Label Prediction Research Articles

Proposal Semantic Relationship Graph Network for Temporal Action Detection

Temporal action detection, a critical task in video activity understanding, is typically divided into two stages: proposal generation and classification. However, most existing methods overlook the importance of information transfer among proposals during classification, often treating each proposal in isolation, which hampers accurate label prediction. In this paper, we propose a novel method for inferring semantic relationships both within and between action proposals, guiding the fusion of action proposal features accordingly. Building on this approach, we introduce the Proposal Semantic Relationship Graph Network (PSRGN), an end-to-end model that leverages intra-proposal semantic relationship graphs to extract cross-scale temporal context and an inter-proposal semantic relationship graph to incorporate complementary neighboring information, significantly improving proposal feature quality and overall detection performance. This is the first method to apply graph structure learning in temporal action detection, adaptively constructing the inter-proposal semantic graph. Extensive experiments on two datasets demonstrate the effectiveness of our approach, achieving state-of-the-art results. Code and results are available at http://github.com/Riiick2011/PSRGN .

A Semantically Enhanced Label Prediction Method for Imbalanced POI Data Category Distribution

A Semantically Enhanced Label Prediction Method for Imbalanced POI Data Category Distribution

AML-DECODER: Advanced Machine Learning for HD-sEMG Signal Classification-Decoding Lateral Epicondylitis in Forearm Muscles.

Innovative algorithms for wearable devices and garments are critical for diagnosing and monitoring disease (such as lateral epicondylitis (LE)) progression. LE affects individuals across various professions and causes daily problems. We analyzed signals from the forearm muscles of 14 healthy controls and 14 LE patients using high-density surface electromyography. We discerned significant differences between groups by employing phase-amplitude coupling (PAC) features. Our study leveraged PAC, Daubechies wavelet with four vanishing moments (db4), and state-of-the-art techniques to train a neural network for the subject's label prediction. Remarkably, PAC features achieved 100% specificity and sensitivity in predicting unseen subjects, while state-of-the-art features lagged with only 35.71% sensitivity and 28.57% specificity, and db4 with 78.57% sensitivity and 85.71 specificity. PAC significantly outperformed the state-of-the-art features (adj. p-value < 0.001) with a large effect size. However, no significant difference was found between PAC and db4 (adj. p-value = 0.147). Also, the Jeffries-Matusita (JM) distance of the PAC was significantly higher than other features (adj. p-value < 0.001), with a large effect size, suggesting PAC features as robust predictors of neuromuscular diseases, offering a profound understanding of disease pathology and new avenues for interpretation. We evaluated the generalization ability of the PAC model using 99.9% confidence intervals and Bayesian credible intervals to quantify prediction uncertainty across subjects. Both methods demonstrated high reliability, with an expected accuracy of 89% in larger, more diverse populations. This study's implications might extend beyond LE, paving the way for enhanced diagnostic tools and deeper insights into the complexities of neuromuscular disorders.

Exploring key factors for long-term vessel incident risk prediction

Factor analysis acts a pivotal role in enhancing maritime safety. Most previous studies conduct factor analysis within the framework of incident-related label prediction, where the developed models can be categorized into short-term and long-term prediction models. The long-term models offer a more strategic approach, enabling more proactive risk management, compared to the short-term ones. Nevertheless, few studies have devoted to rigorously identifying the key factors for the long-term prediction and undertaking comprehensive factor analysis. Hence, this study aims to delve into the key factors for predicting the incident risk levels in the subsequent year given a specific datestamp. The majority of candidate factors potentially contributing to the incident risk are collected from vessels’ historical safety performance data spanning up to five years. An improved embedded feature selection method, which integrates Random Forest classifier with a feature filtering process, is proposed to identify key risk-contributing factors from the candidate pool. A dataset with information of 131,652 vessels collected from 2015 to 2023 is utilized for case study. The results demonstrate superior performances of the proposed method in incident prediction and factor interpretability. Comprehensive analysis is conducted upon the key factors, which could help maritime stakeholders formulate management strategies for incident prevention.

Robust brain MRI image classification with SIBOW-SVM

Primary Central Nervous System tumors in the brain are among the most aggressive diseases affecting humans. Early detection and classification of brain tumor types, whether benign or malignant, glial or non-glial, is critical for cancer prevention and treatment, ultimately improving human life expectancy. Magnetic Resonance Imaging (MRI) is the most effective technique for brain tumor detection, generating comprehensive brain scans. However, human examination can be error-prone and inefficient due to the complexity, size, and location variability of brain tumors. Recently, automated classification techniques using machine learning methods, such as Convolutional Neural Networks (CNNs), have demonstrated significantly higher accuracy than manual screening. However, deep learning-based image classification methods, including CNNs, face challenges in estimating class probabilities without proper model calibration (Guo et al., 2017; Minderer et al., 2021). In this paper, we propose a novel brain tumor image classification method called SIBOW-SVM, which integrates the Bag-of-Features model with SIFT feature extraction and weighted Support Vector Machines. This new approach can effectively extract hidden image features, enabling differentiation of various tumor types, provide accurate label predictions, and estimate probabilities of images belonging to each class, offering high-confidence classification decisions. We have also developed scalable and parallelable algorithms to facilitate the practical implementation of SIBOW-SVM for massive image datasets. To benchmark our method, we apply SIBOW-SVM to a public dataset of brain tumor MRI images containing four classes: glioma, meningioma, pituitary, and normal. Our results demonstrate that the new method outperforms state-of-the-art techniques, including CNNs, in terms of uncertianty quantification, classification accuracy, computational efficiency, and data robustness.

An Industrial Micro Parts Recognition Technology Based on Improved Yolov8

Abstract Aiming at the recognition and classification problem of micro assembly parts in industrial production, the Yolov8 object detection algorithm was used as the base model to improve the C2f class in the backbone. And apply the improved object detection algorithm to the recognition field of detecting assembly parts. By collecting images of gears, nails, and screws from various angles on Roboflow, a training set and a validation set containing three types of gears, nails, and screws are formed. At the same time, in order to improve the robustness of the Yolov8 object detection algorithm, data processing methods such as adding noise and image rotation were adopted for the samples in the data. Due to the use of Yolov8 as the base model, a TXT text file containing label prediction box position information was formed after labelling the dataset on Roboflow. Based on Python (3.10.12) and cloud platform, improved Yolov8 object detection algorithm was trained and network validation was carried out. Use TXT files as input for the model to obtain and learn features of gears, nails, and screws. After 200 epochs, the mAP for detecting gears, nails, and screws increased by about 3% compared to the base model.

Graph-Based Semi-Supervised Learning with Bipartite Graph for Large-Scale Data and Prediction of Unseen Data

Recently, considerable attention has been directed toward graph-based semi-supervised learning (GSSL) as an effective approach for data labeling. Despite the progress achieved by current methodologies, several limitations persist. Firstly, many studies treat all samples equally in terms of weight and influence, disregarding the potential increased importance of samples near decision boundaries. Secondly, the detection of outlier-labeled data is crucial, as it can significantly impact model performance. Thirdly, existing models often struggle with predicting labels for unseen test data, restricting their utility in practical applications. Lastly, most graph-based algorithms rely on affinity matrices that capture pairwise similarities across all data points, thus limiting their scalability to large-scale databases. In this paper, we propose a novel GSSL algorithm tailored for large-scale databases, leveraging anchor points to mitigate the challenges posed by large affinity matrices. Additionally, our method enhances the influence of nodes near decision boundaries by assigning different weights based on their importance and using a mapping function from feature space to label space. Leveraging this mapping function enables direct label prediction for test samples without requiring iterative learning processes. Experimental evaluations on two extensive datasets (Norb and Covtype) demonstrate that our approach is scalable and outperforms existing GSSL methods in terms of performance metrics.

Deep dual incomplete multi-view multi-label classification via label semantic-guided contrastive learning

Multi-view multi-label learning (MVML) aims to train a model that can explore the multi-view information of the input sample to obtain its accurate predictions of multiple labels. Unfortunately, a majority of existing MVML methods are based on the assumption of data completeness, making them useless in practical applications with partially missing views or some uncertain labels. Recently, many approaches have been proposed for incomplete data, but few of them can handle the case of both missing views and labels. Moreover, these few existing works commonly ignore potentially valuable information about unknown labels or do not sufficiently explore latent label information. Therefore, in this paper, we propose a label semantic-guided contrastive learning method named LSGC for the dual incomplete multi-view multi-label classification problem. Concretely, LSGC employs deep neural networks to extract high-level features of samples. Inspired by the observation of exploiting label correlations to improve the feature discriminability, we introduce a graph convolutional network to effectively capture label semantics. Furthermore, we introduce a new sample-label contrastive loss to explore the label semantic information and enhance the feature representation learning. For missing labels, we adopt a pseudo-label filling strategy and develop a weighting mechanism to explore the confidently recovered label information. We validate the framework on five standard datasets and the experimental results show that our method achieves superior performance in comparison with the state-of-the-art methods.

A multi-view ensemble machine learning approach for 3D modeling using geological and geophysical data

Geophysical data are often integrated into geological data for 3D modeling of underground spaces. However, the existing single-view approach means it is difficult to adequately fuse the valid information between the two types of data, and the complexity of lithological decoding and classification is high. To address this issue, a multi-view ensemble machine learning (ML) framework is proposed. Initially, the original dataset of lithology prediction is constructed by aligning geological and geophysical data with different spatial scales. Next, the dataset is divided into three datasets of structural strength, density, and moisture content according to the lithology properties of the geophysical data. The proposed framework is then used to capture the lithologic characteristics under different views to achieve the prediction of lithologic labels. In this process, a self-attentive mechanism is used to adaptively fuse the valid information under each view. To validate the proposed framework, it is applied to a project in Jiaxing, Zhejiang Province, China. Compared with existing ML methods, the proposed multi-view ensemble ML framework improves modeling accuracy and constructs models with low uncertainty. The framework can be extended to other multi-source data fusion tasks across geoscience domains.

Deconfounded hierarchical multi-granularity classification

Hierarchical multi-granularity classification (HMC) assigns labels at varying levels of detail to images using a structured hierarchy that categorizes labels from coarse to fine, such as [“Suliformes”, “Fregatidae”, “Frigatebird”]. Traditional HMC methods typically integrate hierarchical label information into either the model’s architecture or its loss function. However, these approaches often overlook the spurious correlations between coarse-level semantic information and fine-grained labels, which can lead models to rely on these non-causal relationships for making predictions. In this paper, we adopt a causal perspective to address the challenges in HMC, demonstrating how coarse-grained semantics can serve as confounders in fine-grained classification. To comprehensively mitigate confounding bias in HMC, we introduce a novel framework, Deconf-HMC, which consists of three main components: (1) a causal-inspired label prediction module that combines fine-level features with coarse-level prediction outcomes to determine the appropriate labels at each hierarchical level; (2) a representation disentanglement module that minimizes the mutual information between representations of different granularities; and (3) an adversarial training module that restricts the predictive influence of coarse-level representations on fine-level labels, thereby aiming to eliminate confounding bias. Extensive experiments on three widely used datasets demonstrate the superiority of our approach over existing state-of-the-art HMC methods.

MLAWSMOTE: Oversampling in Imbalanced Multi-label Classification with Missing Labels by Learning Label Correlation Matrix

Missing labels in multi-label datasets are a common problem, especially for minority classes, which are more likely to occur. This limitation hinders the performance of classifiers in identifying and extracting information from minority classes. Oversampling is an effective method for addressing imbalanced multi-label problems by generating synthetic instances to create a class-balanced dataset. However, the existing oversampling algorithms mainly focus on the location of the generated data, and there is a lack of design on how to complete the labels of the synthetic data. To address this issue, we propose MLAWSMOTE, a synthetic data generation algorithm based on matrix factorization weights. We introduce a weak supervised learning method in the oversampling method, optimize the weights of features and labels by using label correlation, and iteratively learn the ideal label weights. The mapping relationship between features and labels is learned from the dataset and the label correlation matrix. The oversampling ratio is defined based on the discrepancy between observed labels and the ideal label of synthetic instances. It mitigates the impact of missing minority labels on the model’s predictions. The labeling of synthetic instances is performed based on label prediction, and the potential labeling distribution is complemented. Experimental results on multiple multi-label datasets under different label missing ratios demonstrate the effectiveness of the proposed method in terms of ACC, Hamming loss, MacroF1 and MicroF1. In the validation of the four classifiers, MacroF1 decreased by 24.78%, 17.81%, 3.8% and 19.56%, respectively, with the increase of label loss rate. After applying MLAWSMOTE only decreased by 15.79%, 13.63%, 3.78% and 15.21%.

Towards Better Modeling With Missing Data: A Contrastive Learning-Based Visual Analytics Perspective.

Missing data can pose a challenge for machine learning (ML) modeling. To address this, current approaches are categorized into feature imputation and label prediction and are primarily focused on handling missing data to enhance ML performance. These approaches rely on the observed data to estimate the missing values and therefore encounter three main shortcomings in imputation, including the need for different imputation methods for various missing data mechanisms, heavy dependence on the assumption of data distribution, and potential introduction of bias. This study proposes a Contrastive Learning (CL) framework to model observed data with missing values, where the ML model learns the similarity between an incomplete sample and its complete counterpart and the dissimilarity between other samples. Our proposed approach demonstrates the advantages of CL without requiring any imputation. To enhance interpretability, we introduce CIVis, a visual analytics system that incorporates interpretable techniques to visualize the learning process and diagnose the model status. Users can leverage their domain knowledge through interactive sampling to identify negative and positive pairs in CL. The output of CIVis is an optimized model that takes specified features and predicts downstream tasks. We provide two usage scenarios in regression and classification tasks and conduct quantitative experiments, expert interviews, and a qualitative user study to demonstrate the effectiveness of our approach. In short, this study offers a valuable contribution to addressing the challenges associated with ML modeling in the presence of missing data by providing a practical solution that achieves high predictive accuracy and model interpretability.

Stable isotope labeling and functional gene prediction elucidate the carbon metabolism in fermentative bacteria and microalgae coupling system

The application of the fermentative bacteria and microalgae coupling system in the wastewater treatment has been studied, but there remains few knowledge regarding the organic and inorganic carbon metabolism within this system. In this study, the carbon metabolism of microalgae and fermentative bacteria was elucidated by 13C stable isotope labeling and functional gene prediction, respectively. The 13C glucose and 13C NaHCO3 were used as stable isotope tracers to clarify the organic and inorganic carbon metabolism of microalgae, indicating that approximately 71.5 % of the Acetyl-CoA in microalgae was synthesized from organic carbon sources, while 26.8 % was synthesized through the utilization of inorganic carbon sources. Inorganic carbon sources can enhance the activity of photosynthetic system and facilitate the Calvin cycle. Considering the adequate organic carbon sources and insufficient inorganic carbon sources in the fermentative bacteria and microalgae coupling system, NaHCO3 was added to improve carbon utilization of microalgae. The maximum microalgal lipid yield reached 1130.37 mg/L with 1000 mg/L NaHCO3 supplementation. Functional gene prediction was used to analysis the effect of various carbon composition on the bacterial carbon metabolism. Notably, the additional inorganic carbon sources increased the abundance of bacterial functional genes associated with the fermentation and acetic acids synthesis, which was advantageous for VFAs production and further promoted microalgae growth. This study can gain a deeper understanding of microbial metabolic mechanisms during the operation of fermentative bacteria and microalgae system, and improve its sustained operational stability.

Semi-Supervised Learning for Multi-View Data Classification and Visualization

Data visualization has several advantages, such as representing vast amounts of data and visually demonstrating patterns within it. Manifold learning methods help us estimate lower-dimensional representations of data, thereby enabling more effective visualizations. In data analysis, relying on a single view can often lead to misleading conclusions due to its limited perspective. Hence, leveraging multiple views simultaneously and interactively can mitigate this risk and enhance performance by exploiting diverse information sources. Additionally, incorporating different views concurrently during the graph construction process using interactive visualization approach has improved overall performance. In this paper, we introduce a novel algorithm for joint consistent graph construction and label estimation. Our method simultaneously constructs a unified graph and predicts the labels of unlabeled samples. Furthermore, the proposed approach estimates a projection matrix that enables the prediction of labels for unseen samples. Moreover, it incorporates the information in the label space to further enhance the accuracy. In addition, it merges the information in different views along with the labels to construct a consensus graph. Experimental results conducted on various image databases demonstrate the superiority of our fusion approach compared to using a single view or other fusion algorithms. This highlights the effectiveness of leveraging multiple views and simultaneously constructing a unified graph for improved performance in data classification and visualization tasks in semi-supervised contexts.

Harnessing Fusion Modeling for Enhanced Breast Cancer Classification through Interpretable Artificial Intelligence and In-Depth Explanations

Integrating Artificial Intelligence (AI) into healthcare has shown tremendous promise in several domains, such as prediction, decision-making, and diagnosis. Nevertheless, the widespread belief that AI is incomprehensible and mysterious has restricted its practical use, particularly in medicine. This skepticism highlights the need for methods that provide precise predictions and clear explanations for these projections. The present study presents a hybrid method “BC2” (Breast Cancer Classification) that combines deep learning techniques to accurately detect breast cancer and provide detailed explanations for its predictions. “BC2” exploits the collaborative effects of Convolutional Neural Networks (CNN) and the Light Gradient Boosting Model (LGBM). The automated feature learning process of the model is integrated into many convolutional layers, and LGBM then follows it to provide predictions for class labels. To evaluate the effectiveness of the proposed approach, the authors performed experiments with real-world Breast Cancer Data. The results demonstrate substantial enhancements compared to the current methods in key metrics such as accuracy, precision, recall, and F1 score. The proposed technique obtains an accuracy of 0.982906, a precision of 0.987179, a recall of 0.987179, and an F1-score of 0.987179, demonstrating its exceptional performance. After generating exact forecasts, authors use the “SHAP” (SHapley Additive exPlanation) approach at local and global levels to explain the reasoning behind each projection. This guarantees clarity and improves the comprehensibility of the predictions made by the proposed model. The proposed hybrid model consists of three main components: a feature learning module, a diagnostic module, and an explanation-generating module. The proposed model can improve medical professionals’ diagnostic skills in managing breast cancer patients by clearly explaining its forecasts.

Lightweight bearing fault diagnosis method based on cross-scale learning transformer under imbalanced data*

Abstract Due to the limited amount of failure data in rolling bearing faults, traditional fault diagnosis models encounter challenges such as low diagnostic accuracy and efficiency when dealing with imbalanced data. Additionally, many fault diagnosis models are overly complex and demand high computational resources. To address these issues, a lightweight bearing fault diagnosis method based on cross-scale learnable transformer (CSLT) is proposed for imbalanced data. For difficult-to-classify samples, a learnable generalized focal loss function is defined. The learnable parameters are employed to increase its flexibility, it better addresses the issue of bearing fault diagnosis under imbalanced data conditions. Then, a multi-head broadcasted self-attention mechanism is designed by capturing critical local features of the signal through one-dimensional convolution operations, which not only improves feature extraction capability but also reduces computational complexity. Finally, a dynamic label prediction pruning module is developed to trim redundant labels, which helps in lightening the model and enhancing both feature extraction and diagnostic efficiency. The experimental results demonstrate that the proposed diagnosis method exhibits superior diagnostic precision and efficiency by comparing with other methods.

A Lightweight Multi-Label Classification Method for Urban Green Space in High-Resolution Remote Sensing Imagery

Urban green spaces are an indispensable part of the ecology of cities, serving as the city’s “purifier” and playing a crucial role in promoting sustainable urban development. Therefore, the refined classification of urban green spaces is an important task in urban planning and management. Traditional methods for the refined classification of urban green spaces heavily rely on expert knowledge, often requiring substantial time and cost. Hence, our study presents a multi-label image classification model based on MobileViT. This model integrates the Triplet Attention module, along with the LSTM module, to enhance its label prediction capabilities while maintaining its lightweight characteristic for standalone operation on mobile devices. Trial outcomes in our UGS dataset in this study demonstrate that the approach we used outperforms the baseline by 1.64%, 3.25%, 3.67%, and 2.71% in mAP,F1,precision, and recall, respectively. This indicates that the model can uncover the latent dependencies among labels to enhance the multi-label image classification device’s performance. This study provides a practical solution for the intelligent and detailed classification of urban green spaces, which holds significant importance for the management and planning of urban green spaces.

O-274 An Artificial Intelligence (AI) based screening test can reliably predict embryo ploidy

Abstract Study question Can calibrated AI ploidy screening test results provide reliable, biologically-justified estimates of embryo euploidy? Summary answer The AI ploidy screening test, rooted in the embryo’s morphokinetic profile and clinical metadata, provides reliable probabilistic estimates of euploidy. What is known already Published ploidy algorithms typically provide a binary classification of embryo ploidy (aneuploidy/euploidy). Algorithmic outputs are thresholded; a value above/below the threshold indicates a euploid/aneuploid label, respectively, and the confidence in the label prediction is tested. Experience shows, however, that decision-makers have difficulty interpreting how well these algorithm predictions match the true prevalence of euploidy in their clinics, especially when taking into consideration patient age and embryo quality. An AI embryo ploidy screening tool that uses biologically-relevant inputs to provide reliable euploidy probabilities is needed. Study design, size, duration The AI tool was trained on 5,000 time-lapse video sequences, along with associated clinical parameters: biological age at time of retrieval, Day-5 embryo quality, and morphokinetic parameters ranging from time of pronuclear fading to time to blastulation. Probability calibration was applied and its performance was evaluated using a blind test dataset (N = 708 embryos; euploid=352; aneuploid=356) with known prevalences of euploidy, aneuploidy, and live-birth. Mean ± SD patient age: 35.9 ± 5.4 years. Participants/materials, setting, methods The AI ploidy screening tool used known embryo ploidy status as ground-truth labels; biological age and visual quality parameters were incorporated as continuous input features to maintain biological validity. Reliability curve analysis, which plots the observed frequencies of euploidy in the clinical input data (y-axis) against the predicted probability frequencies by the screening test (x-axis), was used to assess model confidence. Odds ratios (OR) confirmed significance between associations. Main results and the role of chance Logistic regression analysis shows that AI scores are robustly associated with euploidy probability (OR = 2.79 [95% CI = 2.04-3.81] at a threshold of 0.5 when comparing euploid likelihood for high-versus-low AI scores). For embryo cohorts in the blind test set containing ≥1 aneuploid and euploid embryos in the test dataset, (N = 57 cohorts), the highest AI ranked embryo was euploid in 64% of the cohorts. Embryos were divided into four predefined brackets according to their scores (1-32, 33-49, 50-66, 67-99) and euploidy rate per bracket was determined: 28%, 44%, 58%, 71%, respectively. There was a linear association between ascending AI scores and percentage of euploid embryos, with the highest level of model confidence achieved at the tail ends of the scalar; embryos with a score above 66 were 2.5X more likely to be euploid than an embryo with a score below 33. Limitations, reasons for caution This analysis used historical time-lapse sequences. Moving forward, we must evaluate prospective AI use for ploidy screening. Genetic status in utero/birth was not evaluated. Wider implications of the findings A novel AI approach for preimplantation embryo screening using clinical metadata and time-lapse videos can improve our ability to non-invasively predict euploid likelihood prior to confirmatory diagnostic preimplantation genetic testing. Trial registration number Not applicable

Graph-Graph Similarity Network.

Graph learning aims to predict the label for an entire graph. Recently, graph neural network (GNN)-based approaches become an essential strand to learning low-dimensional continuous embeddings of entire graphs for graph label prediction. While GNNs explicitly aggregate the neighborhood information and implicitly capture the topological structure for graph representation, they ignore the relationships among graphs. In this article, we propose a graph-graph (G2G) similarity network to tackle the graph learning problem by constructing a SuperGraph through learning the relationships among graphs. Each node in the SuperGraph represents an input graph, and the weights of edges denote the similarity between graphs. By this means, the graph learning task is then transformed into a classical node label propagation problem. Specifically, we use an adversarial autoencoder to align embeddings of all the graphs to a prior data distribution. After the alignment, we design the G2G similarity network to learn the similarity between graphs, which functions as the adjacency matrix of the SuperGraph. By running node label propagation algorithms on the SuperGraph, we can predict the labels of graphs. Experiments on five widely used classification benchmarks and four public regression benchmarks under a fair setting demonstrate the effectiveness of our method.

Dual separated attention-based graph neural network

Graph neural networks (GNNs) offer a viable solution to model the inter-dependencies among labeled and unlabeled samples in a semi-supervised manner. However, their performances can degrade dramatically when the number of labels is extremely limited, which is due to the limitations of typical graph convolutional network design in most existing GNNs, including over-smoothing, difficulty in extending the propagation step, and failure to preserve the distinctiveness of each node. To address the issues, we propose a dual separated attention-based graph neural network (DSA-GNN) to deal with label scarcity in semi-supervised node classification. Firstly, DSA-GNN decouples feature propagation from representation transformation to alleviate the problems of over-smoothing and overfitting. Secondly, DSA-GNN separates self-representation learning from neighbor-representation learning by two feature extractors with different learnable parameters. As a result, the commonality between connected nodes and the distinctiveness of each node can be both preserved. Thirdly, DSA-GNN incorporates an attention-based label propagation mechanism to refine the label prediction of each node, by aggregating label prediction among the neighborhood based on adaptive edge coefficients. The extensive experimental results on the real-world datasets demonstrate the superiority of DSA-GNN for semi-supervised node classification, especially when the observed labels are extremely limited.