Redundancy Is Not What You Need: An Embedding Fusion Graph Auto-Encoder for Self-Supervised Graph Representation Learning.

Attribute graphs are a crucial data structure for graph communities. However, the presence of redundancy and noise in the attribute graph can impair the aggregation effect of integrating two different heterogeneous distributions of attribute and structural features, resulting in inconsistent and distorted data that ultimately compromises the accuracy and reliability of attribute graph learning. For instance, redundant or irrelevant attributes can result in overfitting, while noisy attributes can lead to underfitting. Similarly, redundant or noisy structural features can affect the accuracy of graph representations, making it challenging to distinguish between different nodes or communities. To address these issues, we propose the embedded fusion graph auto-encoder framework for self-supervised learning (SSL), which leverages multitask learning to fuse node features across different tasks to reduce redundancy. The embedding fusion graph auto-encoder (EFGAE) framework comprises two phases: pretraining (PT) and downstream task learning (DTL). During the PT phase, EFGAE uses a graph auto-encoder (GAE) based on adversarial contrastive learning to learn structural and attribute embeddings separately and then fuses these embeddings to obtain a representation of the entire graph. During the DTL phase, we introduce an adaptive graph convolutional network (AGCN), which is applied to graph neural network (GNN) classifiers to enhance recognition for downstream tasks. The experimental results demonstrate that our approach outperforms state-of-the-art (SOTA) techniques in terms of accuracy, generalization ability, and robustness.

Beyond low-pass filtering on large-scale graphs via Adaptive Filtering Graph Neural Networks.

Graph Neural Networks (GNNs) have emerged as a crucial deep learning framework for graph-structured data. However, existing GNNs suffer from the scalability limitation, which hinders their practical implementation in industrial settings. Many scalable GNNs have been proposed to address this limitation. However, they have been proven to act as low-pass graph filters, which discard the valuable middle- and high-frequency information. This paper proposes a novel graph neural network named Adaptive Filtering Graph Neural Networks (AFGNN), which can capture all frequency information on large-scale graphs. AFGNN consists of two stages. The first stage utilizes low-, middle-, and high-pass graph filters to extract comprehensive frequency information without introducing additional parameters. This computation is a one-time task and is pre-computed before training, ensuring its scalability. The second stage incorporates a node-level attention-based feature combination, enabling the generation of customized graph filters for each node, contrary to existing spectral GNNs that employ uniform graph filters for the entire graph. AFGNN is suitable for mini-batch training, and can enhance scalability and efficiently capture all frequency information from large-scale graphs. We evaluate AFGNN by comparing its ability to capture all frequency information with spectral GNNs, and its scalability with scalable GNNs. Experimental results illustrate that AFGNN surpasses both scalable GNNs and spectral GNNs, highlighting its superiority.

A Dual-Masked Deep Structural Clustering Network With Adaptive Bidirectional Information Delivery.

Structured clustering networks, which alleviate the oversmoothing issue by delivering hidden features from autoencoder (AE) to graph convolutional networks (GCNs), involve two shortcomings for the clustering task. For one thing, they used vanilla structure to learn clustering representations without considering feature and structure corruption; for another thing, they exhibit network degradation and vanishing gradient issues after stacking multilayer GCNs. In this article, we propose a clustering method called dual-masked deep structural clustering network (DMDSC) with adaptive bidirectional information delivery (ABID). Specifically, DMDSC enables generative self-supervised learning to mine deeper interstructure and interfeature correlations by simultaneously reconstructing corrupted structures and features. Furthermore, DMDSC develops an ABID module to establish an information transfer channel between each pairwise layer of AE and GCNs to alleviate the oversmoothing and vanishing gradient problems. Numerous experiments on six benchmark datasets have shown that the proposed DMDSC outperforms the most advanced deep clustering algorithms.

Warm and Harsh Parenting, Self-Kindness and Self-Judgment, and Well-Being: An Examination of Developmental Differences in a Large Sample of Adolescents.

Objectives: This study aimed to examine the associations between warm and harsh parenting and adolescent well-being, and the mediating effects of self-kindness and self-judgment, in relationships. Moreover, this study investigated developmental differences across three adolescence stages (early, middle, and late). Methods: In total, 14,776 Chinese adolescents (mean age = 13.53 ± 2.08, 52.3% males), including individuals in early (10-12 years old, N = 5055), middle (13-15 years old, N = 6714), and late adolescence (16-18 years old, N = 3007) participated in this study. All the adolescents rated their levels of warm and harsh parenting, self-kindness and self-judgment, and well-being. Structural equation modeling (SEM) was adopted to examine the mediation model. Multi-group analysis was conducted to investigate differences in the mediation model across the different developmental stages. Results: Both warm and harsh parenting were related to adolescent well-being through the mediating effects of self-kindness and self-judgment. However, warm parenting exerted a more substantial impact on adolescent well-being. Self-kindness had a more robust mediating effect than self-judgment in relationships. Moreover, harsh parenting had a weaker impact on adolescent well-being in late adolescence than in early and middle adolescence. Warm parenting had a more significant impact on adolescent well-being in early adolescence than in middle and late adolescence. Conclusions: Overall, warm parenting had a more substantial effect than harsh parenting on adolescent well-being. The findings also highlighted the crucial mediating effect of self-kindness in the relationships between parenting and well-being. Moreover, this study also indicated the importance of warm parenting in early adolescence. Intervention programs should focus on enhancing the level of warm parenting to promote self-kindness in adolescents, in order to improve their well-being.

Multi-graph Fusion Graph Convolutional Networks with pseudo-label supervision.

Graph convolutional networks (GCNs) have become a popular tool for learning unstructured graph data due to their powerful learning ability. Many researchers have been interested in fusing topological structures and node features to extract the correlation information for classification tasks. However, it is inadequate to integrate the embedding from topology and feature spaces to gain the most correlated information. At the same time, most GCN-based methods assume that the topology graph or feature graph is compatible with the properties of GCNs, but this is usually not satisfied since meaningless, missing, or even unreal edges are very common in actual graphs. To obtain a more robust and accurate graph structure, we intend to construct an adaptive graph with topology and feature graphs. We propose Multi-graph Fusion Graph Convolutional Networks with pseudo-label supervision (MFGCN), which learn a connected embedding by fusing the multi-graphs and node features. We can obtain the final node embedding for semi-supervised node classification by propagating node features over multi-graphs. Furthermore, to alleviate the problem of labels missing in semi-supervised classification, a pseudo-label generation mechanism is proposed to generate more reliable pseudo-labels based on the similarity of node features. Extensive experiments on six benchmark datasets demonstrate the superiority of MFGCN over state-of-the-art classification methods.

Adaptive graph convolutional clustering network with optimal probabilistic graph.

The graph convolutional network (GCN)-based clustering approaches have achieved the impressive performance due to strong ability of exploiting the topological structure. The adjacency graph seriously affects the clustering performance, especially for non-graph data. Existing approaches usually conduct two independent steps, i.e., constructing a fixed graph structure and then graph embedding representation learning by GCN. However, the constructed graph structure may be unreliable one due to noisy data, resulting in sub-optimal graph embedding representation. In this paper, we propose an adaptive graph convolutional clustering network (AGCCN) to alternatively learn the similarity graph structure and node embedding representation in a unified framework. Our AGCCN learns the weighted adjacency graph adaptively from the node representations by solving the optimization problem of graph learning, in which adaptive and optimal neighbors for each sample are assigned with probabilistic way according to local connectivity. Then, the attribute feature extracted by parallel Auto-Encoder (AE) module is fused into the input of adaptive graph convolution module layer-by-layer to learn the comprehensive node embedding representation and strengthen its representation ability. This also skillfully alleviates the over-smoothing problem of GCN. To further improve the discriminant ability of node representation, a dual self-supervised clustering mechanism is designed to guide model optimization with pseudo-labels information. Extensive experimental results on various real-world datasets consistently show the superiority and effectiveness of the proposed deep graph clustering method.

Pilot Behavior Recognition Based on Multi-Modality Fusion Technology Using Physiological Characteristics.

With the development of the autopilot system, the main task of a pilot has changed from controlling the aircraft to supervising the autopilot system and making critical decisions. Therefore, the human-machine interaction system needs to be improved accordingly. A key step to improving the human-machine interaction system is to improve its understanding of the pilots' status, including fatigue, stress, workload, etc. Monitoring pilots' status can effectively prevent human error and achieve optimal human-machine collaboration. As such, there is a need to recognize pilots' status and predict the behaviors responsible for changes of state. For this purpose, in this study, 14 Air Force cadets fly in an F-35 Lightning II Joint Strike Fighter simulator through a series of maneuvers involving takeoff, level flight, turn and hover, roll, somersault, and stall. Electro cardio (ECG), myoelectricity (EMG), galvanic skin response (GSR), respiration (RESP), and skin temperature (SKT) measurements are derived through wearable physiological data collection devices. Physiological indicators influenced by the pilot's behavioral status are objectively analyzed. Multi-modality fusion technology (MTF) is adopted to fuse these data in the feature layer. Additionally, four classifiers are integrated to identify pilots' behaviors in the strategy layer. The results indicate that MTF can help to recognize pilot behavior in a more comprehensive and precise way.

Improving object detection quality with structural constraints.

Recent researches revealed object detection networks using the simple "classification loss + localization loss" training objective are not effectively optimized in many cases, while providing additional constraints on network features could effectively improve object detection quality. Specifically, some works used constraints on training sample relations to successfully learn discriminative network features. Based on these observations, we propose Structural Constraint for improving object detection quality. Structural constraint supervises feature learning in classification and localization network branches with Fisher Loss and Equi-proportion Loss respectively, by requiring feature similarities of training sample pairs to be consistent with corresponding ground truth label similarities. Structural constraint could be applied to all object detection network architectures with the assist of our Proxy Feature design. Our experiment results showed that structural constraint mechanism is able to optimize object class instances' distribution in network feature space, and consequently detection results. Evaluations on MSCOCO2017 and KITTI datasets showed that our structural constraint mechanism is able to assist baseline networks to outperform modern counterpart detectors in terms of object detection quality.

Long-term renal outcomes of mesangial proliferative lupus nephritis in Chinese patients.

OBJECTIVE: This study aimed to explore the long-term outcomes of mesangial proliferative lupus nephritis (LN class II) and the factors associated with its relapse and histological transformation in Chinese patients. METHODS: 104 SLE patients with biopsy-proven LN class II were included and divided into proteinuria group (proteinuria ≥ 0.4 g/24 h, with or without microscopic hematuria) and hematuria group (microscopic hematuria with proteinuria < 0.4 g/24 h).Patients were treated with glucocorticoid alone (GC monotherapy) or GC in combination with other immunosuppressant (combination therapy). The rates of remission, relapse, histological transformation, end-stage renal disease (ESRD), adverse events, and risk factors related to the outcomes were analyzed. RESULTS: During the median follow-up of 77.5 (IQR 58-116.5) months, all the 104 patients achieved remission. Relapse occurred in 69 cases (66.3%), of which 37 were of renal relapse (35.6%). Histological transformation was found in 14 of the 16 (87.5%) cases who received repeated renal biopsy after renal relapse. At the end of follow-up, 3 (2.9%) patients developed ESRD. There were no significant differences in the rates of relapse, histological transformation, adverse events and in the time from remission to relapse between the proteinuria group and the hematuria group. In contrast, the cumulative relapse rate in the GC monotherapy group was much higher than that in the combination group (P < 0.01). Adverse events occurred in 55 (57.3%) patients during follow-up. CONCLUSIONS: Patients with LN class II have high rates of relapse and renal histological transformation and need optimal maintenance therapy. KEY POINTS: • The rates of relapse and histological transformation are high in patients with LN class II. • Patients with LN class II are suggested to receive combination therapy and consider repeat renal biopsy after renal relapse.

Cellphone usage and addiction among Chinese children and adolescents and related factors / 中国学校卫生

Objective@#To understand the current situation and associated factors of cellphone usage and addiction among Chinese children and adolescents, to provide reference for effective prevention and intervention of cellphone addiction.@*Methods@#Using a stratified random sampling approach, 11 213 children and adolescents and their parents from 31 provinces, municipalities and autonomous regions in China were recruited and surveyed.@*Results@#The median of daily mobile phone use time among Chinese children and adolescents were 120.00 minutes, as reported by either children or parents. Child s age( β =0.12), hedonic( β =0.11) and social( β =0.09) cellphone use motivations positively related to time spent on cellphone( P <0.01). Cellphone related parental communication( β =-0.06) and knowledge( β =-0.03), as well as cellphone usage on instrumental( β =-0.04) or self representation( β =-0.16) motivation negatively related to time spent on cellphone( P <0.05). Child s age( β =-0.04), cellphone related parental communication( β =-0.09) and awareness( β =-0.14), cellphone use on instrumental motivation( β =-0.22) were negatively associated with cellphone addiction among children and adolescents( P <0.05). Cellphone related parental monitoring( β =0.07), as well as cellphone usage on self representation motivation( β =0.03) or hedonic motivation( β =0.29) positively related to cellphone addiction in children and adolescents( P <0.05).@*Conclusion@#Time spent on mobile phone and mobile phone addiction of Chinese children and adolescents are influenced by various internal and external factors, such as the mobile phone use motivation and parenting style.Future school education should help children develop scientific motivation for mobile phone use. Family education should help parents develop positive parenting behaviors such as communication and awareness, so as to reduce the possibility of improper mobile phone use.

Attentive 3D-Ghost Module for Dynamic Hand Gesture Recognition with Positive Knowledge Transfer.

Hand gesture recognition is a challenging topic in the field of computer vision. Multimodal hand gesture recognition based on RGB-D is with higher accuracy than that of only RGB or depth. It is not difficult to conclude that the gain originates from the complementary information existing in the two modalities. However, in reality, multimodal data are not always easy to acquire simultaneously, while unimodal RGB or depth hand gesture data are more general. Therefore, one hand gesture system is expected, in which only unimordal RGB or Depth data is supported for testing, while multimodal RGB-D data is available for training so as to attain the complementary information. Fortunately, a kind of method via multimodal training and unimodal testing has been proposed. However, unimodal feature representation and cross-modality transfer still need to be further improved. To this end, this paper proposes a new 3D-Ghost and Spatial Attention Inflated 3D ConvNet (3DGSAI) to extract high-quality features for each modality. The baseline of 3DGSAI network is Inflated 3D ConvNet (I3D), and two main improvements are proposed. One is 3D-Ghost module, and the other is the spatial attention mechanism. The 3D-Ghost module can extract richer features for hand gesture representation, and the spatial attention mechanism makes the network pay more attention to hand region. This paper also proposes an adaptive parameter for positive knowledge transfer, which ensures that the transfer always occurs from the strong modality network to the weak one. Extensive experiments on SKIG, VIVA, and NVGesture datasets demonstrate that our method is competitive with the state of the art. Especially, the performance of our method reaches 97.87% on the SKIG dataset using only RGB, which is the current best result.

Real-Time Human Action Recognition Using Locally Aggregated Kinematic-Guided Skeletonlet and Supervised Hashing-by-Analysis Model.

3-D action recognition is referred to as the classification of action sequences which consist of 3-D skeleton joints. While many research works are devoted to 3-D action recognition, it mainly suffers from three problems: 1) highly complicated articulation; 2) a great amount of noise; and 3) low implementation efficiency. To tackle all these problems, we propose a real-time 3-D action-recognition framework by integrating the locally aggregated kinematic-guided skeletonlet (LAKS) with a supervised hashing-by-analysis (SHA) model. We first define the skeletonlet as a few combinations of joint offsets grouped in terms of the kinematic principle and then represent an action sequence using LAKS, which consists of a denoising phase and a locally aggregating phase. The denoising phase detects the noisy action data and adjusts it by replacing all the features within it with the features of the corresponding previous frame, while the locally aggregating phase sums the difference between an offset feature of the skeletonlet and its cluster center together over all the offset features of the sequence. Finally, the SHA model combines sparse representation with a hashing model, aiming at promoting the recognition accuracy while maintaining high efficiency. Experimental results on MSRAction3D, UTKinectAction3D, and Florence3DAction datasets demonstrate that the proposed method outperforms state-of-the-art methods in both recognition accuracy and implementation efficiency.

Hydrogen sulfide alleviates the anxiety-like and depressive-like behaviors of type 1 diabetic mice via inhibiting inflammation and ferroptosis.

Studies reported that sodium hydrosulfide (NaHS) can remit the depressive-like and anxiety-like behaviors induced by type 1 diabetes mellitus (T1DM). However, the mechanism is still unclear. In this study, we aimed to investigate the mechanism of NaHS on T1DM. Mice were randomly divided into four groups, including the control group (CON group), DM group, DM + 5.6 mg/kg NaHS group, and CON + 5.6 mg/kg NaHS group. Data showed that NaHS did attenuate the depressive-like and anxiety-like behaviors by OFT, EPM test, FST, and TST. Results suggest that NaHS markedly alleviated the ferroptosis in the prefrontal cortex (PFC) of diabetic mice by reducing iron deposition and oxidative stress, increasing the expression of GPX4 and SLC7A11. Moreover, NaHS could dampen the activation of microglias and the release of pro-inflammatory cytokines, enhance the protein expression of sirtuin 6 (Sirt6) and the interaction between Sirt6 and the acetylation of histoneH3 lysine9 (H3K9ac), and decrease the protein expressions of the Notch1 receptor and H3K9ac. In vitro experiment, NaHS ameliorated the ferroptosis via increasing the protein expressions of SLC7A11, glutathione peroxidase 4 (GPX4), and cystathionine ß-synthase (CBS), reducing the pro-inflammatory cytokines, decreasing the levels of Fe2+, MDA, ROS, and lipid ROS. In conclusion, our results suggested that NaHS did alleviate anxiety-like and depressive-like behaviors. It can inhibit inflammation via modulating Sirt6 and was able to decrease the ferroptosis in the PFC of type 1 diabetic mice and the BV2 cells.

Hyperforin alleviates the psychiatric disorders of adult rats suffered from early maternal separation via activating autophagy.

Rodent animals exposed to early maternal separation (EMS) show abnormal behaviors. Our previous study reported that autophagy is inhibited in the hippocampus of EMS rats, and hyperforin (HYP) alleviates depressive-like and anxious-like behaviors induced by EMS. However, the underlying mechanism of HYP is still unclear. In this study, we tested whether HYP alleviates the psychiatric disorders of EMS rats via activating autophagy. Pups were randomly divided into the control (CON) group, the EMS group, the EMS +3 mg/kg/day HYP (EMS + HYP) group and the EMS + treatment with 3 mg/kg/day fluoxetine (EMS + FT) group. Pups were separated from their mothers for 6 h every day from postnatal day 1 (PD1) to PD21 except pups of the CON group. Besides, HYP and FT were administered from PD22 to PD35 in the EMS + HYP group and the EMS + FT group respectively. Data showed that HYP not only reduced the level of glutamate, decreased the expression of N-methyl-d-aspartate receptor subunit 2B and postsynaptic density-95, but also increased the expression of synaptophysin of EMS rats. Interestingly, the expression of beclin-1 and the ratio of LC3II/LC3I were up-regulated in the EMS + HYP group. Moreover, HYP reduced the expression of the Notch1 receptor and the acetylation of H3K9 of EMS rats. In conclusion, our findings demonstrated that HYP ameliorates the depressive-like and anxious-like behaviors via activating autophagy in the hippocampus of EMS rats.

Berberine inhibits inflammatory mediators and attenuates acute pancreatitis through deactivation of JNK signaling pathways.

Acute pancreatitis (AP) is a life-threatening disease. Berberine (BBR), a well-known plant alkaloid, is reported to have anti-inflammatory activity in many diseases. However, the effects of BBR on AP have not been clearly elucidated. Therefore, the present study aimed to investigate the effects of BBR on cerulein-induced AP in mice. AP was induced by either cerulein or l-arginine. In the BBR treated group, BBR was administered intraperitoneally 1h before the first cerulein or l-arginine injection. Blood samples were obtained to determine serum amylase and lipase activities and nitric oxide production. The pancreas and lung were rapidly removed for examination of histologic changes, myeloperoxidase (MPO) activity, and real-time reverse transcription-polymerase chain reaction. Furthermore, the regulating mechanisms of BBR were evaluated. Treatment of mice with BBR reduced pancreatic injury and activities of amylase, lipase, and pancreatitis-associated lung injury, as well as inhibited several inflammatory parameters such as the expression of pro-inflammatory cytokines and inducible nitric oxide synthesis (iNOS). Furthermore, BBR administration significantly inhibited c-Jun N-terminal kinase (JNK) activation in the cerulein-induced AP. Deactivation of JNK resulted in amelioration of pancreatitis and the inhibition of inflammatory mediators. These results suggest that BBR exerts anti-inflammatory effects on AP via JNK deactivation on mild and severe acute pancreatitis model, and could be a beneficial target in the management of AP.

Clinical-Morphological Features and Outcomes of Lupus Podocytopathy.

BACKGROUND AND OBJECTIVES: Lupus podocytopathy, which is characterized by diffuse foot process effacement without peripheral capillary wall immune deposits and glomerular proliferation, has been described in SLE patients with nephrotic syndrome in case reports and small series. This study aimed to better characterize the incidence, clinical-morphologic features, and outcomes of such patients from a large Chinese cohort. DESIGN, SETTING, PARTICIPANTS, & MEASUREMENTS: Lupus podocytopathy was identified from 3750 biopsies of SLE patients obtained from 2000 to 2013 that showed mild glomerular histology in patients with a clinical sign of nephrotic syndrome. The biopsy results were divided into three groups: glomerular minimal change, mesangial proliferation, and FSGS. RESULTS: Fifty (1.33%) cases were identified as lupus podocytopathy and included minimal change in 13 cases, mesangial proliferation in 28 cases, and FSGS in nine cases. Extensive foot process effacement appeared in all the biopsies and mesangial electron-dense deposits were present in 47 biopsies. All patients demonstrated nephrotic syndrome, and the median proteinuria was 5.72 g/24 h (interquartile range [IQR], 3.82, 6.92). Seventeen (34%) cases presented with AKI. Forty-seven (94%) patients achieved remission after immunosuppressive therapy for a median time of 4 weeks (IQR, 2, 8). Compared with the patients with minimal change and mesangial proliferation, patients with FSGS showed significantly higher incidence of AKI and severe tubule-interstitial injury and a much lower complete remission rate. During follow-up of a median of 62 (IQR, 36, 84) months, renal relapses occurred in 28 (59.6%) patients. No patient died or developed ESRD. CONCLUSIONS: The findings from this cohort study suggest that lupus podocytopathy may represent a special entity of lupus nephritis with distinct clinical-morphologic features. The differences in AKI incidence, tubular injury severity, and response to treatment between the patients with minimal change/mesangial proliferation and those with FSGS patterns indicate two different subtypes of lupus podocytopathy.

Long-term outcome of IgA nephropathy patients with recurrent macroscopic hematuria.

BACKGROUND/AIMS: The long-term renal outcomes of patients with IgA nephropathy (IgAN) who present with recurrent macroscopic hematuria (RMH) have not been described in previous studies. METHODS: Patients with biopsy-proven primary IgAN in Jinling Hospital were divided into three groups according to different patterns of macroscopic hematuria (MH): RMH, isolated MH (IMH), and those without a history of MH (NMH). RESULTS: A total of 1,155 patients were enrolled in the study (158 in the RMH group, 256 in the IMH group, and 741 in the NMH group). At biopsy, patients with RMH were younger, had lower median proteinuria, a lower incidence of hypertension, and a higher estimated glomerular filtration rate than those in the NMH group. Pathologically, patients with RMH had a lower level of mesangial hypercellularity and segmental glomerulosclerosis as well as less tubular atrophy than those with NMH. The demographic and clinical features of patients with IMH fell between patients with RMH and those with NMH. During a median follow-up of 7.9 years, the 5-, 10- and 20-year cumulative renal survival after biopsy, as calculated by K-M methods, were 98, 91, and 91% in the RMH group, 95, 89, and 64% in the IMH group, and 95, 79, and 57% in the NMH group. The renal survival in patients with RMH was significantly better than patients with NMH or IMH. CONCLUSIONS: The long-term prognosis of patients who present with RMH is significantly better than patients with NMH or IMH.