanwhile, D2GOFI explicitly models positive fuzzy interaction between features to enhance the expressiveness of feature relevance evaluation. Finally, a dynamic dual-graph optimization strategy is employed to preserve local manifold structures at the feature and label levels simultaneously, ensuring the feature selection process aligns with the data's inherent patterns. Experimental results demonstrate that D2GOFI achieves superior accuracy and robustness across multiple datasets.

The representation of heterogeneous graph nodes has become a hot research topic due to its diverse applications. However, extant approaches can only give consideration partly to three aspects: node structure, semantics and features. To better integrate these three aspects, a new semi-supervised graph neural network is proposed in this paper, called the CRHGNN ( C entrality-based R elation aware H eterogeneous G raph N eural N etwork). The CRHGNN consists of four components as follows. The first component performs the encoding work and aims to capture the structure and semantics of the nodes. The second and third components perform attention mechanisms and information aggregation, respectively. The CRHGNN learns the mutual attention between nodes and carries out feature learning with less overfitting and fewer oversmoothing problems. The last component performs relation fusion, aiming to obtain a compact representation of the nodes. Experiments are conducted to evaluate the representation learning of nodes on three real-world heterogeneous graph datasets and demonstrate that the proposed model is very competitive in terms of node classification and node clustering tasks .

Noisy data is ubiquitous in quantum computer, greatly affecting the performance of various algorithms. However, existing quantum support vector machine models are not equipped with anti-noise ability, and often deliver low performance when learning accurate hyperplane normal vectors from noisy data. To attack this issue, an anti-noise quantum support vector machine algorithm is developed in this paper. Specifically, a weight factor is first embedded into the hinge loss, so as to construct the objective function of anti-noise support vector machine. And then, an alternative iterative optimization strategy and a quantum circuit are designed for solving the objective function, aiming to obtain the normal vector and intercept of the hyperplane that finally divides the data. Finally, the classification and anti-noise effect of the algorithm are verified on artificial dataset and public dataset. Experimental results show that the proposed algorithm is efficient, yet maintains stable accuracy in noisy data.

Distance function is a main metrics of measuring the affinity between two data points in machine learning. Extant distance functions often provide unreachable distance values in real applications. This can lead to incorrect measure of the affinity between data points. This paper proposes a reachable distance function for KNN classification. The reachable distance function is not a geometric direct-line distance between two data points. It gives a consideration to the class attribute of a training dataset when measuring the affinity between data points. Concretely speaking, the reachable distance between data points includes their class center distance and real distance. Its shape looks like “Z,” and we also call it a Z distance function. In this way, the affinity between data points in the same class is always stronger than that in different classes. Or, the intraclass data points are always closer than those interclass data points. We evaluated the reachable distance with experiments, and demonstrated that the proposed distance function achieved better performance in KNN classification.

Medical Dialogue Information Extraction (MDIE) is a promising task for modern medical care systems, which greatly facilitates the development of many real-world applications such as electronic medical record generation, automatic disease diagnosis, etc. Recent methods have firstly achieved considerable performance in Chinese MDIE but still suffer from some inherent limitations, such as poor exploitation of the inter-dependencies in multiple utterances, weak discrimination of the hard samples. In this paper, we propose a contrastive multi-utterance inference (CMUI) method to address these issues. Specifically, we first use a type-aware encoder to provide an efficient encode mechanism toward different categories. Subsequently, we introduce a selective attention mechanism to explicitly capture the dependencies among utterances, which thus constructs a multi-utterance inference. Finally, a supervised contrastive learning approach is integrated into our framework to improve the recognition ability for the hard samples. Extensive experiments show that our model achieves state-of-the-art performance on a public benchmark Chinese-based dataset and delivers significant performance gain on MDIE as compared with baselines. Specifically, we outperform the state-of-the-art results in F1-score by 2.27%, 0.55% in Recall and 3.61% in Precision (The codes that support the findings of this study are openly available in CMUI at https://github.com/jc4357/CMUI.).

The main challenge of cross-modal retrieval is how to efficiently realize semantic alignment and reduce the heterogeneity gap. However, existing approaches ignore the multi-grained semantic knowledge learning from different modalities. To this end, this paper proposes a novel end-to-end cross-modal representation method, termed as Multi-Graph based Hierarchical Semantic Fusion (MG-HSF). This method is an integration of multi-graph hierarchical semantic fusion with cross-modal adversarial learning, which captures fine-grained and coarse-grained semantic knowledge from cross-modal samples, and generate modalities-invariant representations in a common subspace. To evaluate the performance, extensive experiments are conducted on three benchmarks. The experimental results show that our method is superior than the state-of-the-arts.

Graph convolutional networks (GCN) have been widely used in processing graphs and networks data. However, some recent research experiments show that the existing graph convolutional networks have isseus when integrating node features and topology structure. In order to remedy the weakness, we propose a new GCN architecture. Firstly, the proposed architecture introduces the cross-stitch networks into GCN with improved cross-stitch units. Cross-stitch networks spread information/knowledge between node features and topology structure, and obtains consistent learned representation by integrating information of node features and topology structure at the same time. Therefore, the proposed model can capture various channel information in all images through multiple channels. Secondly, an attention mechanism is to further extract the most relevant information between channel embeddings. Experiments on six benchmark datasets shows that our method outperforms all comparison methods on different evaluation indicators.

Hierarchical graph pooling has shown great potential for capturing high-quality graph representations through the node cluster selection mechanism. However, the current node cluster selection methods have inadequate clustering issues, and their scoring methods rely too much on the node representation, resulting in excessive graph structure information loss during pooling. In this paper, a local capsule pooling network (LCPN) is proposed to alleviate the above issues. Specifically, (i) a local capsule pooling (LCP) is proposed to alleviate the issue of insufficient clustering; (ii) a task-aware readout (TAR) mechanism is proposed to obtain a more expressive graph representation; (iii) a pooling information loss (PIL) term is proposed to further alleviate the information loss caused by pooling during training. Experimental results on the graph classification task, the graph reconstruction task, and the pooled graph adjacency visualization task show the superior performance of the proposed LCPN and demonstrate its effectiveness and efficiency.