Multiple Labels Research Articles

Capturing age-related changes in the tumor microenvironment of luminal breast cancer using the multiplex immunohistochemistry technique, MILAN.

e15089 Background: The ‘immune landscape’ of luminal breast tumors is poorly described to date. We applied the ‘multiple iterative labeling by antibody ‘neodeposition’ (MILAN) technique for in-depth characterization of the tumor microenvironment of luminal-like breast cancer (BC) in relation to age. Methods: Newly diagnosed patients with early luminal BC (grade 2/3, ER+, HER2-, tumor size ≥1,5 cm) were prospectively included in UH Leuven between March 2014 and November 2015 and divided into 3 distinct age categories (young: 35-45 yrs.; middle: 55-65 yrs.; old: ≥70 yrs.). 55 patients (young: n = 12; middle: n = 16; old: n = 27) were included in the cohort. From each patient up to 2 cores from the tumor center (TC) and 2 cores from the invasive front (IF) were sampled from formalin-fixed and paraffin-embedded tumor resection specimens and embedded in tissue microarrays. The MILAN workflow included 20 rounds of three-color indirect immunofluorescence (FITC, TRITC, and Cy5 channel combined with 58 antibodies), image acquisition, and removal/stripping of the antibodies before any other stain. The MILAN data analysis pipeline was applied to study the expression of phenotypic and functional proteins at single-cell level including spatial resolution. These were compared between the different age groups using Wilcoxon's test (FDR correction). Results: We identified 1,5 million cells, of which 59% were epithelial cells (cancer and normal), 29% were immune cells, and the remaining 12% were endothelial or stromal cells. Within the immune cell fraction, T-helper (Th) cells, B-cells, T-cytotoxic (Tcy) cells, and NK-cells significantly decreased with aging while monocytes significantly increased in TC and IF (Table 1). The evaluation of the functional markers in the inflammatory cells revealed a decreased expression with aging of PD-L1 in Tcy cells (p < 0.001 for TC and IF), Th cells (p = 0.003 for IF), T-regulatory (Treg) cells (p = 0.038 for TC), monocytes (p < 0.001 for TC and IF), and B-cells (p = 0.023 for TC). Perforin expression significantly decreased with aging in Th cells (p < 0.05 for TC and IF) and monocytes (p < 0.001 for TC and IF). OX-40 expression significantly decreased with aging in Th cells (p = 0.001 for TC), Treg cells (p = 0.002 for TC), and B-cells (p = 0.012 for TC). Conclusions: Our results suggest important compositional and functional changes in the tumor inflammatory microenvironment during aging in patients with luminal BC. Further analysis on other cell types is warranted. [Table: see text]

Active learning in multi-label image classification with graph convolutional network embedding

Active learning has achieved considerable success in sample selection for deep learning models and has been widely used to address the issue of high-cost sample annotation. However, most of the existing active learning methods focus on single-label image classification and have limited use in multi-label scenarios. To address this issue and take advantage of label associations, we propose an active learning model based on the graph convolutional network (GCN) embedding and loss prediction network. Specifically, we construct a heterogeneous information network (HIN) that uses GCN embeddings to learn multiple label associations, as well as associations between images and labels. We also use a loss prediction network to predict target losses of unlabeled inputs. In addition, we propose a dynamic active coefficient to adjust the proportion of active learning gradually in the training process. Comprehensive multi-label image classification experiments with limited training labels are conducted on the MS-COCO, VOC 2007, and NUS-WIDE datasets. The comparison results demonstrate the superiority of our method compared with conventional methods in terms of both classification accuracy and convergence speed.

High-yield ligation-free assembly of DNA constructs with nucleosome positioning sequence repeats for single-molecule manipulation assays

Force and torque spectroscopy have provided unprecedented insights into the mechanical properties, conformational transitions, and dynamics of DNA and DNA-protein complexes, notably nucleosomes. Reliable single-molecule manipulation measurements require, however, specific and stable attachment chemistries to tether the molecules of interest. Here, we present a functionalization strategy for DNA that enables high-yield production of constructs for torsionally constrained and very stable attachment. The method is based on two subsequent PCRs: first ∼380bp long DNA strands are generated that contain multiple labels, which are used as "megaprimers" in a second PCR to generate ∼kbp long double-stranded DNA constructs with multiple labels at the respective ends. To achieve high-force stability, we use dibenzocyclooctyne-based click chemistry for covalent attachment to the surface and biotin-streptavidin coupling to the bead. The resulting tethers are torsionally constrained and extremely stable under load, with an average lifetime of 70± 3h at 45 pN. The high yield of the approach enables nucleosome reconstitution by salt dialysis on the functionalized DNA, and we demonstrate proof-of-concept measurements on nucleosome assembly statistics and inner turn unwrapping under force. We anticipate that our approach will facilitate a range of studies of DNA interactions and nucleoprotein complexes under forces and torques.

Kinetic Monte Carlo simulation study of the early stages of epitaxial SiC (0001) growth

A kinetic Monte Carlo (kMC) simulation with a cluster-multiple labeling technique was carried out to study the dynamic behavior of Si-C clusters during the early stage of epitaxial growth. In the model, a crystal lattice was established to fix the physical location of atoms and interatomic bonding. Events, such as adatoms adsorption, diffusion on the planar surface, and attachment to and detachment from clusters of adatoms were considered in the model. The Si and C atoms were treated individually to achieve more elaborate information for dynamic behavior of the clusters in the crystal surface. Moreover, a cluster multiple labeling technique, also known as the Hoshen - Kopelman algorithm was used to identify Si-C clusters. Then the properties of the clusters were calculated and analyzed by the algorithm. The simulation results showed that the number of Si-C clusters gradually decreases with increasing substrate temperature and decreasing deposition rate, and that the average size of Si-C clusters at high temperatures is much larger than that at low temperatures. Furthermore, the Si:C ratio on the growing surface has a significant effect on the surface morphology. At Si-rich conditions, the cluster density is relatively high, and the cluster shape is not homogeneously distributed. While with the increase of C mole fraction, the cluster density gradually decreases and the cluster shape becomes more uniform. Therefore, a higher concentration of C in deposition flux is beneficial to the early stage of nucleation. These findings are consistent with predictions for the case of the epitaxial growth of SiC.

Intracellular RNA and DNA tracking by uridine-rich internal loop tagging with fluorogenic bPNA

The most widely used method for intracellular RNA fluorescence labeling is MS2 labeling, which generally relies on the use of multiple protein labels targeted to multiple RNA (MS2) hairpin structures installed on the RNA of interest (ROI). While effective and conveniently applied in cell biology labs, the protein labels add significant mass to the bound RNA, which potentially impacts steric accessibility and native RNA biology. We have previously demonstrated that internal, genetically encoded, uridine-rich internal loops (URILs) comprised of four contiguous UU pairs (8 nt) in RNA may be targeted with minimal structural perturbation by triplex hybridization with 1 kD bifacial peptide nucleic acids (bPNAs). A URIL-targeting strategy for RNA and DNA tracking would avoid the use of cumbersome protein fusion labels and minimize structural alterations to the RNA of interest. Here we show that URIL-targeting fluorogenic bPNA probes in cell media can penetrate cell membranes and effectively label RNAs and RNPs in fixed and live cells. This method, which we call fluorogenic U-rich internal loop (FLURIL) tagging, was internally validated through the use of RNAs bearing both URIL and MS2 labeling sites. Notably, a direct comparison of CRISPR-dCas labeled genomic loci in live U2OS cells revealed that FLURIL-tagged gRNA yielded loci with signal to background up to 7X greater than loci targeted by guide RNA modified with an array of eight MS2 hairpins. Together, these data show that FLURIL tagging provides a versatile scope of intracellular RNA and DNA tracking while maintaining a light molecular footprint and compatibility with existing methods.

Semantic guided level-category hybrid prediction network for hierarchical image classification

Hierarchical classification (HC) assigns each object with multiple labels organized into a hierarchical structure. The existing deep learning-based HC methods usually predict an instance starting from the root node until a leaf node is reached. However, in the real world, images impaired by noise, occlusion, blur, or low resolution may not provide sufficient information for the classification at subordinate levels. To address this issue, we propose a novel Semantic Guided level-category Hybrid Prediction Network (SGHPN) that can jointly perform the level and category prediction in an end-to-end manner. SGHPN comprises two modules: a visual transformer that extracts feature vectors from the input images, and a semantic guided cross-attention module that uses categories word embeddings as queries to guide learning category-specific representations. In order to evaluate the proposed method, we construct two new datasets in which images are at a broad range of quality and thus are labeled to different levels (depths) in the hierarchy according to their individual quality. Experimental results demonstrate the effectiveness of our proposed HC method.

A deep learning approach for medication disposition and corresponding attributes extraction

ObjectiveThis article summarizes our approach to extracting medication and corresponding attributes from clinical notes, which is the focus of track 1 of the 2022 National Natural Language Processing (NLP) Clinical Challenges(n2c2) shared task. MethodsThe dataset was prepared using Contextualized Medication Event Dataset (CMED), including 500 notes from 296 patients. Our system consisted of three components: medication named entity recognition (NER), event classification (EC), and context classification (CC). These three components were built using transformer models with slightly different architecture and input text engineering. A zero-shot learning solution for CC was also explored. ResultsOur best performance systems achieved micro-average F1 scores of 0.973, 0.911, and 0.909 for the NER, EC, and CC, respectively. ConclusionIn this study, we implemented a deep learning-based NLP system and demonstrated that our approach of (1) utilizing special tokens helps our model to distinguish multiple medications mentions in the same context; (2) aggregating multiple events of a single medication into multiple labels improves our model’s performance.

Multi-layered semantic representation network for multi-label image classification

Multi-label image classification is a fundamental and practical task, which aims to assign multiple possible labels to an image. In recent years, many deep convolutional neural network (CNN) based approaches have been proposed which model label correlations to discover semantics of labels and learn semantic representations of images. This paper advances this research direction by improving both the modeling of label correlations and the learning of semantic representations. On the one hand, besides the local semantics of each label, we propose to further explore global semantics shared by multiple labels. On the other hand, existing approaches mainly learn the semantic representations at the last convolutional layer of a CNN. But it has been noted that the image representations of different layers of CNN capture different levels or scales of features and have different discriminative abilities. We thus propose to learn semantic representations at multiple convolutional layers. To this end, this paper designs a Multi-layered Semantic Representation Network (MSRN) which discovers both local and global semantics of labels through modeling label correlations and utilizes the label semantics to guide the semantic representations learning at multiple layers through an attention mechanism. Extensive experiments on five benchmark datasets including VOC2007, VOC2012, MS-COCO, NUS-WIDE, and Apparel show a competitive performance of the proposed MSRN against state-of-the-art models.

Developing a semi-supervised learning and ordinal classification framework for quality level prediction in manufacturing

The authors of this work propose a novel semi-supervised learning framework for quality prediction in manufacturing. Semi-supervised learning is a promising method in neural network applications, where label generation incurs significant time and cost. However, a semi-supervised learning mechanism in the manufacturing industry has not been as popular as a supervised learning method, especially in quality prediction tasks. The proposed framework trains a deep neural network-based model in a self-training scheme that uses filtered unlabeled data based on prediction confidence. In particular, the framework successfully handles ordinal classification tasks by using ordinal label rendering based on a state-of-the-art technique called the soft ordinal vector (SORD) that reflects ordinality in multiple labels. Furthermore, a newly proposed information measure named ordinal entropy, which takes ordinality into account, is used to selectively utilize confident labels among generated pseudo-labels. The proposed framework’s efficacy is validated through a case study using real-world data from the ultraviolet (UV) lamp manufacturing process. The proposed framework has shown better performance in quality prediction than the learned model with only supervision. In addition, various configurations of the proposed framework have been validated with extensive experiments.

Substance P expression and neurochemical characterization in the enteric nervous system of the porcine colon

The pig is a good model for studying intestinal functions and disorders for its homologies with humans such as microbiome composition, size, nutrition being both omnivores and colon fermenters. The enteric nervous system (ENS) of pigs and humans has a multilayered submucosal plexus with an inner submucous plexus (ISP) near the mucosa and an outer plexus (OSP) near the circular muscle in addition to the myenteric plexus (MP) between the muscle layers. We have shown differences in the density and distribution of functionally distinct neurons in different regions and plexuses of the porcine colon. Aim: This study focused on Substance P, a peptide that modulates many functions in the gastrointestinal (GI) tract and plays an important role in neurogenic inflammation. We tested whether there were differences in the density and neurochemical profile of SP neurons in the ISP, OSP and MP of the ascending (AC) and descending (DC) colon of 15 Yucatan minipigs (12M, 3F, 7-months-old, body weight 25-30 kg). We processed colonic wholemounts for multiple labeling immunofluorescence using the HuCD, choline acetyltransferase (ChAT) and neuronal nitric oxide synthase (nNOS) as neuronal markers with high resolution confocal microscopy and Imaris software to quantify the number of neurons/mm2 and the % of total neurons identified by the pan-neuronal marker HuCD. Results: HuCD/SP-immunoreactive (IR) neurons were most abundant in the ISP vs. OSP and MP in both AC and DC (p<0.01-p<0.0001) with highest density in ISP of AC (p<0.0001 vs. DC). SP-IR neurons represent 27-28% of HuCD-IR neurons in ISP followed by OSP (19-22%) and MP (13-17%, p<0.05- p<0.0001) in AC and DC. The highest proportion of SP-IR neurons contains ChAT-IR (62-67% in the MP, 74-76% in the ISP, 70-85% in OSP) with no significant differences in the AC and DC. By contrast, SP-IR neurons containing nNOS-IR were much less abundant in both AC and DC with the highest peak in the OSP 36% vs. 16% in the ISP, p<0.01, and 20% in the MP in AC, whereas in the DC the highest peak was in the ISP (31%) followed by OSP and MP (25% and 18%) without significant differences. There was a small population of SP-/ChAT-/nNOS-IR enteric neurons in all plexuses in AC and DC. Varicose SP-IR fibers were abundant in the ganglia and distributed to the muscle layers and mucosa. Conclusions: SP-IR neurons include excitatory, inhibitory and interneurons based on the co-expression of ChAT, nNOS or both. The different density and neurochemical profile likely reflect different functions of SP-IR neurons such as immune modulation, regulation of mucosal function and control of vascular, myogenic and neurogenic activities. Mapping neuromodulators expression in a pre-clinical model provides the basis for elucidating neuronal circuits underlying GI functions and advance our understanding of human diseases. Supported by NIH SPARC OT2OD24899 & NIH-P30DK41301. This is the full abstract presented at the American Physiology Summit 2023 meeting and is only available in HTML format. There are no additional versions or additional content available for this abstract. Physiology was not involved in the peer review process.

Triplet attention and dual-pool contrastive learning for clinic-driven multi-label medical image classification.

Multi-label classification (MLC) can attach multiple labels on single image, and has achieved promising results on medical images. But existing MLC methods still face challenging clinical realities in practical use, such as: (1) medical risks arising from misclassification, (2) sample imbalance problem among different diseases, (3) inability to classify the diseases that are not pre-defined (unseen diseases). Here, we design a hybrid label to improve the flexibility of MLC methods and alleviate the sample imbalance problem. Specifically, in the labeled training set, we remain independent labels for high-frequency diseases with enough samples and use a hybrid label to merge low-frequency diseases with fewer samples. The hybrid label can also be used to put unseen diseases in practical use. In this paper, we propose Triplet Attention and Dual-pool Contrastive Learning (TA-DCL) for multi-label medical image classification based on the aforementioned label representation. TA-DCL architecture is a triplet attention network (TAN), which combines category-attention, self-attention and cross-attention together to learn high-quality label embeddings for all disease labels by mining effective information from medical images. DCL includes dual-pool contrastive training (DCT) and dual-pool contrastive inference (DCI). DCT optimizes the clustering centers of label embeddings belonging to different disease labels to improve the discrimination of label embeddings. DCI relieves the error classification of sick cases for reducing the clinical risk and improving the ability to detect unseen diseases by contrast of differences. TA-DCL is validated on two public medical image datasets, ODIR and NIH-ChestXray14, showing superior performance than other state-of-the-art MLC methods. Code is available at https://github.com/ZhangYH0502/TA-DCL.

Tag Recommendation System for Marathi News Articles by using Multi-label Classification

Abstract: Multi-label classification is the variant of a classification problem where multiple labels are assigned to each instance. In multi-label classification, the training set is composed of instances each associated with a set of labels, and the task is to predict the label sets of unseen instances through analyzing training instances with known label sets. This paper demonstrates the use of multi-label classification to determine tags for news articles written in the Marathi Language of India. The proposed study uses Binary Relevance (One vs Rest) technique of multi-label classification to establish the tags for the given input of a Marathi news article. Tag recommendation systems for Marathi news articles can greatly enhance the user experience for readers and help them find the articles that are most relevant to their interests.

Learning from crowds with robust logistic regression

Crowdsourcing systems provide an easy way to obtain labels for data. Each instance in data will usually be labeled by multiple crowd labelers who are not experts. Thus, it is very important to design considerate ground truth inference algorithms to infer integrated labels from multiple crowd labels. While almost all ground truth inference algorithms show good performance when the number of crowd labels is large, few algorithms can perform well with few crowd labels. This paper considers how to deal with noise in multiple crowd labels as a key to good ground truth inference. This paper solves ground truth inference using robust classifiers. This paper proposes two versions of ground truth inference algorithm based on robust logistic regression to solve the following two problems: (1) how to embed noise level into the loss function of logistic regression and (2) how to estimate the parameters that model noise level in the crowdsourcing scenario. We call our algorithms robust logistic regression inference (RLRI). By employing the idea of robust classifiers, RLRI can still perform well in the case of a small number of labels. We also theoretically compare the advantages and disadvantages of the two versions of RLRI. Finally, the performance of our algorithms is verified on benchmark and real-world datasets.

Towards Safe Visual Navigation of a Wheelchair Using Landmark Detection

This article presents a method for extracting high-level semantic information through successful landmark detection using 2D RGB images. In particular, the focus is placed on the presence of particular labels (open path, humans, staircase, doorways, obstacles) in the encountered scene, which can be a fundamental source of information enhancing scene understanding and paving the path towards the safe navigation of the mobile unit. Experiments are conducted using a manual wheelchair to gather image instances from four indoor academic environments consisting of multiple labels. Afterwards, the fine-tuning of a pretrained vision transformer (ViT) is conducted, and the performance is evaluated through an ablation study versus well-established state-of-the-art deep architectures for image classification such as ResNet. Results show that the fine-tuned ViT outperforms all other deep convolutional architectures while achieving satisfactory levels of generalization.

Neighborhood Weighted Voting-Based Noise Correction for Crowdsourcing

In crowdsourcing scenarios, we can obtain each instance’s multiple noisy labels set from different crowd workers and then use a ground truth inference algorithm to infer its integrated label. Despite the effectiveness of ground truth inference algorithms, a certain level of noise still remains in the integrated labels. To reduce the impact of noise, many noise correction algorithms have been proposed in recent years. To the best of our knowledge, however, nearly all existing noise correction algorithms only exploit each instance’s own multiple noisy label sets but ignore the multiple noisy label sets of its neighbors. Here neighbors refer to the nearest instances found in the feature space based on the distance metric learning. In this article, we propose neighborhood weighted voting-based noise correction (NWVNC). In NWVNC, we at first take advantage of the multiple noisy label sets of each instance’s neighbors (including itself) to estimate the probability that it belongs to its integrated label. Then, we use the estimated probability to identify and filter noise instances and thus obtain a clean set and a noise set. Finally, we train three heterogeneous classifiers on the clean set and correct the noise instances by the consensus voting of three trained classifiers. The experimental results on 34 simulated and two real-world crowdsourced datasets show that NWVNC significantly outperforms all the other state-of-the-art noise correction algorithms used for comparison.

Global-guided weakly-supervised learning for multi-label image classification

Multi-label classification with region-free labels is attracting increasing attention compared to that with region-based labels due to the time-consuming manual region-labeling process. Existing methods usually employ attention-based technology to discover the conspicuous label-related regions in a weakly-supervised manner with only image-level region-free labels, while the region covering is not precise without exploring global clues of multi-level features. To address this issue, a novel Global-guided Weakly-Supervised Learning (GWSL) method for multi-label classification is proposed. The GWSL first extracts the multi-level features to estimate their global correlation map which is further utilized to guide feature disentanglement in the proposed Feature Disentanglement and Localization (FDL) networks. Specifically, the FDL networks then adaptively combine the different correlated features and localize the fine-grained features for identifying multiple labels. The proposed method is optimized in an end-to-end manner under weakly supervision with only image-level labels. Experimental results demonstrate that the proposed method outperforms the state-of-the-arts for multi-label learning problems on several publicly available image datasets. To facilitate similar researches in the future, the codes are directly available online at https://github.com/Yong-DAI/GWSL.

Cost-constrained feature selection in multilabel classification using an information-theoretic approach

Feature selection is one of the key steps in building a predictive model in multi-label classification. However, most of the existing methods do not take into account information about the costs associated with considered features, such as the costs of performing diagnostic medical tests. We consider a problem of cost-constrained multilabel feature selection, which aims to select a feature subset relevant to multiple labels while satisfying a user-specific maximal admissible budget. This approach allows for building a model with high predictive power, for which the cost of making a prediction for a single instance does not exceed the user-specified budget. In this problem, the balance between the feature subset relevance and its cost should be considered concurrently, which is nontrivial in practice because their optimal balance is unknown. In this paper, we propose a novel criterion for cost-constrained multilabel feature selection that combines the relevance and cost of the candidate feature. The relevance measure is derived using the lower bound of the mutual information between the feature subset and label vector. Moreover, we propose an effective method for determining the cost-factor value that controls the trade-off between relevancy and cost. The experimental results on multilabel datasets with various characteristics demonstrate the superiority of the proposed method over conventional methods.

Multiscale Convolutional Neural Network Based on Channel Space Attention for Gearbox Compound Fault Diagnosis.

Gearboxes are one of the most widely used speed and power transfer elements in rotating machinery. Highly accurate compound fault diagnosis of gearboxes is of great significance for the safe and reliable operation of rotating machinery systems. However, traditional compound fault diagnosis methods treat compound faults as an independent fault mode in the diagnosis process and cannot decouple them into multiple single faults. To address this problem, this paper proposes a gearbox compound fault diagnosis method. First, a multiscale convolutional neural network (MSCNN) is used as a feature learning model, which can effectively mine the compound fault information from vibration signals. Then, an improved hybrid attention module, named the channel-space attention module (CSAM), is proposed. It is embedded into the MSCNN to assign weights to multiscale features for enhancing the feature differentiation processing ability of the MSCNN. The new neural network is named CSAM-MSCNN. Finally, a multilabel classifier is used to output single or multiple labels for recognizing single or compound faults. The effectiveness of the method was verified with two gearbox datasets. The results show that the method possesses higher accuracy and stability than other models for gearbox compound fault diagnosis.

Multilabeled and topological versions of the Hex theorem

The purpose of our paper is to give a purely combinatorial and algorithmic proof of the Hex theorem with multiple labels. Moreover we extend this result for a new class of complexes, called n-essential. We also present a topological version of the Hex theorem and explain its relation to the fixed point property, the Poincaré-Miranda theorem and the covering (Lebesgue) dimension.

5613458 BLOOD MICROPARTICLES: A TOOL OF EVALUATING THE EFFECT OF TRANSFUSION IN SICKLE CELL DISEASE PATIENTS

Background: Blood microparticles (MPs) are phospholipid microvesicles derived from cell membranes. Typically, they range in size from 0.1 μm to 1.0 μm and they are released upon oxidative stress, cell activation, injury or apoptosis. MPs have acquired a biological and clinical interest and are recognized as entities able to modulate many biological functions. Literature suggests that they may have a role in the pathophysiology and downstream manifestations of sickle cell disease (SCD). Aims: This present study was carried out to determine MPs profiles in patients with SCD before and after transfusion and to compare these profiles with healthy subjects. Materials: Twenty healthy donors and twelve polytransfused patients with SCD recruited from National Bone Marrow Transplant Center were recruited in this study. None of the patients was transfused for at least 3 weeks before sampling. The blood samples obtained were analyzed for circulating MPs by flow cytometry. The cellular origin was determined by a multiple labeling with specific antibodies and dyes. Results: A high level of apoptotic, platelets, and endothelial MPs with the predominance of the number of erythrocyte MPs was detected in SCD patients compared to healthy donors. This suggests that erythrocytes MPs may contribute to thrombotic risk in sickle cell patients. Moreover, after transfusion most subjects register a decreased number of different types of MPs. Conclusion: Transfusion is an effective treatment option in patients with SCD, as it seems that it decreases the abnormally elevated levels of MPs.