RESUMO
Entity alignment is a crucial task in knowledge graphs, aiming to match corresponding entities from different knowledge graphs. Due to the scarcity of pre-aligned entities in real-world scenarios, research focused on unsupervised entity alignment has become more popular. However, current unsupervised entity alignment methods suffer from a lack of informative entity guidance, hindering their ability to accurately predict challenging entities with similar names and structures. To solve these problems, we present an unsupervised multi-view contrastive learning framework with an attention-based reranking strategy for entity alignment, named AR-Align. In AR-Align, two kinds of data augmentation methods are employed to provide a complementary view for neighborhood and attribute, respectively. Next, a multi-view contrastive learning method is introduced to reduce the semantic gap between different views of the augmented entities. Moreover, an attention-based reranking strategy is proposed to rerank the hard entities through calculating their weighted sum of embedding similarities on different structures. Experimental results indicate that AR-Align outperforms most both supervised and unsupervised state-of-the-art methods on three benchmark datasets.
RESUMO
The objective of cross-domain sequential recommendation is to forecast upcoming interactions by leveraging past interactions across diverse domains. Most methods aim to utilize single-domain and cross-domain information as much as possible for personalized preference extraction and effective integration. However, on one hand, most models ignore that cross-domain information is composed of multiple single-domains when generating representations. They still treat cross-domain information the same way as single-domain information, resulting in noisy representation generation. Only by imposing certain constraints on cross-domain information during representation generation can subsequent models minimize interference when considering user preferences. On the other hand, some methods neglect the joint consideration of users' long-term and short-term preferences and reduce the weight of cross-domain user preferences to minimize noise interference. To better consider the mutual promotion of cross-domain and single-domains factors, we propose a novel model (C2DREIF) that utilizes Gaussian graph encoders to handle information, effectively constraining the correlation of information and capturing useful contextual information more accurately. It also employs a Top-down transformer to accurately extract user intents within each domain, taking into account the user's long-term and short-term preferences. Additionally, entropy regularized is applied to enhance contrastive learning and mitigate the impact of randomness caused by negative sample composition.
RESUMO
Knowledge tracing (KT) refers to predicting learners' performance in the future according to their historical responses, which has become an essential task in intelligent tutoring systems. Most deep learning-based methods usually model the learners' knowledge states via recurrent neural networks (RNNs) or attention mechanisms. Recently emerging graph neural networks (GNNs) assist the KT model to capture the relationships such as question-skill and question-learner. However, non-pairwise and complex higher-order information among responses is ignored. In addition, a single-channel encoded hidden vector struggles to represent multigranularity knowledge states. To tackle the above problems, we propose a novel KT model named dual-channel adaptive scale hypergraph encoders with cross-view contrastive learning (HyperKT). Specifically, we design an adaptive scale hyperedge distillation component for generating knowledge-aware hyperedges and pattern-aware hyperedges that reflect non-pairwise higher-order features among responses. Then, we propose dual-channel hypergraph encoders to capture multigranularity knowledge states from global and local state hypergraphs. The encoders consist of a simplified hypergraph convolution network and a collaborative hypergraph convolution network. To enhance the supervisory signal in the state hypergraphs, we introduce the cross-view contrastive learning mechanism, which performs among state hypergraph views and their transformed line graph views. Extensive experiments on three real-world datasets demonstrate the superior performance of our HyperKT over the state-of-the-art (SOTA).
RESUMO
Accurate polyp detection is critical for early colorectal cancer diagnosis. Although remarkable progress has been achieved in recent years, the complex colon environment and concealed polyps with unclear boundaries still pose severe challenges in this area. Existing methods either involve computationally expensive context aggregation or lack prior modeling of polyps, resulting in poor performance in challenging cases. In this paper, we propose the Enhanced CenterNet with Contrastive Learning (ECC-PolypDet), a two-stage training & end-to-end inference framework that leverages images and bounding box annotations to train a general model and fine-tune it based on the inference score to obtain a final robust model. Specifically, we conduct Box-assisted Contrastive Learning (BCL) during training to minimize the intra-class difference and maximize the inter-class difference between foreground polyps and backgrounds, enabling our model to capture concealed polyps. Moreover, to enhance the recognition of small polyps, we design the Semantic Flow-guided Feature Pyramid Network (SFFPN) to aggregate multi-scale features and the Heatmap Propagation (HP) module to boost the model's attention on polyp targets. In the fine-tuning stage, we introduce the IoU-guided Sample Re-weighting (ISR) mechanism to prioritize hard samples by adaptively adjusting the loss weight for each sample during fine-tuning. Extensive experiments on six large-scale colonoscopy datasets demonstrate the superiority of our model compared with previous state-of-the-art detectors.
RESUMO
Neural architecture search (NAS) has been applied to design proper 3D networks for medical image segmentation. In order to reduce the computation cost in NAS, researchers tend to adopt weight sharing mechanism to search architectures in a supernet. However, recent studies state that the searched architecture rankings may not be accurate with weight sharing mechanism because the training situations are inconsistent between the searching and training phases. In addition, some NAS algorithms design inflexible supernets that only search operators in a pre-defined backbone and ignore the importance of network topology, which limits the performance of searched architecture. To avoid weight sharing mechanism which may lead to inaccurate results and to comprehensively search network topology and operators, we propose a novel NAS algorithm called NG-NAS. Following the previous studies, we consider the segmentation network as a U-shape structure composed of a set of nodes. Instead of searching from the supernet with a limited search space, our NG-NAS starts from a simple architecture with only 5 nodes, and greedily grows the best candidate node until meeting the constraint. We design 2 kinds of node generations to form various network topological structures and prepare 4 candidate operators for each node. To efficiently evaluate candidate node generations, we use NAS without training strategies. We evaluate our method on several public 3D medical image segmentation benchmarks and achieve state-of-the-art performance, demonstrating the effectiveness of the searched architecture and our NG-NAS. Concretely, our method achieves an average Dice score of 85.11 on MSD liver, 65.70 on MSD brain, and 87.59 in BTCV, which performs much better than the previous SOTA methods.
Assuntos
Algoritmos , Benchmarking , Encéfalo/diagnóstico por imagem , Fígado , Processamento de Imagem Assistida por ComputadorRESUMO
Despite simplicity, stochastic gradient descent (SGD)-like algorithms are successful in training deep neural networks (DNNs). Among various attempts to improve SGD, weight averaging (WA), which averages the weights of multiple models, has recently received much attention in the literature. Broadly, WA falls into two categories: 1) online WA, which averages the weights of multiple models trained in parallel, is designed for reducing the gradient communication overhead of parallel mini-batch SGD and 2) offline WA, which averages the weights of one model at different checkpoints, is typically used to improve the generalization ability of DNNs. Though online and offline WA are similar in form, they are seldom associated with each other. Besides, these methods typically perform either offline parameter averaging or online parameter averaging, but not both. In this work, we first attempt to incorporate online and offline WA into a general training framework termed hierarchical WA (HWA). By leveraging both the online and offline averaging manners, HWA is able to achieve both faster convergence speed and superior generalization performance without any fancy learning rate adjustment. Besides, we also analyze the issues faced by the existing WA methods, and how our HWA addresses them, empirically. Finally, extensive experiments verify that HWA outperforms the state-of-the-art methods significantly.
RESUMO
Self-attention mechanism has been successfully introduced in Graph Neural Networks (GNNs) for graph representation learning and achieved state-of-the-art performances in tasks such as node classification and node attacks. In most existing attention-based GNNs, attention score is only computed between two directly connected nodes with their representation at a single layer. However, this attention score computation method cannot account for its multi-hop neighbors, which supply graph structure information and have influence on many tasks such as link prediction, knowledge graph completion, and adversarial attack as well. In order to address this problem, in this paper, we propose Path Reliability-based Graph Attention Networks (PRGATs), a novel method to incorporate multi-hop neighboring context into attention score computation, enabling to capture longer-range dependencies and large-scale structural information within a single layer. Moreover, path reliability-based attention layer, a core layer of PRGATs, uses a resource-constrain allocation algorithm to compute the reliable path and its attention scores from neighboring nodes to non-neighboring nodes, increasing the receptive field for every message-passing layer. Experimental results on real-world datasets show that, as compared with baselines, our model outperforms existing methods up to 3% on standard node classification and 12% on graph universal adversarial attack.
Assuntos
Algoritmos , Conhecimento , Reprodutibilidade dos Testes , Aprendizagem , Redes Neurais de ComputaçãoRESUMO
Emerging research has highlighted the capacity of microRNA-23a-3p (miR-23a-3p) to alleviate inflammatory pain. However, the molecular mechanism by which miR-23a-3p attenuates inflammatory pain is yet to be fully understood. Hence, the current study aimed to elucidate the mechanism by which miR-23a-3p influences inflammatory pain. Bioinformatics was initially performed to predict the inflammatory pain related downstream targets of miR-23a-3p in macrophage-derived extracellular vesicles (EVs). An animal inflammatory pain model was established using Complete Freund's Adjuvant (CFA). The miR-23a-3p expression was downregulated in the microglia of CFA-induced mice, after which the inflammatory factors were determined by ELISA. FISH and immunofluorescence were performed to analyze the co-localization of miR-23a-3p and microglia. Interestingly, miR-23a-3p was transported to the microglia via M2 macrophage-EVs, which elevated the mechanical allodynia and the thermal hyperalgesia thresholds in mice model. The miR-23a-3p downstream target, USP5, was found to stabilize HDAC2 via deubiquitination to promote its expression while inhibiting the expression of NRF2. Taken together, the key findings of the current study demonstrate that macrophage-derived EVs containing miR-23a-3p regulates the HDAC2/NRF2 axis by decreasing USP5 expression to alleviate inflammatory pain, which may provide novel therapeutic targets for the treatment of inflammatory pain.
Assuntos
Vesículas Extracelulares/metabolismo , Histona Desacetilase 2/metabolismo , Inflamação/metabolismo , Macrófagos/citologia , Fator 2 Relacionado a NF-E2/metabolismo , Dor/metabolismo , Proteases Específicas de Ubiquitina/metabolismo , Animais , Linhagem Celular , Enzimas Desubiquitinantes/metabolismo , Modelos Animais de Doenças , Regulação para Baixo , Estabilidade Enzimática , Vesículas Extracelulares/genética , Inflamação/genética , Inflamação/terapia , Macrófagos/metabolismo , Masculino , Camundongos , Camundongos Endogâmicos C57BL , Microglia/citologia , Microglia/metabolismo , Modelos Biológicos , Dor/genética , Manejo da Dor , Proteases Específicas de Ubiquitina/genética , UbiquitinaçãoRESUMO
This data article compiles the detailed and descriptive experimental data of Wikipedia-based semantic similarity approach called as Neighbourhood Aggregated Semantic Contribution (NASC), presented in Husain, et al. [1]. The JWPL (Java Wikipedia Library)-DataMachine and JWPL WikipediaAPI are used to extract the required Wikipedia features from Wikipedia dump. The dataset presents the disambiguated Wikipedia concepts of the gold standard word similarity benchmarks MC30 (English), RG65es (Spanish) and RG65fr (French) and their associated set of categories in the corresponding Wikipedia category graph (WCG). The dataset also contains the number of ancestors, common ancestors, pages, and common pages in the k-neighbourhood of the associated categories for different levels of parameter k in the English, Spanish, and French WCGs. The presented dataset can be used to assess the semantic similarity between Wikipedia concepts in English (MC30), Spanish (RG65es), and French (RG65fr) languages benchmarks. Moreover, the dataset will be useful for the further analysis and comparison of the taxonomic structures of the English, Spanish, and French WCGs.
RESUMO
The current research progress and the existing problems of uncertain or imprecise knowledge representation and reasoning in description logics are analyzed in this paper. Approximate concepts are introduced to description logics based on rough set theory, and a kind of new rough description logic RDL AC (rough description logic based on approximate concepts) is proposed based on approximate concepts. The syntax, semantics and properties of the RDL AC are given. It is proved that the approximate concept satisfiability (definitely satisfiability and possibly satisfiability) reasoning problem and approximate concepts rough subsumption reasoning problem w.r.t. rough TBox in RDL AC may be reduced to the concept satisfiability reasoning problem in (almost) standard ALC (the description logic that provides the Boolean concept constructors plus the existential and universal restriction constructors). The works of this paper provide logic foundations for approximate ontologies and theoretical foundations for reasoning algorithms of more expressive rough description logics including approximate concepts, number restrictions, nominals, inverse roles and role hierarchies.
RESUMO
It is generally accepted that the management of imprecision and vagueness will yield more intelligent and realistic knowledge-based applications. Description Logics (DLs) are suitable, well-known logics for managing structured knowledge that have gained considerable attention the last decade. The current research progress and the existing problems of uncertain or imprecise knowledge representation and reasoning in DLs are analyzed in this paper. An integration between the theories of fuzzy DLs and rough DLs has been attempted by providing fuzzy rough DLs based on fuzzy rough set theory. The syntax, semantics and properties of fuzzy rough DLs are given. It is proved that the satisfiability, subsumption, entailment and ABox consistency reasoning in fuzzy rough DLs may be reduced to the ABox consistency reasoning in the corresponding fuzzy DLs.