Low-level Labeling Research Articles

Improve label embedding quality through global sensitive GAT for hierarchical text classification

Hierarchical text classification aims to assign text to multiple labels in a label set stored in a tree structure. The current algorithms mainly introduce the priori information of the label hierarchy, but the implicit correlation between labels in the hierarchy is rarely applied. At the same time, we also found that the inherent class imbalance of chain labels will also lead to poor classification effects of lower-level labels through a large number of studies. Therefore, a label structure enhanced hierarchy aware global model (LSE-HiAGM) is proposed. Firstly, the common density coefficient of labels is defined to measure the importance of a pair of labels in the hierarchical structure. Secondly, the common density coefficient is used as the weight of the label to update the topological structure features, so that the label can be linked with all labels globally. Finally, the topological structure feature, text features, and label hierarchical features are fused to make full use of all features to improve the embedding quality of low-level labels. In addition, to alleviate the class imbalance problem, a new loss function is used to constrain the model training. The probability of the label being sampled relative to all the labels of the sample is taken as the weight of the loss function. Therefore, a small penalty is imposed on the upper label and a large penalty on the lower label. A large number of experiments on datasets such as RCV1, WOS and NYT show that LSE-HiAGM performs better than the baseline models in hierarchical text classification.

Guided Learning of Human Sensor Models with Low-Level Grounding

Sensor data often lacks intuitive interpretability in its raw form, unlike language or image data. Furthermore, standard end-to-end training leaves little control over local representation learning. We postulate that guided local representation learning could be used to tackle both issues. In this paper we introduce a novel framework for sensor models which uses low-level grounding for guided learning of human sensor models. Our framework is amenable to different model architecture. We demonstrate our method on two different human activity datasets, one containing labels of low-level actions used in performing high-level activities, and one without any low-level labeling. We provide comprehensive analysis of our framework’s performance across many low-level action subsets and demonstrate how it can be easily adapted to data with no low-level labeling. Our results demonstrate that low-level grounding can be used to improve both the interpretability and performance of sensor models.

Accelerator mass spectrometry systems for 14C in low-level labeling studies

Accelerator mass spectrometry systems for 14C in low-level labeling studies

Multilabel Neighborhood Propagation for Region-Based Image Retrieval

Content-based image retrieval (CBIR) has been an active research topic in the last decade. As one of the promising approaches, graph-based semi-supervised learning has attracted many researchers. However, while the related work mainly focused on global visual features, little attention has been paid to region-based image retrieval (RBIR). In this paper, a framework based on multilabel neighborhood propagation is proposed for RBIR, which can be characterized by three key properties: (1) For graph construction, in order to determine the edge weights robustly and automatically, mixture distribution is introduced into the Earth mover's distance (EMD) and a linear programming framework is involved. (2) Multiple low-level labels for each image can be obtained based on a generative model, and the correlations among different labels are explored when the labels are propagated simultaneously on the weighted graph. (3) By introducing multilayer semantic representation (MSR) and support vector machine (SVM) into the long-term learning, more exact weighted graph for label propagation and more meaningful high-level labels to describe the images can be calculated. Experimental results, including comparisons with the state-of-the-art retrieval systems, demonstrate the effectiveness of our proposal.

Osr1 expression demarcates a multi-potent population of intermediate mesoderm that undergoes progressive restriction to an Osr1-dependent nephron progenitor compartment within the mammalian kidney

The mammalian metanephric kidney is derived from the intermediate mesoderm. In this report, we use molecular fate mapping to demonstrate that the majority of cell types within the metanephric kidney arise from an Osr1 + population of metanephric progenitor cells. These include the ureteric epithelium of the collecting duct network, the cap mesenchyme and its nephron epithelia derivatives, the interstitial mesenchyme, vasculature and smooth muscle. Temporal fate mapping shows a progressive restriction of Osr1 + cell fates such that at the onset of active nephrogenesis, Osr1 activity is restricted to the Six2 + cap mesenchyme nephron progenitors. However, low-level labeling of Osr1 + cells suggests that the specification of interstitial mesenchyme and cap mesenchyme progenitors occurs within the Osr1 + population prior to the onset of metanephric development. Furthermore, although Osr1 + progenitors give rise to much of the kidney, Osr1 function is only essential for the development of the nephron progenitor compartment. These studies provide new insights into the cellular origins of metanephric kidney structures and lend support to a model where Osr1 function is limited to establishing the nephron progenitor pool.

Localization of taurine transporters, taurine, and (3)H taurine accumulation in the rat retina, pituitary, and brain.

The nervous system contains an abundance of taurine, a neuroactive sulfonic acid. Antibodies were generated against two cloned high-affinity taurine transporters, referred to in this study as TAUT-1 and TAUT-2. The distribution of such was compared with the distribution of taurine in the rat brain, pituitary, and retina. The cellular pattern of [(3)H] taurine uptake in brain slices, pituitary slices, and retinas was examined by autoradiography. TAUT-2 was predominantly associated with glial cells, including the Bergmann glial cells of the cerebellum and astrocytes in brain areas such as hippocampus. Low-level labeling for TAUT-2 was also observed in some neurones such as CA1 pyramidal cells. TAUT-1 distribution was more limited; in the posterior pituitary TAUT-1 was associated with the pituicytes but was absent from glial cells in the intermediate and anterior lobes. Conversely, in the brain TAUT-1 was associated with cerebellar Purkinje cells and, in the retina, with photoreceptors and bipolar cells. Our data suggest that intracellular taurine levels in glial cells and neurons may be regulated in part by specific high-affinity taurine transporters. The heterogeneous distribution of taurine and its transporters in the brain does not reconcile well with the possibility that taurine acts solely as a ubiquitous osmolyte in nervous tissues.

Expression of N‐Methyl‐D‐Aspartate receptor subunit mRNA in the human brain: Mesencephalic dopaminergic neurons

Evidence is accumulating that glutamate-mediated excitotoxicity plays an important role in neuronal degeneration in Parkinson's disease (PD). In addition, alterations in excitatory amino acid neurotransmission in the basal ganglia contribute to the clinical manifestations of motor dysfunction. However, detailed knowledge of the anatomical distribution and subtype specificity of glutamate receptors in the dopamine neurons of human substantia nigra (SN) has been lacking. In order to test the hypothesis that selective expression of particular N-methyl-D-aspartate receptor (NR) subunit mRNA contributes to the differential vulnerability of specific neuronal populations to excitotoxic injury in PD, we have used a quantitative dual label, in situ hybridization technique with ribonucleotide probes to examine the cellular distribution of NR subunit mRNA in postmortem human mesencephalic dopaminergic neurons from subjects with no known neurological disorder. Analysis of both film autoradiograms and emulsion-dipped sections demonstrated significant labeling of nigral neurons for each NR subunit. Neuronal labeling was most intense for the NR1 and NR2D subunits, with low level labeling for the remaining subunits. In addition, we examined four subregions of the ventral mesencephalon for differential expression of NR subunit mRNA. For all NR subunits, the pars lateralis (PL) exhibited the most intense signal, while neurons of the ventral tier substantia nigra pars compacta (SNpc) failed to demonstrate a preponderance of a particular subunit. These results demonstrate that NRs are expressed to a significant degree in dopaminergic neurons of the SN and that their distribution does not correlate with the characteristic pattern of neuronal degeneration in PD. J. Comp. Neurol. 390:91–101, 1998. © 1998 Wiley-Liss, Inc.

Expression of N-Methyl-D-Aspartate receptor subunit mRNA in the human brain: Mesencephalic dopaminergic neurons

Evidence is accumulating that glutamate-mediated excitotoxicity plays an important role in neuronal degeneration in Parkinson's disease (PD). In addition, alterations in excitatory amino acid neurotransmission in the basal ganglia contribute to the clinical manifestations of motor dysfunction. However, detailed knowledge of the anatomical distribution and subtype specificity of glutamate receptors in the dopamine neurons of human substantia nigra (SN) has been lacking. In order to test the hypothesis that selective expression of particular N-methyl-D-aspartate receptor (NR) subunit mRNA contributes to the differential vulnerability of specific neuronal populations to excitotoxic injury in PD, we have used a quantitative dual label, in situ hybridization technique with ribonucleotide probes to examine the cellular distribution of NR subunit mRNA in postmortem human mesencephalic dopaminergic neurons from subjects with no known neurological disorder. Analysis of both film autoradiograms and emulsion-dipped sections demonstrated significant labeling of nigral neurons for each NR subunit. Neuronal labeling was most intense for the NR1 and NR2D subunits, with low level labeling for the remaining subunits. In addition, we examined four subregions of the ventral mesencephalon for differential expression of NR subunit mRNA. For all NR subunits, the pars lateralis (PL) exhibited the most intense signal, while neurons of the ventral tier substantia nigra pars compacta (SNpc) failed to demonstrate a preponderance of a particular subunit. These results demonstrate that NRs are expressed to a significant degree in dopaminergic neurons of the SN and that their distribution does not correlate with the characteristic pattern of neuronal degeneration in PD.

Synchronized overproduction of AMPA, kainate, and NMDA glutamate receptors during human spinal cord development.

Quantitative receptor autoradiography was used to map the distribution in the developing human spinal cord of the three types of ionotropic glutamate receptors. N-methyl-D-Aspartate (NMDA) receptors were labeled with [3H]glutamate, kainic acid (KA) receptors were labeled with [3H]KA, and alpha-amino-3-hydroxy-5-methyl-4-isoxazole proprionate (AMPA) receptors were labeled with [3H]AMPA. In the adult, labeling of all three receptor subtypes is largely restricted to the substantia gelatinosa (SG) in the dorsal horn, with very low level labeling elsewhere in the spinal gray matter. In marked distinction, in late fetal life, high level ligand binding is seen throughout the spinal gray matter. In early postnatal life, binding sites diminish in all regions, but least so in the SG, until the adult pattern emerges. Thus a coordinated transient high level of ionotropic glutamate receptor expression occurs within the developing spinal cord. Saturation analysis of ligand binding shows that the affinity of [3H]KA and [3H]AMPA binding is not developmentally regulated. In contrast, the affinity of [3H]glutamate binding to the NMDA receptor in the fetal ventral horn is three-fold greater than in the adult ventral horn. Thus, in addition to quantitative changes in glutamate receptor expression, qualitative changes occur in the expression of NMDA receptors during development. The distinct glutamate receptor phenotype of fetal and early postnatal spinal cord cells suggests that alterations in the excitable properties of these cells plays an important role in activity-dependent development and in susceptibility to excitotoxic injury.

DNESYS-an expert system for automatic extraction of drainage networks from digital elevation data

The problems that arose in earlier programs to map drainage systems are analyzed in detail. An expert system called the drainage Network Extraction System (DNESYS) is described. It uses both local operator and global reasoning to extract drainage networks and ridge lines. A stream representation called a parameterized directed graph (PDG) is constructed to model a drainage system. The construction of the model begins with an initial pixel labeling procedure. Then, a network tracing and property measurement procedure converts the two-dimensional low-level labeling information into a symbolic database for high-level processing. The application of Dempster-Shafer evidence theory allows evidence collection and uncertain reasoning to be performed against the DNESYS knowledge-base that contains the drainage system model and the organized expert knowledge. By discarding erroneous information and supplying missing information, DNESYS produces a complete PDG which can be converted into the final drainage system. >

Isolation of the two components of Condylactis toxin

The two toxic components from the sea anemone Condylactis gigantea have been isolated and purified using gel filtration, ion-exchange chromatography, and preparative isoelectric focusing. The molecular weight of both components is estimated by gel filtration to be 4500. The isoelectric points of the two toxins were determined to be 4.8 and 5.8 with LD 50 values of 19 and 58 μg/kg, respectively, when injected into Armadillidium vulgare, a terrestrial crustacean. Low-level labeling with 125I using lactoperoxidase was effected for the purpose of estimating homogeneity.