Cross domain recommendation (CDR) is one popular research topic in recommender systems. This article focuses on a scenario for CDR where different domains share the same set of users but no overlapping items. The majority recent methods have explored shared-user representation to transfer knowledge across domains. However, idea resorts learning overlapped features user preferences and suppresses domain-specific features. Other works try capture by an MLP mapping require heuristic human...

Anchor selection or learning has become a critical component in large-scale multi-view clustering. Existing anchor-based methods, which either select-then-fix initialize-then-optimize with orthogonality, yield promising performance. However, these methods still suffer from instability of initialization insufficient depiction data distribution. Moreover, the desired properties anchors clustering remain unspecified. To address issues, this paper first formalizes characteristics anchors, namely...

Anchor based pair-wised multi-view clustering often assumes data are paired, and has demonstrated significant advancements in recent years. However, this presumption is easily violated, commonly unpaired fully practical applications due to the influence of collection storage processes. Addressing large-scale through anchor learning remains a research gap. The absence pairing disrupts consistency complementarity multiple views, posing challenges powerful meaningful anchors bipartite graphs...

Network alignment is a critical task in wide variety of fields. Many existing works leverage on representation learning to accomplish this without eliminating domain bias induced by domain-dependent features, which yield inferior performance. This paper proposes unified deep architecture ( <italic xmlns:mml="http://www.w3.org/1998/Math/MathML" xmlns:xlink="http://www.w3.org/1999/xlink">DANA</i> ) obtain domain-invariant for network via an adversarial classifier. Specifically, we employ the...

Unsupervised multi-view bipartite graph clustering (MVBGC) is a fast-growing research, due to promising scalability in large-scale tasks. Although many variants are proposed by various strategies, common design construct the directly from input data, i.e. only consider unidirectional "encoding" process. However, "encoding-decoding" mechanism popular for deep learning, most representative one auto-encoder (AE). Enlightened this, this paper rethinks existing MVBGC paradigms and transfers into...

BACKGROUND: Gray matter (GM) and white (WM) impairments are both associated with raised blood pressure (BP), although whether elevated BP is differentially the GM WM aging process remains inadequately examined. METHODS: We included 37 327 participants diffusion-weighted imaging (DWI) 39 630 T1-weighted scans from UK Biobank. was classified into 4 categories: normal BP, high-normal grade 1, 2 hypertension. Brain age gaps (BAGs) for (BAG ) were derived T1 separately using...

The reliability of facility location problem has aroused wide concern recently. Many researchers focus on reliable and robust systems design under component failures have obtained promising performance. However, the target a system are to large degree adversely affected by edge in network, which remains deep study. In this paper, we systems’ subject failures. For system, formulate two models based classical uncapacitated fixed-charge deterministic stochastic cases. specific example,...

Cross domain recommendation (CDR) is one popular research topic in recommender systems. This paper focuses on a scenario for CDR where different domains share the same set of users but no overlapping items. The majority recent methods have explored shared-user representation to transfer knowledge across domains. However, idea resorts learn overlapped features user preferences and suppresses domain-specific features. Other works try capture by an MLP mapping require heuristic human choosing...

Inspired by the impressive success of contrastive learning (CL), a variety graph augmentation strategies have been employed to learn node representations in self-supervised manner. Existing methods construct samples adding perturbations structure or attributes. Although results are achieved, it is rather blind wealth prior information assumed: with increase perturbation degree applied on original graph: 1) similarity between and generated augmented gradually decreases 2) discrimination all...

Label embedding plays an important role in many real-world applications. To enhance the label relatedness captured by embeddings, multiple contexts can be adopted. However, these are heterogeneous and often partially observed practical tasks, imposing significant challenges to capture overall among labels. In this paper, we propose a general Partial Heterogeneous Context Embedding (PHCLE) framework address challenges. Categorizing into two groups, relational context descriptive context,...

Driver mental fatigue leads to thousands of traffic accidents. The increasing quality and availability low-cost electroencephalogram (EEG) systems offer possibilities for practical monitoring. However, non-data-driven methods, designed practical, complex situations, usually rely on handcrafted data statistics EEG signals. To reduce human involvement, we introduce a data-driven methodology online detection: self-weight ordinal regression (SWORE). Reaction time (RT), referring the length...

Abstract As the brain ages, it almost invariably accumulates vascular pathology, which differentially affects cerebral white matter. A rich body of research has investigated link between risk factors and brain. One less studied questions is that among various modifiable factors, most debilitating one for matter health? specific age was developed to evaluate overall health from diffusion weighted imaging, using a three-dimensional convolutional neural network deep learning model in both...

Brain age has been widely investigated by using the whole brain image. However, of some specific regions, such as those related to hippocampus, remains underexplored. This study developed prediction models for left and right hippocampus-centred regions interest (hippocampus ROI) three-dimensional convolutional neural networks (3D-CNN) based on MRI scans from 31,370 healthy participants in UK Biobank. The hippocampus ROI (HA) gap was calculated subtracting chronological predicted HA....

Worst-case fairness with off-the-shelf demographics achieves group parity by maximizing the model utility of worst-off group. Nevertheless, demographic information is often unavailable in practical scenarios, which impedes use such a direct max-min formulation. Recent advances have reframed this learning problem introducing lower bound minimal partition ratio, denoted as $\alpha$, side information, referred to ``$\alpha$-sized worst-case fairness'' paper. We first justify significance...