Class Of Representations Research Articles

Clean-Label Backdoor Watermarking for Dataset Copyright Protection via Trigger Optimization

High-quality datasets are essential for training high-performance models, while the process of collection, cleaning, and labeling is costly. As a result, datasets are considered valuable intellectual property. However, when security mechanisms are symmetry-breaking, creating exploitable vulnerabilities, unauthorized use or data leakage can infringe on the copyright of dataset owners. In this study, we design a method to mount clean-label dataset watermarking based on trigger optimization, aiming to protect the copyright of the dataset from infringement. We first perform iterative optimization of the trigger based on a surrogate model, with targets class samples guiding the updates. The process ensures that the optimized triggers contain robust feature representations of the watermark target class. A watermarked dataset is obtained by embedding optimized triggers into randomly selected samples from the watermark target class. If an adversary trains a model with the watermarked dataset, our watermark will manipulate the model’s output. By observing the output of the suspect model on samples with triggers, it can be determined whether the model was trained on the watermarked dataset. The experimental results demonstrate that the proposed method exhibits high imperceptibility and strong robustness against pruning and fine-tuning attacks. Compared to existing methods, the proposed method significantly improves effectiveness at very low watermarking rates.

A cognitive few-shot learning for medical diagnosis: A case study on cleft lip and palate and Parkinson’s disease

This paper presents a novel cognitive few-shot learning (CFSL) for the diagnosis of cleft lip and palate and Parkinson’s diseases. The proposed method utilizes computational analysis of paralinguistic features to expedite the diagnostic process. Unlike other methods that rely on complex and fragmented representations, CFSL trains itself to recognize patterns that are easily interpretable by humans. Rather than learning a single, unstructured metric space, CFSL combines the outputs of individual landmark (LM) learners by mapping LMs into semi-formation spaces. In order to assess the effectiveness of CFSL, we conducted a comparative analysis with seven distinct FSL-based models, including momentum contrastive learning for FSL (MCFSL), self-updating FSL (SUFSL), mutual info multi-attention FSL (MAMIFSL), dual class representation FSL (DCRFSL), Improved FSL (IFSL), meta-knowledge for FSL (MKFSL), and prototypical networks (ProtoNet), using three popular datasets, namely GPRS, CIEMPIESS, and PC-GITA. The findings indicate that CFSL demonstrated superior performance compared to the highest-performing baseline frameworks for the 5-shot (5-sh) and 1-shot (1-sh), having an average enhancement of 4.39% and 4.49%, respectively. CFSL demonstrated better performance than the ProtoNet baseline in both 1-sh and 5-sh across all datasets, with an improvement of 12.966% and 11.033%, respectively. In addition, we performed ablation tests to assess the effects of variables such as the density of LMs, the structure of the network, the distance measure used, and the positioning of LMs. The CFSL approach, if adopted in hospitals, has the potential to enhance the precision and efficiency of diagnosis for cleft lip and palate as well as Parkinson’s disease.

DyConfidMatch: Dynamic thresholding and re-sampling for 3D semi-supervised learning

Semi-supervised learning (SSL) leverages limited labeled and abundant unlabeled data but often faces challenges with data imbalance, especially in 3D contexts. This study investigates class-level confidence as an indicator of learning status in 3D SSL, proposing a novel method that utilizes dynamic thresholding to better use unlabeled data, particularly from underrepresented classes. A re-sampling strategy is also introduced to mitigate bias towards well-represented classes, ensuring equitable class representation. Through extensive experiments in 3D SSL, our method surpasses state-of-the-art counterparts in classification and detection tasks, highlighting its effectiveness in tackling data imbalance. This approach presents a significant advancement in SSL for 3D datasets, providing a robust solution for data imbalance issues.

A Semantic Segmentation Method for Winter Wheat in North China Based on Improved HRNet

Winter wheat is one of the major crops for global food security. Accurate statistics of its planting area play a crucial role in agricultural policy formulation and resource management. However, the existing semantic segmentation methods for remote sensing images are subjected to limitations in dealing with noise, ambiguity, and intra-class heterogeneity, posing a negative impact on the segmentation performance of the spatial distribution and area of winter wheat fields in practical applications. In response to the above challenges, we proposed an improved HRNet-based semantic segmentation model in this paper. First, this model incorporates a semantic domain module (SDM), which improves the model’s precision of pixel-level semantic parsing and reduces the interference from noise through multi-confidence scale class representation. Second, a nested attention module (NAM) is embedded, which enhances the model’s capability of recognizing correct correlations in pixel classes. The experimental results show that the proposed model achieved a mean intersection over union (mIoU) of 80.51%, a precision of 88.64%, a recall of 89.14%, an overall accuracy (OA) of 90.12%, and an F1-score of 88.89% on the testing set. Compared to traditional methods, our model demonstrated better segmentation performance in winter wheat semantic segmentation tasks. The achievements of this study not only provide an effective tool and technical support for accurately measuring the area of winter wheat fields, but also have important practical value and profound strategic significance for optimizing agricultural resource allocation and achieving precision agriculture.

Gender Representation in Class Ten Nepali Language Textbook

The present study is based on the analysis of gender representation in textbook. It aims to analyze the situation of gender representation in the texts included in the grade X compulsory Nepali textbook of CDC, GON Nepal. Qualitative framework has been adopted in this study. After an in-depth study of the lessons of the Nepali textbook of grade X, the data has been noted and tabulated. For this, the paradigms created by Logsdon (1985) for the analysis of gender representation have been made a basis. It has been analyzed using the textual method. Out of the seventeen texts included in the textbook of class ten, only one text is written by a female author. In terms of occupation, women are shown only in a limited field compared to men. Examples, contexts and events include a wide range of masculine names and a limited range of feminine names. Overall, it is seen that the textbook is not balanced in terms of gender representation. This study has guided the revision of this book in the future to maintain a balance in it and to pay attention to the role of textbooks in developing the sense of gender equality in children.

Купечество Вологодской губернии в структуре городского самоуправления в последней трети XIX века

The paper deals with the correlation between the proportion of merchants in the total urban population of the Vologda province in the second third of the XIX century and the degree of representation of the merchant class in local self-government after the approval of the Town Regulations in 1870 and then in 1892. Based on materials published in the “Vologda Provincial Gazette” in the 1870s–1890s, the share of the merchant class in the number of electors in the towns of the Vologda province and the share of the merchant class in the number of elected members of the municipal Dumas were calculated and summarized in tabular and graphical form. The results of the analysis not only confirmed the widespread participation of merchants both in electoral processes as voters and in local self-government bodies – municipal Dumas, but also revealed certain patterns. Thus, the more developed the town was in terms of trade and industry, the more representative was the share of merchants among the members of the Duma, even despite the modest share of the merchant class in the total population of the town. And, conversely, in towns where trade and industry were rather of local importance, the share of merchants among members of the municipal Dumas was only slightly higher than their share in the total population of the town. The changing role of the town in the economic life of the province was also reflected in the degree of representation of the merchant class in the municipal Duma.

BENN: Balanced Ensemble Neural Network for Handling Class Imbalance in Big Data

ABSTRACTClass imbalance is a critical challenge in big data analytics, often leading to biased predictive models. This imbalance can lead to biased models that perform well on the majority class but poorly on the minority class. Many machine learning models tend to be biased towards the majority class because they aim to minimise overall error, often leading to poor performance on the minority class. This paper presents the balanced ensemble neural network, a novel solution to effectively address class imbalance in big data. Balanced ensemble neural network combines the robust capabilities of neural networks with the power of ensemble learning, incorporating class balancing strategies to ensure fair representation of minority classes. The methodology involves integrating multiple neural networks, each trained on balanced subsets of data using techniques like Synthetic Minority Over‐sampling Technique and Random Undersampling. This integration aims to leverage the strengths of individual networks while reducing their inherent biases. Our extensive experiments across various datasets reveal that BENN achieves an AUC‐ROC score of 0.94, surpassing other models such as random forest (0.88), support vector (0.84) and single neural net (0.80). It was also observed that BENN's performance is better compared to traditional neural network models and standard ensemble methods in key metrics like accuracy, precision, recall, F1‐score and AUC‐ROC. The results specifically highlight BENN's effectiveness in accurately classifying instances of minority classes, a notable challenge in many existing models. These findings underscore BENN's potential as a substantial advancement in handling class imbalance within big data environments, offering a promising direction for future research and application in machine learning.

Representation of Women and Class in Shakespeare’s <i>Othello</i>

Shakespeare’s plays reflect his society by imitating events that shaped the socio-economic, historical and political development of the Elizabethan era. This study examines the representation of women and class in the tragic play; Othello. Shakespeare juxtaposes male and female characters to understand their roles seeking to understand the complexities that shapes gender and class in society. He lived in a patriarchal society where women were subjugated, relegated, and oppressed. They were only good as daughters or wives; Shakespeare in some of his plays creates strong female characters that exhibit a new perspective in the representation of women who break away from the norm by their actions or inactions in Elizabethan England. The theory used for the analysis is New Historicism which seeks to reveal and challenge the depiction of women and class during the reign of Queen Elizabeth the first. New historicism is a practice of interpreting texts in their historical context or concerning the time and place in which they were written according to Greenblatt. Women and class have often been misrepresented in Shakespeare’s time but in his play Othello, they are found to break down the shackles of patriarchy to make choices and stand by it.

On Matrix Representations of Homeomorphism Classes

In this study, we investigate the matrix representations of homeomorphism classes. Considering well-known concepts such as the matrix of ones, column-sum, row-sum, one's complement, Hadamard product, and regular addition for matrices, we explore binary matrices' relationships with subsets of a finite set. The main results establish connections between matrix operations and set operations, providing insights into the structure of homeomorphism classes. The paper concludes with the formulation of a topology on a set based on specific matrix conditions.

Multivariate prototype representation for domain-generalized incremental learning

Deep learning models often suffer from catastrophic forgetting when fine-tuned with samples of new classes. This issue becomes even more challenging when there is a domain shift between training and testing data. In this paper, we address the critical yet less explored Domain-Generalized Class-Incremental Learning (DGCIL) task. We propose a DGCIL approach designed to memorize old classes, adapt to new classes, and reliably classify objects from unseen domains. Specifically, our loss formulation maintains classification boundaries while suppressing domain-specific information for each class. Without storing old exemplars, we employ knowledge distillation and estimate the drift of old class prototypes as incremental training progresses. Our prototype representations are based on multivariate Normal distributions, with means and covariances continually adapted to reflect evolving model features, providing effective representations for old classes. We then sample pseudo-features for these old classes from the adapted Normal distributions using Cholesky decomposition. Unlike previous pseudo-feature sampling strategies that rely solely on average mean prototypes, our method captures richer semantic variations. Experiments on several benchmarks demonstrate the superior performance of our method compared to the state of the art.

The timing and mode of southern Andean human migrations

Abstract While recent genomic and isotopic information show that migration has been pervasive along human history, southern Andean archaeology has largely overlooked its importance in shaping human trajectories of sociocultural change. Building on previous isotopic research that identified the presence of migrant farmers in the Uspallata Valley (Mendoza, Argentina), we present chronological and bioarchaeological results that help to characterize the timing and mode of human migrations in the southern Andes. The burials with migrants show the representation of the different age classes, including a high abundance of children, as well as both men and women, suggesting that family groups were likely involved. The Bayesian modeling of 16 direct dates for migrants indicates that these migrations started between 1210–1275 CE (median 1255 CE) and finished at 1320–1425 CE (median 1360 CE), indicating that there is nearly no overlap between the commencement of this migration phase and the southwards expansion of the Inka Empire. The model defines a diachronic process that lasted between 55 and 195 years, implying that migration to Uspallata was a multi-generational process that involved between two and eight generations (median of four generations). Our contextual, bioarchaeological and chronological evidence indicates that the conditions fostering migration to Uspallata were sustained through time, inviting to explore persisting push-pull dynamics acting during this period. 87Sr/86Sr results show that migration occurred across the daily territories of these groups and may have involved movement across social or ethnic frontiers.

Category Mapping of Emergency Supplies Classification Standard Based on BERT-TextCNN

In recent years, the escalation in emergency occurrences has underscored the pressing need for expedient responses in delivering essential supplies. Efficient integration and precise allocation of emergency resources under joint government–enterprise stockpiling models are pivotal for enhancing emergency response effectiveness and minimizing economic repercussions. However, current research predominantly focuses on contract coordination and cost-sharing within these joint reserve modes, overlooking significant discrepancies in emergency supply classification standards between government and enterprise sectors, as well as the asymmetry in cross-sectoral and cross-regional supply information. This oversight critically impedes the timeliness and accuracy of emergency supply responses. In practice, manual judgment has been used to match the same materials under differing classification standards between government and enterprise reserves. Still, this approach is inefficient and prone to high error rates. To mitigate these challenges, this study proposes a methodology leveraging the BERT pre-trained language model and TextCNN neural network to establish a robust mapping relationship between these classification criteria. The approach involves abstracting textual representations of both taxonomical classes, generating comparable sentence vectors via average pooling, and calculating cosine similarity scores to facilitate precise classification mapping. Illustrated with China’s Classification and Coding of Emergency Supplies standards and Global Product Classification standards, empirical validation on annotated data demonstrates the BERT-TextCNN model’s exceptional accuracy of 98.22%, surpassing other neural network methodologies such as BERT-CNN, BERT-RNN, BERT-BiLSTM, etc. This underscores the potential of advanced neural network techniques in enhancing emergency supply management across diverse sectors and regions.

Revolutionizing breast ultrasound diagnostics with EfficientNet-B7 and Explainable AI

Breast cancer is a leading cause of mortality among women globally, necessitating precise classification of breast ultrasound images for early diagnosis and treatment. Traditional methods using CNN architectures such as VGG, ResNet, and DenseNet, though somewhat effective, often struggle with class imbalances and subtle texture variations, leading to reduced accuracy for minority classes such as malignant tumors. To address these issues, we propose a methodology that leverages EfficientNet-B7, a scalable CNN architecture, combined with advanced data augmentation techniques to enhance minority class representation and improve model robustness. Our approach involves fine-tuning EfficientNet-B7 on the BUSI dataset, implementing RandomHorizontalFlip, RandomRotation, and ColorJitter to balance the dataset and improve model robustness. The training process includes early stopping to prevent overfitting and optimize performance metrics. Additionally, we integrate Explainable AI (XAI) techniques, such as Grad-CAM, to enhance the interpretability and transparency of the model’s predictions, providing visual and quantitative insights into the features and regions of ultrasound images influencing classification outcomes. Our model achieves a classification accuracy of 99.14%, significantly outperforming existing CNN-based approaches in breast ultrasound image classification. The incorporation of XAI techniques enhances our understanding of the model’s decision-making process, thereby increasing its reliability and facilitating clinical adoption. This comprehensive framework offers a robust and interpretable tool for the early detection and diagnosis of breast cancer, advancing the capabilities of automated diagnostic systems and supporting clinical decision-making processes.

“Des œuvres solides et vivantes qu’il rêvait”: Rethinking Claude Lantier’s Chef d’œuvre

Abstract: Scholars have recently prioritized how the suppressed Paris Commune haunts the narrative, the characters, and even descriptions of the bloody, dying foods of Les Halles in Zola’s Le Ventre de Paris . This article considers one of those still lifes, Claude Lantier’s self-proclaimed triumph, a reorganisation of his aunt Lisa’s charcuterie display case intended to cause indigestion. Claude is an aesthetic insurgent who denies any social or political significance for his work, but we should not take him at his word. The fictional artist misreads the anxiety that his display produces as its audience’s inability to appreciate modernism. In fact, Claude is the bad viewer. He misses the symbolism of his work, the impact of its visceral appeal, and the socio-political importance of both. This article argues that we should not also overlook these by privileging the formal stakes of Claude’s intervention or its failure as painting. Zola rivals his character in his own representation of violence, nausea, and class war through food. In the novel’s terms, Claude’s display is a genuine success that raises questions about the role and conditions of art-making in the context of clashing Fats and Thins.

Semantic Interaction Meta-Learning Based on Patch Matching Metric.

Metric-based meta-learning methods have demonstrated remarkable success in the domain of few-shot image classification. However, their performance is significantly contingent upon the choice of metric and the feature representation for the support classes. Current approaches, which predominantly rely on holistic image features, may inadvertently disregard critical details necessary for novel tasks, a phenomenon known as "supervision collapse". Moreover, relying solely on visual features to characterize support classes can prove to be insufficient, particularly in scenarios involving limited sample sizes. In this paper, we introduce an innovative framework named Patch Matching Metric-based Semantic Interaction Meta-Learning (PatSiML), designed to overcome these challenges. To counteract supervision collapse, we have developed a patch matching metric strategy based on the Transformer architecture to transform input images into a set of distinct patch embeddings. This approach dynamically creates task-specific embeddings, facilitated by a graph convolutional network, to formulate precise matching metrics between the support classes and the query image patches. To enhance the integration of semantic knowledge, we have also integrated a label-assisted channel semantic interaction strategy. This strategy merges word embeddings with patch-level visual features across the channel dimension, utilizing a sophisticated language model to combine semantic understanding with visual information. Our empirical findings across four diverse datasets reveal that the PatSiML method achieves a classification accuracy improvement of 0.65% to 21.15% over existing methodologies, underscoring its robustness and efficacy.

Representation of sequence classes by operator ideals

It is well known that weakly p-summable sequences in a Banach space E are associated to bounded operators from ℓp⁎ to E, and unconditionally p-summable sequences in E are associated to compact operators from ℓp⁎ to E. Generalizing these results to a quite wide environment, we characterize the classes of Banach spaces-valued sequences that are associated to (or represented by) some Banach operator ideal. Using these characterizations, we decide, among all sequence classes that usually appear in the literature, which are represented by some Banach operator ideal and which are not. Moreover, to each class that is represented by some Banach operator ideal, we show an ideal that represents it. Illustrative examples and additional applications are provided.

Customised-sampling approach for pipe failure prediction in water distribution networks

This paper presents a new methodology for addressing imbalanced class data for failure prediction in Water Distribution Networks (WDNs). The proposed methodology relies on existing approaches including under-sampling, over-sampling, and class weighting as primary strategies. These techniques aim to treat the imbalanced datasets by adjusting the representation of minority and majority classes. Under-sampling reduces data in the majority class, over-sampling adds data to the minority class, and class weighting assigns unequal weights based on class counts to balance the influence of each class during machine learning (ML) model training. In this paper, the mentioned approaches were used at levels other than “balance point” to construct pipe failure prediction models for a WDN with highly imbalanced data. F1-score, and AUC–ROC, were selected to evaluate model performance. Results revealed that under-sampling above the balance point yields the highest F1-score, while over-sampling below the balance point achieves optimal results. Employing class weights during training and prediction emphasises the efficacy of lower weights than the balance. Combining under-sampling and over-sampling to the same ratio for both majority and minority classes showed limited improvement. However, a more effective predictive model emerged when over-sampling the minority class and under-sampling the majority class to different ratios, followed by applying class weights to balance data.

Shape-shifting versions of class in Australia and the pursuit of equity in public health.

The COVID-19 pandemic and current cost of living crisis have highlighted socioeconomically patterned health disparities, bringing renewed focus on equity in public health. Despite political rhetoric invoking cultural narratives of egalitarianism and opportunities for class mobility, social class remains a significant factor in health outcomes in the Australian context. For social scientists, class (despite robust critiques) is a key analytical concept that has been theoretically broadened to encompass social and cultural practices (habitus). In public health, however, concepts of social disadvantage have expanded toward frames such as health equity and socioeconomic status in ways that can obscure 'class' and habitus. Understandings and operationalization of concepts of class and equity not only impact collaborative and interdisciplinary relationships, but also the framing of public health problems and health promotion interventions and policies. In this article, we draw on our experiences as anthropologists conducting ethnography in and of Australian health promotion programs to map and re-evaluate the intersection of concepts of social class and equity. We trace how representations of class emerged in these programs, and the versions of class and equity that materialized across different public health contexts. We argue for a conceptual repositioning of class that recognizes its shape-shifting qualities and of its materializations in different politics, disciplines and everyday contexts. In doing so, we highlight 'class' as a salient dimension of the design, implementation and evaluation of health promotion programs.

Some Properties of (M,k)-quasi Paranormal Operators on Hilbert Spaces

Let H be a complex Hilbert space and let T represent a bounded linear operator on H. In this paper we introduce, a new class of non-normal operators, the (M,k)-quasi paranormal operator. An operator T is said to be (M,k)-quasi paranormal operator, for a non-negative integer k and a real positive number M, if it satisfies: ||Tk+1x||2 < M ||Tk+2x|| ||Tkx||, for every x in H. This new class of operators is a generalization of some of the non-normal operators, such as the k-quasi paranormal and M-paranormal operators. We prove the basic properties, the structural and spectral properties and also the matrix representation of this new class of operators.

SFPL: Sample-specific fine-grained prototype learning for imbalanced medical image classification

Imbalanced classification is a common and difficult task in many medical image analysis applications. However, most existing approaches focus on balancing feature distribution and classifier weights between classes, while ignoring the inner-class heterogeneity and the individuality of each sample. In this paper, we proposed a sample-specific fine-grained prototype learning (SFPL) method to learn the fine-grained representation of the majority class and learn a cosine classifier specifically for each sample such that the classification model is highly tuned to the individual’s characteristic. SFPL first builds multiple prototypes to represent the majority class, and then updates the prototypes through a mixture weighting strategy. Moreover, we proposed a uniform loss based on set representations to make the fine-grained prototypes distribute uniformly. To establish associations between fine-grained prototypes and cosine classifier, we propose a selective attention aggregation module to select the effective fine-grained prototypes for final classification. Extensive experiments on three different tasks demonstrate that SFPL outperforms the state-of-the-art (SOTA) methods. Importantly, as the imbalance ratio increases from 10 to 100, the improvement of SFPL over SOTA methods increases from 2.2% to 2.4%; as the training data decreases from 800 to 100, the improvement of SFPL over SOTA methods increases from 2.2% to 3.8%.