Domain generalization (DG) aims to learn a model that generalizes well unseen target domains utilizing multiple source without re-training. Most existing DG works are based on convolutional neural networks (CNNs). However, the local operation of convolution kernel makes focus too much representations (e.g., texture), which inherently causes more prone overfit and hampers its ability. Recently, several MLP-based methods have achieved promising results in supervised learning tasks by global...

Deep Neural Networks have exhibited considerable success in various visual tasks. However, when applied to unseen test datasets, state-of-the-art models often suffer performance degradation due domain shifts. In this paper, we introduce a novel approach for generalization from perspective of enhancing the robustness channels feature maps We observe that trained on source domains contain substantial number exhibit unstable activations across different domains, which are inclined capture...

To segment medical images with distribution shifts, domain generalization (DG) has emerged as a promising setting to train models on source domains that can generalize unseen target domains. Existing DG methods are mainly based CNN or ViT architectures. Recently, advanced state space models, represented by Mamba, have shown results in various supervised image segmentation. The success of Mamba is primarily owing its ability capture long-range dependencies while keeping linear complexity...

The main purposes of this study are to analyse the evaluation tailings dam stability under multiple factors and prevent accidents more effectively by proposing a composite risk analysis model. model combining TOPSIS bow tie is presented in paper. Firstly, new formula was adopted calculate integrated weights based on subjective objective theory introduced. Secondly, taking uranium reservoir south China as an example, index values constant 10 dams determined according eight aspects...

Domain generalization (DG) aims to learn a generic model from multiple observed source domains that generalizes well arbitrary unseen target without further training. The major challenge in DG is the inevitably faces severe overfitting issue due domain gap between and domains. To mitigate this problem, some dropout-based methods have been proposed resist by discarding part of representation intermediate layers. However, we observe most these only conduct dropout operation specific layers,...

Deep Neural Networks have exhibited considerable success in various visual tasks. However, when applied to unseen test datasets, state-of-the-art models often suffer performance degradation due domain shifts. In this paper, we introduce a novel approach for generalization from perspective of enhancing the robustness channels feature maps We observe that trained on source domains contain substantial number exhibit unstable activations across different domains, which are inclined capture...

Domain generalization (DG) intends to train a model on multiple source domains ensure that it can generalize well an arbitrary unseen target domain. The acquisition of domain-invariant representations is pivotal for DG as they possess the ability capture inherent semantic information data, mitigate influence domain shift, and enhance capability model. Adopting perspectives, such sample feature, proves be effective. perspective facilitates data augmentation through manipulation techniques,...

Internal diffusion of molecules in porous materials plays an important role many chemical processes. However, the pore capacity cannot be measured by conventional catalyst characterization methods. In present paper, a factor, ratio diffusion-constriction factor to tortuosity materials, was proposed measure ability pores inside solid and method for measuring using well-defined uniform size material as reference. The calculated based on effective coefficients reference materials. measurement...

A bidisperse model for transient diffusion and adsorption processes in porous materials is presented this paper. The mathematical solved by numerical methods based on finite elements combined with the linear driving force approximation. criterion to identify controlling mechanism (macropore diffusion, micropore or both) proposed. effects of different isotherms (linear, Freundlich, Langmuir) concentration profiles curves fractional uptake versus time are investigated. In addition, influences...

Domain generalization (DG) intends to train a model on multiple source domains ensure that it can generalize well an arbitrary unseen target domain. The acquisition of domain-invariant representations is pivotal for DG as they possess the ability capture inherent semantic information data, mitigate influence domain shift, and enhance capability model. Adopting perspectives, such sample feature, proves be effective. perspective facilitates data augmentation through manipulation techniques,...

Domain generalization (DG) aims to enhance the model robustness against domain shifts without accessing target domains. A prevalent category of methods for DG is data augmentation, which focuses on generating virtual samples simulate shifts. However, existing augmentation techniques in are mainly tailored convolutional neural networks (CNNs), with limited exploration token-based architectures, i.e., vision transformer (ViT) and multi-layer perceptrons (MLP) models. In this paper, we study...

Abstract Manifold-learning is particularly useful to resolve the complex cellular state space from single-cell RNA sequences. While current manifold-learning methods provide insights into cell fate by inferring graph-based trajectory at level, challenges remain retrieve interpretable biology underlying diverse states. Here, we described MGPfact XMBD , a model-based framework and capable factorize development trajectories independent bifurcation processes of gene sets, thus enables inference...

Domain generalization (DG) aims to enhance the model robustness against domain shifts without accessing target domains. A prevalent category of methods for DG is data augmentation, which focuses on generating virtual samples simulate shifts. However, existing augmentation techniques in are mainly tailored convolutional neural networks (CNNs), with limited exploration token-based architectures, i.e., vision transformer (ViT) and multi-layer perceptrons (MLP) models. In this paper, we study...

Revealing the microdynamic mechanism and thermodynamic properties of gelation process waxy crude oil at microscale plays an important role in ensuring safe production transportation oil. Considering influence different paraffin configurations with same molecular formula on gelling oils, three models systems were established, including paraffins n-alkanes isoparaffins distributions branched chains, as well light oils asphaltenes. Analysis energy changes, aggregation behavior, a system during...

Domain Generalization (DG) aims to enable models generalize unseen target domains by learning from multiple source domains. Existing DG methods primarily rely on convolutional neural networks (CNNs), which inherently learn texture biases due their limited receptive fields, making them prone overfitting While some works have introduced transformer-based (ViTs) for leverage the global field, these incur high computational costs quadratic complexity of self-attention. Recently, advanced state...