A generalization of supervised single-label learning based on the assumption that each sample in a dataset may belong to more than one class simultaneously is called multi-label learning.The main objective of this wor...A generalization of supervised single-label learning based on the assumption that each sample in a dataset may belong to more than one class simultaneously is called multi-label learning.The main objective of this work is to create a novel framework for learning and classifying imbalancedmulti-label data.This work proposes a framework of two phases.The imbalanced distribution of themulti-label dataset is addressed through the proposed Borderline MLSMOTE resampling method in phase 1.Later,an adaptive weighted l21 norm regularized(Elastic-net)multilabel logistic regression is used to predict unseen samples in phase 2.The proposed Borderline MLSMOTE resampling method focuses on samples with concurrent high labels in contrast to conventional MLSMOTE.The minority labels in these samples are called difficult minority labels and are more prone to penalize classification performance.The concurrentmeasure is considered borderline,and labels associated with samples are regarded as borderline labels in the decision boundary.In phase II,a novel adaptive l21 norm regularized weighted multi-label logistic regression is used to handle balanced data with different weighted synthetic samples.Experimentation on various benchmark datasets shows the outperformance of the proposed method and its powerful predictive performances over existing conventional state-of-the-art multi-label methods.展开更多
Purpose:Many science,technology and innovation(STI)resources are attached with several different labels.To assign automatically the resulting labels to an interested instance,many approaches with good performance on t...Purpose:Many science,technology and innovation(STI)resources are attached with several different labels.To assign automatically the resulting labels to an interested instance,many approaches with good performance on the benchmark datasets have been proposed for multi-label classification task in the literature.Furthermore,several open-source tools implementing these approaches have also been developed.However,the characteristics of real-world multi-label patent and publication datasets are not completely in line with those of benchmark ones.Therefore,the main purpose of this paper is to evaluate comprehensively seven multi-label classification methods on real-world datasets.Research limitations:Three real-world datasets differ in the following aspects:statement,data quality,and purposes.Additionally,open-source tools designed for multi-label classification also have intrinsic differences in their approaches for data processing and feature selection,which in turn impacts the performance of a multi-label classification approach.In the near future,we will enhance experimental precision and reinforce the validity of conclusions by employing more rigorous control over variables through introducing expanded parameter settings.Practical implications:The observed Macro F1 and Micro F1 scores on real-world datasets typically fall short of those achieved on benchmark datasets,underscoring the complexity of real-world multi-label classification tasks.Approaches leveraging deep learning techniques offer promising solutions by accommodating the hierarchical relationships and interdependencies among labels.With ongoing enhancements in deep learning algorithms and large-scale models,it is expected that the efficacy of multi-label classification tasks will be significantly improved,reaching a level of practical utility in the foreseeable future.Originality/value:(1)Seven multi-label classification methods are comprehensively compared on three real-world datasets.(2)The TextCNN and TextRCNN models perform better on small-scale datasets with more complex hierarchical structure of labels and more balanced document-label distribution.(3)The MLkNN method works better on the larger-scale dataset with more unbalanced document-label distribution.展开更多
Complementary-label learning(CLL)aims at finding a classifier via samples with complementary labels.Such data is considered to contain less information than ordinary-label samples.The transition matrix between the tru...Complementary-label learning(CLL)aims at finding a classifier via samples with complementary labels.Such data is considered to contain less information than ordinary-label samples.The transition matrix between the true label and the complementary label,and some loss functions have been developed to handle this problem.In this paper,we show that CLL can be transformed into ordinary classification under some mild conditions,which indicates that the complementary labels can supply enough information in most cases.As an example,an extensive misclassification error analysis was performed for the Kernel Ridge Regression(KRR)method applied to multiple complementary-label learning(MCLL),which demonstrates its superior performance compared to existing approaches.展开更多
The world produces vast quantities of high-dimensional multi-semantic data.However,extracting valuable information from such a large amount of high-dimensional and multi-label data is undoubtedly arduous and challengi...The world produces vast quantities of high-dimensional multi-semantic data.However,extracting valuable information from such a large amount of high-dimensional and multi-label data is undoubtedly arduous and challenging.Feature selection aims to mitigate the adverse impacts of high dimensionality in multi-label data by eliminating redundant and irrelevant features.The ant colony optimization algorithm has demonstrated encouraging outcomes in multi-label feature selection,because of its simplicity,efficiency,and similarity to reinforcement learning.Nevertheless,existing methods do not consider crucial correlation information,such as dynamic redundancy and label correlation.To tackle these concerns,the paper proposes a multi-label feature selection technique based on ant colony optimization algorithm(MFACO),focusing on dynamic redundancy and label correlation.Initially,the dynamic redundancy is assessed between the selected feature subset and potential features.Meanwhile,the ant colony optimization algorithm extracts label correlation from the label set,which is then combined into the heuristic factor as label weights.Experimental results demonstrate that our proposed strategies can effectively enhance the optimal search ability of ant colony,outperforming the other algorithms involved in the paper.展开更多
The success of deep transfer learning in fault diagnosis is attributed to the collection of high-quality labeled data from the source domain.However,in engineering scenarios,achieving such high-quality label annotatio...The success of deep transfer learning in fault diagnosis is attributed to the collection of high-quality labeled data from the source domain.However,in engineering scenarios,achieving such high-quality label annotation is difficult and expensive.The incorrect label annotation produces two negative effects:1)the complex decision boundary of diagnosis models lowers the generalization performance on the target domain,and2)the distribution of target domain samples becomes misaligned with the false-labeled samples.To overcome these negative effects,this article proposes a solution called the label recovery and trajectory designable network(LRTDN).LRTDN consists of three parts.First,a residual network with dual classifiers is to learn features from cross-domain samples.Second,an annotation check module is constructed to generate a label anomaly indicator that could modify the abnormal labels of false-labeled samples in the source domain.With the training of relabeled samples,the complexity of diagnosis model is reduced via semi-supervised learning.Third,the adaptation trajectories are designed for sample distributions across domains.This ensures that the target domain samples are only adapted with the pure-labeled samples.The LRTDN is verified by two case studies,in which the diagnosis knowledge of bearings is transferred across different working conditions as well as different yet related machines.The results show that LRTDN offers a high diagnosis accuracy even in the presence of incorrect annotation.展开更多
The Nutri-Score is a 5-color front-of-pack nutrition label designed to provide consumers with an easily understandable guideline to the healthiness of food products.The impact that the Nutri-Score may have on consumer...The Nutri-Score is a 5-color front-of-pack nutrition label designed to provide consumers with an easily understandable guideline to the healthiness of food products.The impact that the Nutri-Score may have on consumers'choices is unclear since different experimental paradigms have found vastly different effect sizes.In the present study,we have investigated how student participants change a hypothetical personal 1-daydietary plan after a learning phase during which they learn about the Nutri-Scores of the available food items.Participants were instructed to compose a healthy diet plan in order that the question of whether the NutriScore would improve their ability to compose a healthy dietary plan could be investigated,independent of the question of whether they would apply this knowledge in their ordinary lives.We found a substantial(Cohen's d=0.86)positive impact on nutritional quality(as measured by the Nutrient Profiling System score of the Food Standards Agency)and a medium-sized(Cohen's d=0.43)reduction of energy content.Energy content reduction was larger for participants who had initially composed plans with higher energy content.The results suggest that the Nutri-Score has the potential to guide consumers to healthier food choices.It remains unclear,however,whether this potential will be reflected in real-life dietary choices.展开更多
Achieving increasingly finely targeted drug delivery to organs,tissues,cells,and even to intracellular biomacromolecules is one of the core goals of nanomedicines.As the delivery destination is refined to cellular and...Achieving increasingly finely targeted drug delivery to organs,tissues,cells,and even to intracellular biomacromolecules is one of the core goals of nanomedicines.As the delivery destination is refined to cellular and subcellular targets,it is essential to explore the delivery of nanomedicines at the molecular level.However,due to the lack of technical methods,the molecular mechanism of the intracellular delivery of nanomedicines remains unclear to date.Here,we develop an enzyme-induced proximity labeling technology in nanoparticles(nano-EPL)for the real-time monitoring of proteins that interact with intracellular nanomedicines.Poly(lactic-co-glycolic acid)nanoparticles coupled with horseradish peroxidase(HRP)were fabricated as a model(HRP(+)-PNPs)to evaluate the molecular mechanism of nano delivery in macrophages.By adding the labeling probe biotin-phenol and the catalytic substrate H_(2)O_(2)at different time points in cellular delivery,nano-EPL technology was validated for the real-time in situ labeling of proteins interacting with nanoparticles.Nano-EPL achieves the dynamic molecular profiling of 740 proteins to map the intracellular delivery of HRP(+)-PNPs in macrophages over time.Based on dynamic clustering analysis of these proteins,we further discovered that different organelles,including endosomes,lysosomes,the endoplasmic reticulum,and the Golgi apparatus,are involved in delivery with distinct participation timelines.More importantly,the engagement of these organelles differentially affects the drug delivery efficiency,reflecting the spatial–temporal heterogeneity of nano delivery in cells.In summary,these findings highlight a significant methodological advance toward understanding the molecular mechanisms involved in the intracellular delivery of nanomedicines.展开更多
The halopyrimidine 5-bromo-2′-deoxyuridine(BrdU)is an exogenous marker of DNA synthesis.Since the introduction of monoclonal antibodies against BrdU,an increasing number of methodologies have been used for the immuno...The halopyrimidine 5-bromo-2′-deoxyuridine(BrdU)is an exogenous marker of DNA synthesis.Since the introduction of monoclonal antibodies against BrdU,an increasing number of methodologies have been used for the immunodetection of this synthesized bromine-tagged base analogue into replicating DNA.BrdU labeling is widely used for identifying neuron precursors and following their fate during the embryonic,perinatal,and adult neurogenesis in a variety of vertebrate species including birds,reptiles,and mammals.Due to BrdU toxicity,its incorporation into replicating DNA presents adverse consequences on the generation,survival,and settled patterns of cells.This may lead to false results and misinterpretation in the identification of proliferative neuroblasts.In this review,I will indicate the detrimental effects of this nucleoside during the development of the central nervous system,as well as the reliability of BrdU labeling to detect proliferating neuroblasts.Moreover,it will show factors influencing BrdU immunodetection and the contribution of this nucleoside to the study of prenatal,perinatal,and adult neurogenesis.Human adult neurogenesis will also be discussed.It is my hope that this review serves as a reference for those researchers who focused on detecting cells that are in the synthetic phase of the cell cycle.展开更多
Battery pack capacity estimation under real-world operating conditions is important for battery performance optimization and health management,contributing to the reliability and longevity of batterypowered systems.Ho...Battery pack capacity estimation under real-world operating conditions is important for battery performance optimization and health management,contributing to the reliability and longevity of batterypowered systems.However,complex operating conditions,coupling cell-to-cell inconsistency,and limited labeled data pose great challenges to accurate and robust battery pack capacity estimation.To address these issues,this paper proposes a hierarchical data-driven framework aimed at enhancing the training of machine learning models with fewer labeled data.Unlike traditional data-driven methods that lack interpretability,the hierarchical data-driven framework unveils the“mechanism”of the black box inside the data-driven framework by splitting the final estimation target into cell-level and pack-level intermediate targets.A generalized feature matrix is devised without requiring all cell voltages,significantly reducing the computational cost and memory resources.The generated intermediate target labels and the corresponding features are hierarchically employed to enhance the training of two machine learning models,effectively alleviating the difficulty of learning the relationship from all features due to fewer labeled data and addressing the dilemma of requiring extensive labeled data for accurate estimation.Using only 10%of degradation data,the proposed framework outperforms the state-of-the-art battery pack capacity estimation methods,achieving mean absolute percentage errors of 0.608%,0.601%,and 1.128%for three battery packs whose degradation load profiles represent real-world operating conditions.Its high accuracy,adaptability,and robustness indicate the potential in different application scenarios,which is promising for reducing laborious and expensive aging experiments at the pack level and facilitating the development of battery technology.展开更多
Recently,weak supervision has received growing attention in the field of salient object detection due to the convenience of labelling.However,there is a large performance gap between weakly supervised and fully superv...Recently,weak supervision has received growing attention in the field of salient object detection due to the convenience of labelling.However,there is a large performance gap between weakly supervised and fully supervised salient object detectors because the scribble annotation can only provide very limited foreground/background information.Therefore,an intuitive idea is to infer annotations that cover more complete object and background regions for training.To this end,a label inference strategy is proposed based on the assumption that pixels with similar colours and close positions should have consistent labels.Specifically,k-means clustering algorithm was first performed on both colours and coordinates of original annotations,and then assigned the same labels to points having similar colours with colour cluster centres and near coordinate cluster centres.Next,the same annotations for pixels with similar colours within each kernel neighbourhood was set further.Extensive experiments on six benchmarks demonstrate that our method can significantly improve the performance and achieve the state-of-the-art results.展开更多
Multimodal sentiment analysis utilizes multimodal data such as text,facial expressions and voice to detect people’s attitudes.With the advent of distributed data collection and annotation,we can easily obtain and sha...Multimodal sentiment analysis utilizes multimodal data such as text,facial expressions and voice to detect people’s attitudes.With the advent of distributed data collection and annotation,we can easily obtain and share such multimodal data.However,due to professional discrepancies among annotators and lax quality control,noisy labels might be introduced.Recent research suggests that deep neural networks(DNNs)will overfit noisy labels,leading to the poor performance of the DNNs.To address this challenging problem,we present a Multimodal Robust Meta Learning framework(MRML)for multimodal sentiment analysis to resist noisy labels and correlate distinct modalities simultaneously.Specifically,we propose a two-layer fusion net to deeply fuse different modalities and improve the quality of the multimodal data features for label correction and network training.Besides,a multiple meta-learner(label corrector)strategy is proposed to enhance the label correction approach and prevent models from overfitting to noisy labels.We conducted experiments on three popular multimodal datasets to verify the superiority of ourmethod by comparing it with four baselines.展开更多
BACKGROUND Panic disorder(PD)involves emotion dysregulation,but its underlying mechanisms remain poorly understood.Previous research suggests that implicit emotion regulation may play a central role in PD-related emot...BACKGROUND Panic disorder(PD)involves emotion dysregulation,but its underlying mechanisms remain poorly understood.Previous research suggests that implicit emotion regulation may play a central role in PD-related emotion dysregulation and symptom maintenance.However,there is a lack of studies exploring the neural mechanisms of implicit emotion regulation in PD using neurophysiological indicators.AIM To study the neural mechanisms of implicit emotion regulation in PD with eventrelated potentials(ERP).METHODS A total of 25 PD patients and 20 healthy controls(HC)underwent clinical evaluations.The study utilized a case-control design with random sampling,selecting participants for the case group from March to December 2018.Participants performed an affect labeling task,using affect labeling as the experimental condition and gender labeling as the control condition.ERP and behavioral data were recorded to compare the late positive potential(LPP)within and between the groups.RESULTS Both PD and HC groups showed longer reaction times and decreased accuracy under the affect labeling.In the HC group,late LPP amplitudes exhibited a dynamic pattern of initial increase followed by decrease.Importantly,a significant group×condition interaction effect was observed.Simple effect analysis revealed a reduction in the differences of late LPP amplitudes between the affect labeling and gender labeling conditions in the PD group compared to the HC group.Furthermore,among PD patients under the affect labeling,the late LPP was negatively correlated with disease severity,symptom frequency,and intensity.CONCLUSION PD patients demonstrate abnormalities in implicit emotion regulation,hampering their ability to mobilize cognitive resources for downregulating negative emotions.The late LPP amplitude in response to affect labeling may serve as a potentially valuable clinical indicator of PD severity.展开更多
Recent advances in utilizing ^(17)O isotopic labeling methods for solid-state nuclear magnetic resonance(NMR)investigations of metal oxides for lithium-ion batteries have yielded extensive insights into their structur...Recent advances in utilizing ^(17)O isotopic labeling methods for solid-state nuclear magnetic resonance(NMR)investigations of metal oxides for lithium-ion batteries have yielded extensive insights into their structural and dynamic details.Herein,we commence with a brief introduction to recent research on lithium-ion battery oxide materials studied using ^(17)O solid-state NMR spectroscopy.Then we delve into a review of ^(17)O isotopic labeling methods for tagging oxygen sites in both the bulk and surfaces of metal oxides.At last,the unresolved problems and the future research directions for advancing the ^(17)O labeling technique are discussed.展开更多
Non-descriptive and convenient labels are uninformative and unfairly project blame onto patients.The language clinicians use in the Electronic Medical Record,research,and clinical settings shapes biases and subsequent...Non-descriptive and convenient labels are uninformative and unfairly project blame onto patients.The language clinicians use in the Electronic Medical Record,research,and clinical settings shapes biases and subsequent behaviors of all providers involved in the enterprise of transplantation.Terminology such as noncompliant and nonadherent serve as a reason for waitlist inactivation and limit access to life-saving transplantation.These labels fail to capture all the circum-stances surrounding a patient’s inability to follow their care regimen,trivialize social determinants of health variables,and bring unsubstantiated subjectivity into decisions regarding organ allocation.Furthermore,insufficient Medicare coverage has forced patients to ration or stop taking medication,leading to allograft failure and their subsequent diagnosis of noncompliant.We argue that perpetuating non-descriptive language adds little substantive information,in-creases subjectivity to the organ allocation process,and plays a major role in reduced access to transplantation.For patients with existing barriers to care,such as racial/ethnic minorities,these effects may be even more drastic.Transplant committees must ensure thorough documentation to correctly encapsulate the entirety of a patient’s position and give voice to an already vulnerable population.展开更多
BACKGROUND Global and regional cerebral blood flow(CBF)changes in patients with unilateral internal carotid artery occlusion(ICAO)are unclear when the dual post-labeling delays(PLD)arterial spin labeling(ASL)magnetic ...BACKGROUND Global and regional cerebral blood flow(CBF)changes in patients with unilateral internal carotid artery occlusion(ICAO)are unclear when the dual post-labeling delays(PLD)arterial spin labeling(ASL)magnetic resonance imaging(MRI)technique is used.Manual delineation of regions of interest for CBF measurement is time-consuming and laborious.AIM To assess global and regional CBF changes in patients with unilateral ICAO with the ASL-MRI perfusion technique.METHODS Twenty hospitalized patients with ICAO and sex-and age-matched controls were included in the study.Regional CBF was measured by Dr.Brain's ASL software.The present study evaluated differences in global,middle cerebral artery(MCA)territory,anterior cerebral artery territory,and Alberta Stroke Program Early Computed Tomography Score(ASPECTS)regions(including the caudate nucleus,lentiform nucleus,insula ribbon,internal capsule,and M1-M6)and brain lobes(including frontal,parietal,temporal,and insular lobes)between ICAO patients and controls at PLD 1.5 s and PLD 2.5 s.RESULTS When comparing CBF between ICAO patients and controls,the global CBF in ICAO patients was lower at both PLD 1.5 s and PLD 2.5 s;the CBF on the occluded side was lower in 15 brain regions at PLD 1.5 s,and it was lower in 9 brain regions at PLD 2.5 s;the CBF in the contralateral hemisphere was lower in the caudate nucleus and internal capsule at PLD 1.5 s and in M6 at PLD 2.5 s.The global CBF in ICAO patients was lower at PLD 1.5 s than at PLD 2.5 s.The ipsilateral CBF at PLD 1.5 s was lower than that at PLD 2.5 s in 15 regions,whereas the contralateral CBF was lower at PLD 1.5 s than at PLD 2.5 s in 12 regions.The ipsilateral CBF was lower than the contralateral CBF in 15 regions at PLD 1.5 s,and in M6 at PLD 2.5 s.CONCLUSION Unilateral ICAO results in hypoperfusion in the global and MCA territories,especially in the ASPECTS area.Dual PLD settings prove more suitable for accurate CBF quantification in ICAO.展开更多
In the field of medical images,pixel-level labels are time-consuming and expensive to acquire,while image-level labels are relatively easier to obtain.Therefore,it makes sense to learn more information(knowledge)from ...In the field of medical images,pixel-level labels are time-consuming and expensive to acquire,while image-level labels are relatively easier to obtain.Therefore,it makes sense to learn more information(knowledge)from a small number of hard-to-get pixel-level annotated images to apply to different tasks to maximize their usefulness and save time and training costs.In this paper,using Pixel-Level Labeled Images forMulti-Task Learning(PLDMLT),we focus on grading the severity of fundus images for Diabetic Retinopathy(DR).This is because,for the segmentation task,there is a finely labeled mask,while the severity grading task is without classification labels.To this end,we propose a two-stage multi-label learning weakly supervised algorithm,which generates initial classification pseudo labels in the first stage and visualizes heat maps at all levels of severity using Grad-Cam to further provide medical interpretability for the classification task.A multitask model framework with U-net as the baseline is proposed in the second stage.A label update network is designed to alleviate the gradient balance between the classification and segmentation tasks.Extensive experimental results show that our PLDMLTmethod significantly outperforms other stateof-the-art methods in DR segmentation on two public datasets,achieving up to 98.897%segmentation accuracy.In addition,our method achieves comparable competitiveness with single-task fully supervised learning in the DR severity grading task.展开更多
In the realm of Multi-Label Text Classification(MLTC),the dual challenges of extracting rich semantic features from text and discerning inter-label relationships have spurred innovative approaches.Many studies in sema...In the realm of Multi-Label Text Classification(MLTC),the dual challenges of extracting rich semantic features from text and discerning inter-label relationships have spurred innovative approaches.Many studies in semantic feature extraction have turned to external knowledge to augment the model’s grasp of textual content,often overlooking intrinsic textual cues such as label statistical features.In contrast,these endogenous insights naturally align with the classification task.In our paper,to complement this focus on intrinsic knowledge,we introduce a novel Gate-Attention mechanism.This mechanism adeptly integrates statistical features from the text itself into the semantic fabric,enhancing the model’s capacity to understand and represent the data.Additionally,to address the intricate task of mining label correlations,we propose a Dual-end enhancement mechanism.This mechanism effectively mitigates the challenges of information loss and erroneous transmission inherent in traditional long short term memory propagation.We conducted an extensive battery of experiments on the AAPD and RCV1-2 datasets.These experiments serve the dual purpose of confirming the efficacy of both the Gate-Attention mechanism and the Dual-end enhancement mechanism.Our final model unequivocally outperforms the baseline model,attesting to its robustness.These findings emphatically underscore the imperativeness of taking into account not just external knowledge but also the inherent intricacies of textual data when crafting potent MLTC models.展开更多
基金partly supported by the Technology Development Program of MSS(No.S3033853)by the National Research Foundation of Korea(NRF)grant funded by the Korea government(MSIT)(No.2021R1A4A1031509).
文摘A generalization of supervised single-label learning based on the assumption that each sample in a dataset may belong to more than one class simultaneously is called multi-label learning.The main objective of this work is to create a novel framework for learning and classifying imbalancedmulti-label data.This work proposes a framework of two phases.The imbalanced distribution of themulti-label dataset is addressed through the proposed Borderline MLSMOTE resampling method in phase 1.Later,an adaptive weighted l21 norm regularized(Elastic-net)multilabel logistic regression is used to predict unseen samples in phase 2.The proposed Borderline MLSMOTE resampling method focuses on samples with concurrent high labels in contrast to conventional MLSMOTE.The minority labels in these samples are called difficult minority labels and are more prone to penalize classification performance.The concurrentmeasure is considered borderline,and labels associated with samples are regarded as borderline labels in the decision boundary.In phase II,a novel adaptive l21 norm regularized weighted multi-label logistic regression is used to handle balanced data with different weighted synthetic samples.Experimentation on various benchmark datasets shows the outperformance of the proposed method and its powerful predictive performances over existing conventional state-of-the-art multi-label methods.
基金the Natural Science Foundation of China(Grant Numbers 72074014 and 72004012).
文摘Purpose:Many science,technology and innovation(STI)resources are attached with several different labels.To assign automatically the resulting labels to an interested instance,many approaches with good performance on the benchmark datasets have been proposed for multi-label classification task in the literature.Furthermore,several open-source tools implementing these approaches have also been developed.However,the characteristics of real-world multi-label patent and publication datasets are not completely in line with those of benchmark ones.Therefore,the main purpose of this paper is to evaluate comprehensively seven multi-label classification methods on real-world datasets.Research limitations:Three real-world datasets differ in the following aspects:statement,data quality,and purposes.Additionally,open-source tools designed for multi-label classification also have intrinsic differences in their approaches for data processing and feature selection,which in turn impacts the performance of a multi-label classification approach.In the near future,we will enhance experimental precision and reinforce the validity of conclusions by employing more rigorous control over variables through introducing expanded parameter settings.Practical implications:The observed Macro F1 and Micro F1 scores on real-world datasets typically fall short of those achieved on benchmark datasets,underscoring the complexity of real-world multi-label classification tasks.Approaches leveraging deep learning techniques offer promising solutions by accommodating the hierarchical relationships and interdependencies among labels.With ongoing enhancements in deep learning algorithms and large-scale models,it is expected that the efficacy of multi-label classification tasks will be significantly improved,reaching a level of practical utility in the foreseeable future.Originality/value:(1)Seven multi-label classification methods are comprehensively compared on three real-world datasets.(2)The TextCNN and TextRCNN models perform better on small-scale datasets with more complex hierarchical structure of labels and more balanced document-label distribution.(3)The MLkNN method works better on the larger-scale dataset with more unbalanced document-label distribution.
基金Supported by the Indigenous Innovation’s Capability Development Program of Huizhou University(HZU202003,HZU202020)Natural Science Foundation of Guangdong Province(2022A1515011463)+2 种基金the Project of Educational Commission of Guangdong Province(2023ZDZX1025)National Natural Science Foundation of China(12271473)Guangdong Province’s 2023 Education Science Planning Project(Higher Education Special Project)(2023GXJK505)。
文摘Complementary-label learning(CLL)aims at finding a classifier via samples with complementary labels.Such data is considered to contain less information than ordinary-label samples.The transition matrix between the true label and the complementary label,and some loss functions have been developed to handle this problem.In this paper,we show that CLL can be transformed into ordinary classification under some mild conditions,which indicates that the complementary labels can supply enough information in most cases.As an example,an extensive misclassification error analysis was performed for the Kernel Ridge Regression(KRR)method applied to multiple complementary-label learning(MCLL),which demonstrates its superior performance compared to existing approaches.
基金supported by National Natural Science Foundation of China(Grant Nos.62376089,62302153,62302154,62202147)the key Research and Development Program of Hubei Province,China(Grant No.2023BEB024).
文摘The world produces vast quantities of high-dimensional multi-semantic data.However,extracting valuable information from such a large amount of high-dimensional and multi-label data is undoubtedly arduous and challenging.Feature selection aims to mitigate the adverse impacts of high dimensionality in multi-label data by eliminating redundant and irrelevant features.The ant colony optimization algorithm has demonstrated encouraging outcomes in multi-label feature selection,because of its simplicity,efficiency,and similarity to reinforcement learning.Nevertheless,existing methods do not consider crucial correlation information,such as dynamic redundancy and label correlation.To tackle these concerns,the paper proposes a multi-label feature selection technique based on ant colony optimization algorithm(MFACO),focusing on dynamic redundancy and label correlation.Initially,the dynamic redundancy is assessed between the selected feature subset and potential features.Meanwhile,the ant colony optimization algorithm extracts label correlation from the label set,which is then combined into the heuristic factor as label weights.Experimental results demonstrate that our proposed strategies can effectively enhance the optimal search ability of ant colony,outperforming the other algorithms involved in the paper.
基金the National Key R&D Program of China(2022YFB3402100)the National Science Fund for Distinguished Young Scholars of China(52025056)+4 种基金the National Natural Science Foundation of China(52305129)the China Postdoctoral Science Foundation(2023M732789)the China Postdoctoral Innovative Talents Support Program(BX20230290)the Open Foundation of Hunan Provincial Key Laboratory of Health Maintenance for Mechanical Equipment(2022JXKF JJ01)the Fundamental Research Funds for Central Universities。
文摘The success of deep transfer learning in fault diagnosis is attributed to the collection of high-quality labeled data from the source domain.However,in engineering scenarios,achieving such high-quality label annotation is difficult and expensive.The incorrect label annotation produces two negative effects:1)the complex decision boundary of diagnosis models lowers the generalization performance on the target domain,and2)the distribution of target domain samples becomes misaligned with the false-labeled samples.To overcome these negative effects,this article proposes a solution called the label recovery and trajectory designable network(LRTDN).LRTDN consists of three parts.First,a residual network with dual classifiers is to learn features from cross-domain samples.Second,an annotation check module is constructed to generate a label anomaly indicator that could modify the abnormal labels of false-labeled samples in the source domain.With the training of relabeled samples,the complexity of diagnosis model is reduced via semi-supervised learning.Third,the adaptation trajectories are designed for sample distributions across domains.This ensures that the target domain samples are only adapted with the pure-labeled samples.The LRTDN is verified by two case studies,in which the diagnosis knowledge of bearings is transferred across different working conditions as well as different yet related machines.The results show that LRTDN offers a high diagnosis accuracy even in the presence of incorrect annotation.
文摘The Nutri-Score is a 5-color front-of-pack nutrition label designed to provide consumers with an easily understandable guideline to the healthiness of food products.The impact that the Nutri-Score may have on consumers'choices is unclear since different experimental paradigms have found vastly different effect sizes.In the present study,we have investigated how student participants change a hypothetical personal 1-daydietary plan after a learning phase during which they learn about the Nutri-Scores of the available food items.Participants were instructed to compose a healthy diet plan in order that the question of whether the NutriScore would improve their ability to compose a healthy dietary plan could be investigated,independent of the question of whether they would apply this knowledge in their ordinary lives.We found a substantial(Cohen's d=0.86)positive impact on nutritional quality(as measured by the Nutrient Profiling System score of the Food Standards Agency)and a medium-sized(Cohen's d=0.43)reduction of energy content.Energy content reduction was larger for participants who had initially composed plans with higher energy content.The results suggest that the Nutri-Score has the potential to guide consumers to healthier food choices.It remains unclear,however,whether this potential will be reflected in real-life dietary choices.
基金supported by Natural Science Foundation of Beijing Municipality(L212013)National Key Research and Development Program of China(No.2022YFA1206104)+2 种基金AI+Health Collaborative Innovation Cultivation Project(Z211100003521002)National Natural Science Foundation of China(81971718,82073786,81872809,U20A20412,81821004)Beijing Natural Science Foundation(7222020).
文摘Achieving increasingly finely targeted drug delivery to organs,tissues,cells,and even to intracellular biomacromolecules is one of the core goals of nanomedicines.As the delivery destination is refined to cellular and subcellular targets,it is essential to explore the delivery of nanomedicines at the molecular level.However,due to the lack of technical methods,the molecular mechanism of the intracellular delivery of nanomedicines remains unclear to date.Here,we develop an enzyme-induced proximity labeling technology in nanoparticles(nano-EPL)for the real-time monitoring of proteins that interact with intracellular nanomedicines.Poly(lactic-co-glycolic acid)nanoparticles coupled with horseradish peroxidase(HRP)were fabricated as a model(HRP(+)-PNPs)to evaluate the molecular mechanism of nano delivery in macrophages.By adding the labeling probe biotin-phenol and the catalytic substrate H_(2)O_(2)at different time points in cellular delivery,nano-EPL technology was validated for the real-time in situ labeling of proteins interacting with nanoparticles.Nano-EPL achieves the dynamic molecular profiling of 740 proteins to map the intracellular delivery of HRP(+)-PNPs in macrophages over time.Based on dynamic clustering analysis of these proteins,we further discovered that different organelles,including endosomes,lysosomes,the endoplasmic reticulum,and the Golgi apparatus,are involved in delivery with distinct participation timelines.More importantly,the engagement of these organelles differentially affects the drug delivery efficiency,reflecting the spatial–temporal heterogeneity of nano delivery in cells.In summary,these findings highlight a significant methodological advance toward understanding the molecular mechanisms involved in the intracellular delivery of nanomedicines.
文摘The halopyrimidine 5-bromo-2′-deoxyuridine(BrdU)is an exogenous marker of DNA synthesis.Since the introduction of monoclonal antibodies against BrdU,an increasing number of methodologies have been used for the immunodetection of this synthesized bromine-tagged base analogue into replicating DNA.BrdU labeling is widely used for identifying neuron precursors and following their fate during the embryonic,perinatal,and adult neurogenesis in a variety of vertebrate species including birds,reptiles,and mammals.Due to BrdU toxicity,its incorporation into replicating DNA presents adverse consequences on the generation,survival,and settled patterns of cells.This may lead to false results and misinterpretation in the identification of proliferative neuroblasts.In this review,I will indicate the detrimental effects of this nucleoside during the development of the central nervous system,as well as the reliability of BrdU labeling to detect proliferating neuroblasts.Moreover,it will show factors influencing BrdU immunodetection and the contribution of this nucleoside to the study of prenatal,perinatal,and adult neurogenesis.Human adult neurogenesis will also be discussed.It is my hope that this review serves as a reference for those researchers who focused on detecting cells that are in the synthetic phase of the cell cycle.
基金supported by the National Outstanding Youth Science Fund Project of National Natural Science Foundation of China[Grant No.52222708]the Natural Science Foundation of Beijing Municipality[Grant No.3212033]。
文摘Battery pack capacity estimation under real-world operating conditions is important for battery performance optimization and health management,contributing to the reliability and longevity of batterypowered systems.However,complex operating conditions,coupling cell-to-cell inconsistency,and limited labeled data pose great challenges to accurate and robust battery pack capacity estimation.To address these issues,this paper proposes a hierarchical data-driven framework aimed at enhancing the training of machine learning models with fewer labeled data.Unlike traditional data-driven methods that lack interpretability,the hierarchical data-driven framework unveils the“mechanism”of the black box inside the data-driven framework by splitting the final estimation target into cell-level and pack-level intermediate targets.A generalized feature matrix is devised without requiring all cell voltages,significantly reducing the computational cost and memory resources.The generated intermediate target labels and the corresponding features are hierarchically employed to enhance the training of two machine learning models,effectively alleviating the difficulty of learning the relationship from all features due to fewer labeled data and addressing the dilemma of requiring extensive labeled data for accurate estimation.Using only 10%of degradation data,the proposed framework outperforms the state-of-the-art battery pack capacity estimation methods,achieving mean absolute percentage errors of 0.608%,0.601%,and 1.128%for three battery packs whose degradation load profiles represent real-world operating conditions.Its high accuracy,adaptability,and robustness indicate the potential in different application scenarios,which is promising for reducing laborious and expensive aging experiments at the pack level and facilitating the development of battery technology.
文摘Recently,weak supervision has received growing attention in the field of salient object detection due to the convenience of labelling.However,there is a large performance gap between weakly supervised and fully supervised salient object detectors because the scribble annotation can only provide very limited foreground/background information.Therefore,an intuitive idea is to infer annotations that cover more complete object and background regions for training.To this end,a label inference strategy is proposed based on the assumption that pixels with similar colours and close positions should have consistent labels.Specifically,k-means clustering algorithm was first performed on both colours and coordinates of original annotations,and then assigned the same labels to points having similar colours with colour cluster centres and near coordinate cluster centres.Next,the same annotations for pixels with similar colours within each kernel neighbourhood was set further.Extensive experiments on six benchmarks demonstrate that our method can significantly improve the performance and achieve the state-of-the-art results.
基金supported by STI 2030-Major Projects 2021ZD0200400National Natural Science Foundation of China(62276233 and 62072405)Key Research Project of Zhejiang Province(2023C01048).
文摘Multimodal sentiment analysis utilizes multimodal data such as text,facial expressions and voice to detect people’s attitudes.With the advent of distributed data collection and annotation,we can easily obtain and share such multimodal data.However,due to professional discrepancies among annotators and lax quality control,noisy labels might be introduced.Recent research suggests that deep neural networks(DNNs)will overfit noisy labels,leading to the poor performance of the DNNs.To address this challenging problem,we present a Multimodal Robust Meta Learning framework(MRML)for multimodal sentiment analysis to resist noisy labels and correlate distinct modalities simultaneously.Specifically,we propose a two-layer fusion net to deeply fuse different modalities and improve the quality of the multimodal data features for label correction and network training.Besides,a multiple meta-learner(label corrector)strategy is proposed to enhance the label correction approach and prevent models from overfitting to noisy labels.We conducted experiments on three popular multimodal datasets to verify the superiority of ourmethod by comparing it with four baselines.
基金Supported by The National Natural Science Foundation of China,No.81871080the Key R&D Program of Jining(Major Program),No.2023YXNS004+2 种基金the National Natural Science Foundation of China,No.81401486the Natural Science Foundation of Liaoning Province of China,No.20170540276the Medicine and Health Science Technology Development Program of Shandong Province,No.202003070713.
文摘BACKGROUND Panic disorder(PD)involves emotion dysregulation,but its underlying mechanisms remain poorly understood.Previous research suggests that implicit emotion regulation may play a central role in PD-related emotion dysregulation and symptom maintenance.However,there is a lack of studies exploring the neural mechanisms of implicit emotion regulation in PD using neurophysiological indicators.AIM To study the neural mechanisms of implicit emotion regulation in PD with eventrelated potentials(ERP).METHODS A total of 25 PD patients and 20 healthy controls(HC)underwent clinical evaluations.The study utilized a case-control design with random sampling,selecting participants for the case group from March to December 2018.Participants performed an affect labeling task,using affect labeling as the experimental condition and gender labeling as the control condition.ERP and behavioral data were recorded to compare the late positive potential(LPP)within and between the groups.RESULTS Both PD and HC groups showed longer reaction times and decreased accuracy under the affect labeling.In the HC group,late LPP amplitudes exhibited a dynamic pattern of initial increase followed by decrease.Importantly,a significant group×condition interaction effect was observed.Simple effect analysis revealed a reduction in the differences of late LPP amplitudes between the affect labeling and gender labeling conditions in the PD group compared to the HC group.Furthermore,among PD patients under the affect labeling,the late LPP was negatively correlated with disease severity,symptom frequency,and intensity.CONCLUSION PD patients demonstrate abnormalities in implicit emotion regulation,hampering their ability to mobilize cognitive resources for downregulating negative emotions.The late LPP amplitude in response to affect labeling may serve as a potentially valuable clinical indicator of PD severity.
基金supported by National Key R&D Program of China(2021YFA1502803)the National Natural Science Foundation of China(NSFC)(21972066,91745202)+3 种基金NSFC-Royal Society Joint Program(21661130149)L.P.thanks the Royal Society and Newton Fund for a Royal Society-Newton Advanced Fellowshipsupported by the Research Funds for the Frontiers Science Centre for Critical Earth Material Cycling,Nanjing Universitya Project Funded by the Priority Academic Program Development of Jiangsu Higher Education Institutions.
文摘Recent advances in utilizing ^(17)O isotopic labeling methods for solid-state nuclear magnetic resonance(NMR)investigations of metal oxides for lithium-ion batteries have yielded extensive insights into their structural and dynamic details.Herein,we commence with a brief introduction to recent research on lithium-ion battery oxide materials studied using ^(17)O solid-state NMR spectroscopy.Then we delve into a review of ^(17)O isotopic labeling methods for tagging oxygen sites in both the bulk and surfaces of metal oxides.At last,the unresolved problems and the future research directions for advancing the ^(17)O labeling technique are discussed.
文摘Non-descriptive and convenient labels are uninformative and unfairly project blame onto patients.The language clinicians use in the Electronic Medical Record,research,and clinical settings shapes biases and subsequent behaviors of all providers involved in the enterprise of transplantation.Terminology such as noncompliant and nonadherent serve as a reason for waitlist inactivation and limit access to life-saving transplantation.These labels fail to capture all the circum-stances surrounding a patient’s inability to follow their care regimen,trivialize social determinants of health variables,and bring unsubstantiated subjectivity into decisions regarding organ allocation.Furthermore,insufficient Medicare coverage has forced patients to ration or stop taking medication,leading to allograft failure and their subsequent diagnosis of noncompliant.We argue that perpetuating non-descriptive language adds little substantive information,in-creases subjectivity to the organ allocation process,and plays a major role in reduced access to transplantation.For patients with existing barriers to care,such as racial/ethnic minorities,these effects may be even more drastic.Transplant committees must ensure thorough documentation to correctly encapsulate the entirety of a patient’s position and give voice to an already vulnerable population.
基金Supported by The Key Research and Development Program Projects of Shaanxi Province of China,No.S2023-YF-YBSF-0273Natural Science Foundation of Shaanxi Province of China,No.2022JQ-900.
文摘BACKGROUND Global and regional cerebral blood flow(CBF)changes in patients with unilateral internal carotid artery occlusion(ICAO)are unclear when the dual post-labeling delays(PLD)arterial spin labeling(ASL)magnetic resonance imaging(MRI)technique is used.Manual delineation of regions of interest for CBF measurement is time-consuming and laborious.AIM To assess global and regional CBF changes in patients with unilateral ICAO with the ASL-MRI perfusion technique.METHODS Twenty hospitalized patients with ICAO and sex-and age-matched controls were included in the study.Regional CBF was measured by Dr.Brain's ASL software.The present study evaluated differences in global,middle cerebral artery(MCA)territory,anterior cerebral artery territory,and Alberta Stroke Program Early Computed Tomography Score(ASPECTS)regions(including the caudate nucleus,lentiform nucleus,insula ribbon,internal capsule,and M1-M6)and brain lobes(including frontal,parietal,temporal,and insular lobes)between ICAO patients and controls at PLD 1.5 s and PLD 2.5 s.RESULTS When comparing CBF between ICAO patients and controls,the global CBF in ICAO patients was lower at both PLD 1.5 s and PLD 2.5 s;the CBF on the occluded side was lower in 15 brain regions at PLD 1.5 s,and it was lower in 9 brain regions at PLD 2.5 s;the CBF in the contralateral hemisphere was lower in the caudate nucleus and internal capsule at PLD 1.5 s and in M6 at PLD 2.5 s.The global CBF in ICAO patients was lower at PLD 1.5 s than at PLD 2.5 s.The ipsilateral CBF at PLD 1.5 s was lower than that at PLD 2.5 s in 15 regions,whereas the contralateral CBF was lower at PLD 1.5 s than at PLD 2.5 s in 12 regions.The ipsilateral CBF was lower than the contralateral CBF in 15 regions at PLD 1.5 s,and in M6 at PLD 2.5 s.CONCLUSION Unilateral ICAO results in hypoperfusion in the global and MCA territories,especially in the ASPECTS area.Dual PLD settings prove more suitable for accurate CBF quantification in ICAO.
文摘In the field of medical images,pixel-level labels are time-consuming and expensive to acquire,while image-level labels are relatively easier to obtain.Therefore,it makes sense to learn more information(knowledge)from a small number of hard-to-get pixel-level annotated images to apply to different tasks to maximize their usefulness and save time and training costs.In this paper,using Pixel-Level Labeled Images forMulti-Task Learning(PLDMLT),we focus on grading the severity of fundus images for Diabetic Retinopathy(DR).This is because,for the segmentation task,there is a finely labeled mask,while the severity grading task is without classification labels.To this end,we propose a two-stage multi-label learning weakly supervised algorithm,which generates initial classification pseudo labels in the first stage and visualizes heat maps at all levels of severity using Grad-Cam to further provide medical interpretability for the classification task.A multitask model framework with U-net as the baseline is proposed in the second stage.A label update network is designed to alleviate the gradient balance between the classification and segmentation tasks.Extensive experimental results show that our PLDMLTmethod significantly outperforms other stateof-the-art methods in DR segmentation on two public datasets,achieving up to 98.897%segmentation accuracy.In addition,our method achieves comparable competitiveness with single-task fully supervised learning in the DR severity grading task.
基金supported by National Natural Science Foundation of China(NSFC)(Grant Nos.62162022,62162024)the Key Research and Development Program of Hainan Province(Grant Nos.ZDYF2020040,ZDYF2021GXJS003)+2 种基金the Major Science and Technology Project of Hainan Province(Grant No.ZDKJ2020012)Hainan Provincial Natural Science Foundation of China(Grant Nos.620MS021,621QN211)Science and Technology Development Center of the Ministry of Education Industry-University-Research Innovation Fund(2021JQR017).
文摘In the realm of Multi-Label Text Classification(MLTC),the dual challenges of extracting rich semantic features from text and discerning inter-label relationships have spurred innovative approaches.Many studies in semantic feature extraction have turned to external knowledge to augment the model’s grasp of textual content,often overlooking intrinsic textual cues such as label statistical features.In contrast,these endogenous insights naturally align with the classification task.In our paper,to complement this focus on intrinsic knowledge,we introduce a novel Gate-Attention mechanism.This mechanism adeptly integrates statistical features from the text itself into the semantic fabric,enhancing the model’s capacity to understand and represent the data.Additionally,to address the intricate task of mining label correlations,we propose a Dual-end enhancement mechanism.This mechanism effectively mitigates the challenges of information loss and erroneous transmission inherent in traditional long short term memory propagation.We conducted an extensive battery of experiments on the AAPD and RCV1-2 datasets.These experiments serve the dual purpose of confirming the efficacy of both the Gate-Attention mechanism and the Dual-end enhancement mechanism.Our final model unequivocally outperforms the baseline model,attesting to its robustness.These findings emphatically underscore the imperativeness of taking into account not just external knowledge but also the inherent intricacies of textual data when crafting potent MLTC models.
