Colorectal cancer is the second and third most common in women men, respectively. Pathological diagnosis "gold standard" for tumor diagnosis. Accurate segmentation of glands from tissue images a crucial step assisting pathologists their The typical methods gland form dense image representation, ignoring its texture multi-scale attention information. Therefore, we utilize Gabor-based module to extract information at different scales directions histopathology images. This paper also designs...

Abstract Approximate reasoning systems facilitate fuzzy inference through activating if–then rules in which attribute values are imprecisely described. Fuzzy rule interpolation (FRI) supports such with sparse bases where certain observations may not match any existing rules, manipulation of that bear similarity an unmatched observation. This differs from classical rule-based requires direct pattern matching between and the given rules. FRI techniques have been continuously investigated for...

Convolutional neural networks (CNNs) have been widely utilized in many computer vision tasks. However, CNNs a fixed reception field and lack the ability of long-range perception, which is crucial to human pose estimation. Transformer architecture has adopted applications recently proven be highly effective architecture. We are interested exploring its capability estimation, thus propose novel model based on transformer, enhanced with feature pyramid fusion structure. More specifically, we...

Significant progress has been made in video captioning recent years. However, most existing methods directly learn from all given captions without distinguishing the styles of captions. The large diversity these might bring ambiguity to model learning. To address this issue, we propose a style-aware two-stage learning framework. In first stage, is trained with separate styles, including length style (short, medium, long), action (single or multiple actions), and object (one more). For...

<title>Abstract</title> Grammatical Error Correction (GEC) aims to generate a correct sentence from an erroneous one. However, due the scarcity of high-quality datasets, it still faces challenges in achieving deep semantic understanding sentences. This paper proposes Adaptive Syntax-Enhanced (ASynGEC) model that incorporates adaptive dependency syntax information into GEC framework. During processing phase, word-level syntactic is chosen as input. captures rich relationships between words....

In recent years, methods based on heterogeneous graph neural networks (HGNNs) have been widely used for embedding graphs (HGs) due to their ability effectively encode the rich information from HGs into low-dimensional node embeddings. Existing HGNNs focus neighbor aggregation and semantic fusion while neglecting HG structure learning paradigms. However, original data might lack features, which existing models may not account for. Additionally, exclusively relying a single supervised approach...

In this paper, we propose a deep wavelet neural network (DWNN) model to approximate the natural phenomena that are described by some classical PDEs. Concretely, introduce wavelets architecture obtain fine feature description and extraction. That is, constructs expansion layer based on family of vanishing momentum wavelets. Second, Gaussian error function is considered as activation owing its fast convergence rate zero-centered output. Third, design cost considering residual governing...

.For the nonlinear conservative system, how to design an efficient scheme preserve as manyinvariants possible is a challenging task. The aim of this paper construct finite difference/spectral method for Klein–Gordon–Schrödinger (KGS) system on infinite domains \(\mathbb{R}^d\) ( \(d=1, 2\) , and 3) conserve three kinds most important invariants, namely, mass, energy, momentum. Regarding mass- momentum-conservation laws \(d+1\) globally physical constraints, we elaborately combine exponential...

The nonlinear propagation of ion-acoustic waves is theoretically reported in a collisional plasma containing strongly coupled ions and nonthermal electrons featuring Tsallis distribution. For this purpose, the integro-differential form generalized hydrodynamic model used to investigate strong-coupling effect. modified complex Ginzburg–Landau equation with linear dissipative term derived for potential wave amplitude regime, modulation instability examined. When effect neglected, reduces...

In scientific computing, neural networks have been widely used to solve partial differential equations (PDEs). this paper, we propose a novel RBF-assisted hybrid network for approximating solutions PDEs. Inspired by the tendency of physics-informed (PINNs) become local approximations after training, proposed method utilizes radial basis function (RBF) provide normalization and localization properties input data. The objective strategy is assist in solving PDEs more effectively. During...

Unsupervised domain adaptation is a problem which exploits the knowledge learned from resource-rich to obtain an accurate classifier for resource-poor domain. Most of existing methods lift performance by reducing differences between distributions, such as difference marginal probability conditional or both. However, all these consider two distributions be equally important, could lead poor classification in practical applications. Therefore, balanced factor required weigh compensate degraded...