The Dung beetle optimization (DBO) algorithm, devised by Jiankai Xue in 2022, is known for its strong capabilities and fast convergence. However, it does have certain limitations, including insufficiently random population initialization, slow search speed, inadequate global capabilities. Drawing inspiration from the mathematical properties of Sinh Cosh functions, we proposed a new metaheuristic Sinh–Cosh Beetle Optimization (SCDBO). By leveraging functions to disrupt initial distribution...

Predicting drug–target interactions (DTIs) is a crucial step in the development of new drugs and drug repurposing. In this paper, we propose novel prediction model called MCF-DTI. The utilizes SMILES representation sequence features targets, employing multi-scale convolutional neural network (MSCNN) with parallel shared-weight modules to extract from side. For target side, it combines MSCNN Transformer capture both local global effectively. extracted are then weighted fused, enabling...

Few-shot learning (FSL) poses a significant challenge in classifying unseen classes with limited samples, primarily stemming from the scarcity of data. Although numerous generative approaches have been investigated for FSL, their generation process often results entangled outputs, exacerbating distribution shift inherent FSL. Consequently, this considerably hampers overall quality generated samples. Addressing concern, we present pioneering framework called DisGenIB, which leverages an...

In the age of scholarly big data, efficiently navigating and analyzing vast corpus scientific literature is a significant challenge. This paper introduces specialized pre-trained BERT-based language model, termed SPBERT, which enhances natural processing tasks specifically tailored to domain analysis. Our method employs novel neural network embedding technique that leverages textual components, such as keywords, titles, abstracts, full texts, represent papers in vector space. By integrating...

This study introduces a deep learning framework based on SMILES representations of chemical structures to predict drug-drug interactions (DDIs). The model extracts Morgan fingerprints and key molecular descriptors, transforming them into raw graphical features for input modified ResNet18 architecture. residual network, enhanced with regularization techniques, efficiently addresses training issues such as gradient vanishing exploding, resulting in superior predictive performance. Experimental...

In recent years, the use of deep learning methods for drug-target interaction (DTI) prediction has become mainstream research direction. Drugs, targets, and other related biological chemical properties have constructed a very complex network structure. How to effectively extract features predict target big challenge. Graph Convolutional Neural Network (GCN) is one effective networks. It extends convolution operation from traditional European space non-Euclidean space, can simultaneously...

Researchers recently investigated to explain Graph Neural Networks (GNNs) on the access a task-specific GNN, which may hinder their wide applications in practice. Specifically, explanation methods are incapable of explaining pretrained GNNs whose downstream tasks usually inaccessible, not mention giving explanations for transferable knowledge GNNs. Additionally, only consider target models' output label space, coarse-grained and insufficient reflect model's internal logic. To address these...

Recent studies have revealed that GNNs are vulnerable to adversarial attacks. Most existing robust graph learning methods measure model robustness based on label information, rendering them infeasible when information is not available. A straightforward direction employ the widely used Infomax technique from typical Unsupervised Graph Representation Learning (UGRL) learn unsupervised representations. Nonetheless, directly transplanting UGRL may involve a biased assumption. In light of...

An unsupervised learning network is developed by incorporating the idea of non-linear mapping (NLM) into a backpropagation (BP) algorithm. This performs process 2iteratively adjusting its parameters to minimize an appropriate criterion using generalized BP (GBP) generalization makes algorithms more competent for many supervised and tasks provided that has been designed. Results numerical simulation real data show proposed technique promising approach visualize multidimensional clusters...

본 논문에서는 인공 신경망의 일종인 Extreme Learning Machine의 학습 알고리즘을 기반으로 하여 노이즈에 강한 특성을 보이는 퍼지 집합 이론을 이용한 새로운 패턴 분류기를 제안 한다. 기존 신경망에 비해 학습속도가 매우 빠르며, 모델의 일반화 성능이 우수하다고 알려진 분류기에 적용하여 분류기의 속도와 분류 성능을 개선 제안된 퍼지패턴 평가하기 위하여, 다양한 머신 러닝 데이터 집합을 사용한다. In this paper, we introduce a new pattern classifier which is based on the learning algorithm of Machine sort artificial neural networks and fuzzy set theory well known as being robust to noise. The used in faster than conventional networks. key advantage...

Due to the superior performance of Graph Neural Networks (GNNs) in various domains, there is an increasing interest GNN explanation problem "\emph{which fraction input graph most crucial decide model's decision?}" Existing methods focus on supervised settings, \eg, node classification and classification, while for unsupervised graph-level representation learning still unexplored. The opaqueness representations may lead unexpected risks when deployed high-stake decision-making scenarios. In...

Cross-modal retrieval is crucial in understanding latent correspondences across modalities. However, existing methods implicitly assume well-matched training data, which impractical as real-world data inevitably involves imperfect alignments, i.e., noisy correspondences. Although some works explore similarity-based strategies to address such noise, they suffer from sub-optimal similarity predictions influenced by modality-exclusive information (MEI), e.g., background noise images and...

Recent studies have shown that graph convolution networks (GCNs) are vulnerable to carefully designed attacks, which aim cause misclassification of a specific node on the with unnoticeable perturbations. However, vast majority existing works cannot handle large-scale graphs because their high time complexity. Additionally, mainly focus manipulating nodes graph, while in practice, attackers usually do not privilege modify information nodes. In this paper, we develop more scalable framework...

Web service is a service-oriented architecture which can achieve real platform-independent and language-independent. Ontology formal representation of the knowledge by set concepts within domain relationships between those concepts. In this paper, we analyzed key technologies services discus one such approach that involves adding semantics to WSDL UDDI using OWL ontologies in order describe discover Services accurately for us. Based on some problems current Web-based Learning system, propose...