Prototype Learning for Medical Time Series Classification via Human-Machine Collaboration.

Deep neural networks must address the dual challenge of delivering high-accuracy predictions and providing user-friendly explanations. While deep models are widely used in the field of time series modeling, deciphering the core principles that govern the models' outputs remains a significant challenge. This is crucial for fostering the development of trusted models and facilitating domain expert validation, thereby empowering users and domain experts to utilize them confidently in high-risk decision-making contexts (e.g., decision-support systems in healthcare). In this work, we put forward a deep prototype learning model that supports interpretable and manipulable modeling and classification of medical time series (i.e., ECG signal). Specifically, we first optimize the representation of single heartbeat data by employing a bidirectional long short-term memory and attention mechanism, and then construct prototypes during the training phase. The final classification outcomes (i.e., normal sinus rhythm, atrial fibrillation, and other rhythm) are determined by comparing the input with the obtained prototypes. Moreover, the proposed model presents a human-machine collaboration mechanism, allowing domain experts to refine the prototypes by integrating their expertise to further enhance the model's performance (contrary to the human-in-the-loop paradigm, where humans primarily act as supervisors or correctors, intervening when required, our approach focuses on a human-machine collaboration, wherein both parties engage as partners, enabling more fluid and integrated interactions). The experimental outcomes presented herein delineate that, within the realm of binary classification tasks-specifically distinguishing between normal sinus rhythm and atrial fibrillation-our proposed model, albeit registering marginally lower performance in comparison to certain established baseline models such as Convolutional Neural Networks (CNNs) and bidirectional long short-term memory with attention mechanisms (Bi-LSTMAttns), evidently surpasses other contemporary state-of-the-art prototype baseline models. Moreover, it demonstrates significantly enhanced performance relative to these prototype baseline models in the context of triple classification tasks, which encompass normal sinus rhythm, atrial fibrillation, and other rhythm classifications. The proposed model manifests a commendable prediction accuracy of 0.8414, coupled with macro precision, recall, and F1-score metrics of 0.8449, 0.8224, and 0.8235, respectively, achieving both high classification accuracy as well as good interpretability.

Time-ordering japonica/geng genomes analysis indicates the importance of large structural variants in rice breeding.

Temperate japonica/geng (GJ) rice yield has significantly improved due to intensive breeding efforts, dramatically enhancing global food security. However, little is known about the underlying genomic structural variations (SVs) responsible for this improvement. We compared 58 long-read assemblies comprising cultivated and wild rice species in the present study, revealing 156 319 SVs. The phylogenomic analysis based on the SV dataset detected the putatively selected region of GJ sub-populations. A significant portion of the detected SVs overlapped with genic regions were found to influence the expression of involved genes inside GJ assemblies. Integrating the SVs and causal genetic variants underlying agronomic traits into the analysis enables the precise identification of breeding signatures resulting from complex breeding histories aimed at stress tolerance, yield potential and quality improvement. Further, the results demonstrated genomic and genetic evidence that the SV in the promoter of LTG1 is accounting for chilling sensitivity, and the increased copy numbers of GNP1 were associated with positive effects on grain number. In summary, the current study provides genomic resources for retracing the properties of SVs-shaped agronomic traits during previous breeding procedures, which will assist future genetic, genomic and breeding research on rice.

Enabling Timely Medical Intervention by Exploring Health-Related Multivariate Time Series with a Hybrid Attentive Model.

Modern healthcare practice, especially in intensive care units, produces a vast amount of multivariate time series of health-related data, e.g., multi-lead electrocardiogram (ECG), pulse waveform, blood pressure waveform and so on. As a result, timely and accurate prediction of medical intervention (e.g., intravenous injection) becomes possible, by exploring such semantic-rich time series. Existing works mainly focused on onset prediction at the granularity of hours that was not suitable for medication intervention in emergency medicine. This research proposes a Multi-Variable Hybrid Attentive Model (MVHA) to predict the impending need of medical intervention, by jointly mining multiple time series. Specifically, a two-level attention mechanism is designed to capture the pattern of fluctuations and trends of different time series. This work applied MVHA to the prediction of the impending intravenous injection need of critical patients at the intensive care units. Experiments on the MIMIC Waveform Database demonstrated that the proposed model achieves a prediction accuracy of 0.8475 and an ROC-AUC of 0.8318, which significantly outperforms baseline models.

Simultaneous Improvement of Grain Yield and Quality through Manipulating Two Type C G Protein Gamma Subunits in Rice.

Heterotrimeric G protein signaling is an evolutionarily conserved mechanism in diverse organisms that mediates intracellular responses to external stimuli. In rice, the G proteins are involved in the regulation of multiple important agronomic traits. In this paper, we present our finding that two type C G protein gamma subunits, DEP1 and GS3, antagonistically regulated grain yield and grain quality. The DEP1 gene editing we conducted, significantly increased the grain number per panicle but had a negative impact on taste value, texture properties, and chalkiness-related traits. The GS3 gene editing decreased grain number per panicle but significantly increased grain length. In addition, the GS3 gene-edited plants showed improved taste value, appearance, texture properties, and Rapid Visco Analyser (RVA) profiles. To combine the advantages of both gs3 and dep1, we conducted a molecular design breeding at the GS3 locus of a "super rice" variety, SN265, which has a truncated dep1 allele with erect panicle architecture, high-yield performance, and which is of mediocre eating quality. The elongated grain size of the sn265/gs3 gene-edited plants further increased the grain yield. More importantly, the texture properties and RVA profiles were significantly improved, and the taste quality was enhanced. Beyond showcasing the combined function of dep1 and gs3, this paper presents a strategy for the simultaneous improvement of rice grain yield and quality through manipulating two type C G protein gamma subunits in rice.

Investigation of the determinants for misinformation correction effectiveness on social media during COVID-19 pandemic.

The rapid dissemination of misinformation in social media during the COVID-19 pandemic triggers panic and threatens the pandemic preparedness and control. Correction is a crucial countermeasure to debunk misperceptions. However, the effective mechanism of correction on social media is not fully verified. Previous works focus on psychological theories and experimental studies, while the applicability of conclusions to the actual social media is unclear. This study explores determinants governing the effectiveness of misinformation corrections on social media with a combination of a data-driven approach and related theories on psychology and communication. Specifically, referring to the Backfire Effect, Source Credibility, and Audience's role in dissemination theories, we propose five hypotheses containing seven potential factors (regarding correction content and publishers' influence), e.g., the proportion of original misinformation and warnings of misinformation. Then, we obtain 1487 significant COVID-19 related corrections on Microblog between January 1st, 2020 and April 30th, 2020, and conduct annotations, which characterize each piece of correction based on the aforementioned factors. We demonstrate several promising conclusions through a comprehensive analysis of the dataset. For example, mentioning excessive original misinformation in corrections would not undermine people's believability within a short period after reading; warnings of misinformation in a demanding tone make correction worse; determinants of correction effectiveness vary among different topics of misinformation. Finally, we build a regression model to predict correction effectiveness. These results provide practical suggestions on misinformation correction on social media, and a tool to guide practitioners to revise corrections before publishing, leading to ideal efficacies.

Curcumin upregulates the Nrf2 system by repressing inflammatory signaling-mediated Keap1 expression in insulin-resistant conditions.

Both oxidative stress and inflammation contribute to the development of insulin resistance (IR). Curcumin (Cur) not only has an anti-inflammatory effect but also has an antioxidative stress effect via the activation of NF-E2-related factor 2 (Nrf2). Since there is close cross-communication between inflammation and oxidative stress, we examined whether Cur could modulate Nrf2 function via its anti-inflammatory ability and investigated its underlying mechanism. In this study, we show that Cur inhibits inflammatory signaling and Kelch-like ECH-associated protein 1 (Keap1) expression, which is accompanied by the activation of the Nrf2 system. We further identified that the proinflammatory cytokine tumor necrosis factor alpha (TNFα) could stimulate Keap1 synthesis and increase Nrf2 polyubiquitination, but these effects could be significantly inhibited by Cur treatment. This study demonstrates that Cur-induced Nrf2 activation occurs through the inhibition of inflammatory signaling-mediated upregulation of Keap1, contributing to its beneficial effects on redox homeostasis and insulin sensitivity.

Metformin exerts glucose-lowering action in high-fat fed mice via attenuating endotoxemia and enhancing insulin signaling.

AIM: Accumulating evidence shows that lipopolysaccharides (LPS) derived from gut gram-negative bacteria can be absorbed, leading to endotoxemia that triggers systemic inflammation and insulin resistance. In this study we examined whether metformin attenuated endotoxemia, thus improving insulin signaling in high-fat diet fed mice. METHODS: Mice were fed a high-fat diet for 18 weeks to induce insulin resistance. One group of the mice was treated with oral metformin (100 mg·kg(-1)·d(-1)) for 4 weeks. Another group was treated with LPS (50 µg·kg(-1)·d(-1), sc) for 5 days followed by the oral metformin for 10 d. Other two groups received a combination of antibiotics for 7 d or a combination of antibiotics for 7 d followed by the oral metformin for 4 weeks, respectively. Glucose metabolism and insulin signaling in liver and muscle were evaluated, the abundance of gut bacteria, gut permeability and serum LPS levels were measured. RESULTS: In high-fat fed mice, metformin restored the tight junction protein occludin-1 levels in gut, reversed the elevated gut permeability and serum LPS levels, and increased the abundance of beneficial bacteria Lactobacillus and Akkermansia muciniphila. Metformin also increased PKB Ser473 and AMPK T172 phosphorylation, decreased MDA contents and redox-sensitive PTEN protein levels, activated the anti-oxidative Nrf2 system, and increased IκBα in liver and muscle of the mice. Treatment with exogenous LPS abolished the beneficial effects of metformin on glucose metabolism, insulin signaling and oxidative stress in liver and muscle of the mice. Treatment with antibiotics alone produced similar effects as metformin did. Furthermore, the beneficial effects of antibiotics were addictive to those of metformin. CONCLUSION: Metformin administration attenuates endotoxemia and enhances insulin signaling in high-fat fed mice, which contributes to its anti-diabetic effects.

FFA-induced hepatic insulin resistance in vivo is mediated by PKCδ, NADPH oxidase, and oxidative stress.

Fat-induced hepatic insulin resistance plays a key role in the pathogenesis of type 2 diabetes in obese individuals. Although PKC and inflammatory pathways have been implicated in fat-induced hepatic insulin resistance, the sequence of events leading to impaired insulin signaling is unknown. We used Wistar rats to investigate whether PKCδ and oxidative stress play causal roles in this process and whether this occurs via IKKß- and JNK-dependent pathways. Rats received a 7-h infusion of Intralipid plus heparin (IH) to elevate circulating free fatty acids (FFA). During the last 2 h of the infusion, a hyperinsulinemic-euglycemic clamp with tracer was performed to assess hepatic and peripheral insulin sensitivity. An antioxidant, N-acetyl-L-cysteine (NAC), prevented IH-induced hepatic insulin resistance in parallel with prevention of decreased IκBα content, increased JNK phosphorylation (markers of IKKß and JNK activation, respectively), increased serine phosphorylation of IRS-1 and IRS-2, and impaired insulin signaling in the liver without affecting IH-induced hepatic PKCδ activation. Furthermore, an antisense oligonucleotide against PKCδ prevented IH-induced phosphorylation of p47(phox) (marker of NADPH oxidase activation) and hepatic insulin resistance. Apocynin, an NADPH oxidase inhibitor, prevented IH-induced hepatic and peripheral insulin resistance similarly to NAC. These results demonstrate that PKCδ, NADPH oxidase, and oxidative stress play a causal role in FFA-induced hepatic insulin resistance in vivo and suggest that the pathway of FFA-induced hepatic insulin resistance is FFA → PKCδ → NADPH oxidase and oxidative stress → IKKß/JNK → impaired hepatic insulin signaling.

Adaptive Memory Broad Learning System for Unsupervised Time Series Anomaly Detection.

Time series anomaly detection is the process of identifying anomalies within time series data. The primary challenge of this task lies in the necessity for the model to comprehend the characteristics of time-independent and abnormal data patterns. In this study, a novel algorithm called adaptive memory broad learning system (AdaMemBLS) is proposed for time series anomaly detection. This algorithm leverages the rapid inference capabilities of the broad learning algorithm and the memory bank's capacity to differentiate between normal and abnormal data. Furthermore, an incremental algorithm based on multiple data augmentation techniques is introduced and applied to multiple ensemble learners, thereby enhancing the model's effectiveness in learning the characteristics of time series data. To bolster the model's anomaly detection capabilities, a more diverse ensemble approach and a discriminative anomaly score are recommended. Extensive experiments conducted on various real-world datasets demonstrate that the proposed method exhibits superior inference speed and more accurate anomaly detection compared to the existing competitors. A detailed experimental investigation is presented to elucidate the effectiveness of the proposed method and the underlying reasons for its efficacy.

Adaptive Subspace Optimization Ensemble Method for High-Dimensional Imbalanced Data Classification.

It is hard to construct an optimal classifier for high-dimensional imbalanced data, on which the performance of classifiers is seriously affected and becomes poor. Although many approaches, such as resampling, cost-sensitive, and ensemble learning methods, have been proposed to deal with the skewed data, they are constrained by high-dimensional data with noise and redundancy. In this study, we propose an adaptive subspace optimization ensemble method (ASOEM) for high-dimensional imbalanced data classification to overcome the above limitations. To construct accurate and diverse base classifiers, a novel adaptive subspace optimization (ASO) method based on adaptive subspace generation (ASG) process and rotated subspace optimization (RSO) process is designed to generate multiple robust and discriminative subspaces. Then a resampling scheme is applied on the optimized subspace to build a class-balanced data for each base classifier. To verify the effectiveness, our ASOEM is implemented based on different resampling strategies on 24 real-world high-dimensional imbalanced datasets. Experimental results demonstrate that our proposed methods outperform other mainstream imbalance learning approaches and classifier ensemble methods.

View-Aware Collaborative Learning for Survival Prediction and Subgroup Identification.

Advances of high throughput experimental methods have led to the availability of more diverse omic datasets in clinical analysis applications. Different types of omic data reveal different cellular aspects and contribute to the understanding of disease progression from these aspects. While survival prediction and subgroup identification are two important research problems in clinical analysis, their performance can be further boosted by taking advantages of multiple omics data through multi-view learning. However, these two tasks are generally studied separately, and the possibility that they could reinforce each other by collaborative learning has not been adequately considered. In light of this, we propose a View-aware Collaborative Learning (VaCoL) method to jointly boost the performance of survival prediction and subgroup identification by integration of multiple omics data. Specifically, survival analysis and affinity learning, which respectively perform survival prediction and subgroup identification, are integrated into a unified optimization framework to learn the two tasks in a collaborative way. In addition, by considering the diversity of different types of data, we make use of the log-rank test statistic to evaluate the importance of different views. As a result, the proposed approach can adaptively learn the optimal weight for each view during training. Empirical results on several real datasets show that our method is able to significantly improve the performance of survival prediction and subgroup identification. A detailed model analysis study is also provided to show the effectiveness of the proposed collaborative learning and view-weight learning approaches.

Trust assessment for mobile crowdsensing via device fingerprinting.

Mobile crowdsensing leverages the ubiquitous sensors of smart devices to facilitate various sensing applications. Users who participate in contributing data usually get rewards from the task requester, while there is a potential risk that someone would preempt the task and provide a forged sensing report for seeking revenue with minimal effort. Thus, trust assessment is essential to identify those irregular sensing reports. The existing methods mainly consider users' reputations and estimate the trustworthiness upon the difference from the aggregated result. However, they still face a severe problem when a majority of reports are invalid or low-quality caused by the repeated submission, e.g., a user can switch multiple accounts on a single device to repeatedly submit forged reports. To tackle this problem, we design a trust assessment scheme with an enhanced device fingerprinting algorithm. Briefly, to reduce the influence of the repeated sensing reports, we first compute their unique fingerprints derived from the intrinsic characteristics of sensors and assign an initial trust weight for each report. Then, to improve the accuracy of the assessment, we further compute the similarity of the reports to obtain their final trust values. Extensive evaluations are conducted to justify the effectiveness of our proposed design.

A review of curcumin in food preservation: Delivery system and photosensitization.

As a natural polyphenol, curcumin has been used as an alternative to synthetic preservatives in food preservation. Different from previous reviews that mainly focus on the pH-responsive discoloration of curcumin to detect changes in food quality in real time, this paper focuses on the perspective of the delivery system and photosensitization of curcumin for food preservation. The delivery system is an effective means to overcome the challenges of curcumin like instability, hydrophobicity, and low bioavailability. Curcumin as a photosensitizer can effectively sterilize to preserve food. The practical fresh-keeping effects of the delivery system and photosensitization of curcumin on foods (fruits/vegetables, animal-derived food, and grain) were summarized comprehensively, including shelf-life extension, maintenance of physicochemical properties, nutritional quality, and sensory. Future research should focus on the development of novel curcumin-loaded materials used for food preservation, and most importantly, the biosafety and accumulation toxicity associated with these materials should be explored.

Self-Guided Partial Graph Propagation for Incomplete Multiview Clustering.

In this work, we study a more realistic challenging scenario in multiview clustering (MVC), referred to as incomplete MVC (IMVC) where some instances in certain views are missing. The key to IMVC is how to adequately exploit complementary and consistency information under the incompleteness of data. However, most existing methods address the incompleteness problem at the instance level and they require sufficient information to perform data recovery. In this work, we develop a new approach to facilitate IMVC based on the graph propagation perspective. Specifically, a partial graph is used to describe the similarity of samples for incomplete views, such that the issue of missing instances can be translated into the missing entries of the partial graph. In this way, a common graph can be adaptively learned to self-guide the propagation process by exploiting the consistency information, and the propagated graph of each view is in turn used to refine the common self-guided graph in an iterative manner. Thus, the associated missing entries can be inferred through graph propagation by exploiting the consistency information across all views. On the other hand, existing approaches focus on the consistency structure only, and the complementary information has not been sufficiently exploited due to the data incompleteness issue. By contrast, under the proposed graph propagation framework, an exclusive regularization term can be naturally adopted to exploit the complementary information in our method. Extensive experiments demonstrate the effectiveness of the proposed method in comparison with state-of-the-art methods. The source code of our method is available at the https://github.com/CLiu272/TNNLS-PGP.

Feature-based 3D motif filtering for ribosomal RNA.

MOTIVATION: RNA 3D motifs are recurrent substructures in an RNA subunit and are building blocks of the RNA architecture. They play an important role in binding proteins and consolidating RNA tertiary structures. RNA 3D motif searching consists of two steps: candidate generation and candidate filtering. We proposed a novel method, known as Feature-based RNA Motif Filtering (FRMF), for identifying motifs based on a set of moment invariants and the Earth Mover's Distance in the second step. RESULTS: A positive set of RNA motifs belonging to six characteristic types, with eight subtypes occurring in HM 50S, is compiled by us. The proposed method is validated on this representative set. FRMF successfully finds most of the positive fragments. Besides the proposed new method and the compiled positive set, we also recognize some new motifs, in particular a π-turn and some non-standard A-minor motifs are found. These newly discovered motifs provide more information about RNA structure conformation. AVAILABILITY: Matlab code can be downloaded from www.cs.cityu.edu.hk/~yingshen/FRMF.html CONTACT: cshswong@cityu.edu.hk SUPPLEMENTARY INFORMATION: Supplementary data are available at Bioinformatics online.

Vitamin D attenuates high fat diet-induced hepatic steatosis in rats by modulating lipid metabolism.

BACKGROUND: Vitamin D has been reported to be reversely associated with type 2 diabetes and metabolic syndrome and is involved in modulation of lipid metabolism. The purpose of the present study was to determine whether 1,25-dihydroxyvitamin D(3) (1,25(OH)(2) D(3) ) has a protective effect on high fat diet (HFD)-induced hepatic steatosis in rats and to elucidate its underlying molecular mechanisms. MATERIALS AND METHODS: Male Sprague-Dawley (SD) rats were fed with normal fat diet, HFD or HFD with intraperitoneal injection of 1, 2.5 and 5 µg/kg 1,25(OH)(2) D(3) , respectively, each 2 days for 8 weeks. Serum lipid profile and liver triglyceride were determined. Hepatic histology was examined by haematoxylin/eosin (H&E) and Oil Red O stainings. Hepatic gene expression involved in lipogenesis and lipid oxidation was analysed by quantitative reverse transcription-polymerase chain reaction (RT-PCR). RESULTS: The administration of 1,25(OH)(2) D(3) prevented HFD-induced body weight gain and reduced liver weight. 1,25(OH)(2) D(3) attenuated hepatic steatosis in a dose-dependent manner along with improved serum lipid profile. Furthermore, 1,25(OH)(2) D(3) downregulated mRNA expression of sterol regulatory element binding protein-1c (SREBP-1c) and its target genes acetyl-CoA carboxylase (ACC) and fatty acid synthase (FAS) involved in lipogenesis. Peroxisome proliferator-activated receptor α (PPARα) and its target gene carnitine palmitoyltransferase-1 (CPT-1) involved in hepatic fatty acid (FA) oxidation were upregulated by 1,25(OH)(2) D(3) . CONCLUSIONS: These results suggest that the preventing effect of 1,25(OH)(2) D(3) against HFD-induced hepatic steatosis is related to the inhibition of lipogenesis and the promotion of FA oxidation in rat liver.

Classifier Ensemble Based on Multiview Optimization for High-Dimensional Imbalanced Data Classification.

High-dimensional class imbalanced data have plagued the performance of classification algorithms seriously. Because of a large number of redundant/invalid features and the class imbalanced issue, it is difficult to construct an optimal classifier for high-dimensional imbalanced data. Classifier ensemble has attracted intensive attention since it can achieve better performance than an individual classifier. In this work, we propose a multiview optimization (MVO) to learn more effective and robust features from high-dimensional imbalanced data, based on which an accurate and robust ensemble system is designed. Specifically, an optimized subview generation (OSG) in MVO is first proposed to generate multiple optimized subviews from different scenarios, which can strengthen the classification ability of features and increase the diversity of ensemble members simultaneously. Second, a new evaluation criterion that considers the distribution of data in each optimized subview is developed based on which a selective ensemble of optimized subviews (SEOS) is designed to perform the subview selective ensemble. Finally, an oversampling approach is executed on the optimized view to obtain a new class rebalanced subset for the classifier. Experimental results on 25 high-dimensional class imbalanced datasets indicate that the proposed method outperforms other mainstream classifier ensemble methods.

Progressive Ensemble Kernel-Based Broad Learning System for Noisy Data Classification.

The broad learning system (BLS) is an algorithm that facilitates feature representation learning and data classification. Although weights of BLS are obtained by analytical computation, which brings better generalization and higher efficiency, BLS suffers from two drawbacks: 1) the performance depends on the number of hidden nodes, which requires manual tuning, and 2) double random mappings bring about the uncertainty, which leads to poor resistance to noise data, as well as unpredictable effects on performance. To address these issues, a kernel-based BLS (KBLS) method is proposed by projecting feature nodes obtained from the first random mapping into kernel space. This manipulation reduces the uncertainty, which contributes to performance improvements with the fixed number of hidden nodes, and indicates that manually tuning is no longer needed. Moreover, to further improve the stability and noise resistance of KBLS, a progressive ensemble framework is proposed, in which the residual of the previous base classifiers is used to train the following base classifier. We conduct comparative experiments against the existing state-of-the-art hierarchical learning methods on multiple noisy real-world datasets. The experimental results indicate our approaches achieve the best or at least comparable performance in terms of accuracy.

Clustering Ensemble Based on Hybrid Multiview Clustering.

As an effective method for clustering applications, the clustering ensemble algorithm integrates different clustering solutions into a final one, thus improving the clustering efficiency. The key to designing the clustering ensemble algorithm is to improve the diversities of base learners and optimize the ensemble strategies. To address these problems, we propose a clustering ensemble framework that consists of three parts. First, three view transformation methods, including random principal component analysis, random nearest neighbor, and modified fuzzy extension model, are used as base learners to learn different clustering views. A random transformation and hybrid multiview learning-based clustering ensemble method (RTHMC) is then designed to synthesize the multiview clustering results. Second, a new random subspace transformation is integrated into RTHMC to enhance its performance. Finally, a view-based self-evolutionary strategy is developed to further improve the proposed method by optimizing random subspace sets. Experiments and comparisons demonstrate the effectiveness and superiority of the proposed method for clustering different kinds of data.

Adaptive Dense Ensemble Model for Text Classification.

Text classification has been widely explored in natural language processing. In this article, we propose a novel adaptive dense ensemble model (AdaDEM) for text classification, which includes local ensemble stage (LES) and global dense ensemble stage (GDES). To strengthen the classification ability and robustness of the enhanced layer, we propose a selective ensemble model based on enhanced attention convolutional neural networks (EnCNNs). To increase the diversity of the ensemble system, these EnCNNs are generated by using two manners: 1) different sample subsets and 2) different granularity kernels. Then, an evaluation criterion that considers both accuracy and diversity is proposed in LES to obtain effective integration results. Furthermore, to make better use of information flow, we develop an adaptive dense ensemble structure with multiple enhanced layers in GDES to mitigate the issue that there may be redundant or invalid enhanced layers in the cascade structure. We conducted extensive experiments against state-of-the-art methods on multiple real-world datasets, including long and short texts, which has verified the effectiveness and generality of our method.