Deep Ordinal Classification in Forest Areas Using Light Detection and Ranging Point Clouds.

Recent advances in Deep Learning and aerial Light Detection And Ranging (LiDAR) have offered the possibility of refining the classification and segmentation of 3D point clouds to contribute to the monitoring of complex environments. In this context, the present study focuses on developing an ordinal classification model in forest areas where LiDAR point clouds can be classified into four distinct ordinal classes: ground, low vegetation, medium vegetation, and high vegetation. To do so, an effective soft labeling technique based on a novel proposed generalized exponential function (CE-GE) is applied to the PointNet network architecture. Statistical analyses based on Kolmogorov-Smirnov and Student's t-test reveal that the CE-GE method achieves the best results for all the evaluation metrics compared to other methodologies. Regarding the confusion matrices of the best alternative conceived and the standard categorical cross-entropy method, the smoothed ordinal classification obtains a more consistent classification compared to the nominal approach. Thus, the proposed methodology significantly improves the point-by-point classification of PointNet, reducing the errors in distinguishing between the middle classes (low vegetation and medium vegetation).

Learning Ordinal-Hierarchical Constraints for Deep Learning Classifiers.

Real-world classification problems may disclose different hierarchical levels where the categories are displayed in an ordinal structure. However, no specific deep learning (DL) models simultaneously learn hierarchical and ordinal constraints while improving generalization performance. To fill this gap, we propose the introduction of two novel ordinal-hierarchical DL methodologies, namely, the hierarchical cumulative link model (HCLM) and hierarchical-ordinal binary decomposition (HOBD), which are able to model the ordinal structure within different hierarchical levels of the labels. In particular, we decompose the hierarchical-ordinal problem into local and global graph paths that may encode an ordinal constraint for each hierarchical level. Thus, we frame this problem as simultaneously minimizing global and local losses. Furthermore, the ordinal constraints are set by two approaches ordinal binary decomposition (OBD) and cumulative link model (CLM) within each global and local function. The effectiveness of the proposed approach is measured on four real-use case datasets concerning industrial, biomedical, computer vision, and financial domains. The extracted results demonstrate a statistically significant improvement to state-of-the-art nominal, ordinal, and hierarchical approaches.

Proteomic Approach during the Induction of Somatic Embryogenesis in Coffea canephora.

Plant growth regulators (PGR) are essential for somatic embryogenesis (SE) in different species, and Coffea canephora is no exception. In our study model, previously, we have been able to elucidate the participation of various genes involved in SE by using different strategies; however, until now, we have not used a proteomic approach. This research seeks to contribute to understanding the primary cellular pathways involved in developing SE in C. canephora. The process of our model consists of two stages: (1) preconditioning in MS medium with auxin (NAA) and cytokinin (KIN), and (2) induction in Yasuda liquid medium added with cytokinin (BA). Therefore, in this study, we analyzed different days of the SE induction process using shotgun label-free proteomics. An amount of 1630 proteins was found among different sampling days of the process, of which the majority were accumulated during the induction stage. We found that some of the most enriched pathways during this process were the biosynthesis of amino acids and secondary metabolites. Eighteen proteins were found related to auxin homeostasis and two to cytokinin metabolism, such as ABC, BIG, ILR, LOG, and ARR. Ten proteins and transcription factors related to SE were also identified, like SERK1, SKP1, nuclear transcription factor Y, MADS-box, and calreticulin, and 19 related to other processes of plant development, among which the 14-3-3 and PP2A proteins stand out. This is the first report on the proteomic approach to elucidate the mechanisms that operate during the induction of SE in C. canephora. So, our findings provide the groundwork for future, more in-depth research. Data are available via ProteomeXchange Consortium via the PRIDE partner repository with the dataset identifier PXD047172.

Activation Functions for Convolutional Neural Networks: Proposals and Experimental Study.

Activation functions lie at the core of every neural network model from shallow to deep convolutional neural networks. Their properties and characteristics shape the output range of each layer and, thus, their capabilities. Modern approaches rely mostly on a single function choice for the whole network, usually ReLU or other similar alternatives. In this work, we propose two new activation functions and analyze their properties and compare them with 17 different function proposals from recent literature on six distinct problems with different characteristics. The objective is to shed some light on their comparative performance. The results show that the proposed functions achieved better performance than the most commonly used ones.

Infectious episodes during pregnancy, at particular mucosal sites, increase specific IgA1 or IgA2 subtype levels in human colostrum.

BACKGROUND: Colostrum is the primary source of maternal immunoglobulin A (IgA) for the newborn. IgA participates in protection and regulation mechanisms of the immune response at the neonate's mucosa. Several studies have evaluated infectious diseases and vaccine protocols effects during pregnancy on maternal milk IgA levels, with the aim to understand lactation protecting effect on newborn. However, most of their results demonstrated that there were no differences in the total IgA levels. In humans, IgA has two subclasses (IgA1 and IgA2), they have an anatomical distribution among mucosal compartments, their levels vary after antigen stimulation and are also seen to describe differential affinities in colostrum. Although there are differences between IgA subclasses in several compartments, these studies have excluded specific colostrum IgA1 and IgA2 determination. METHODS: We analyzed data from 900 women in Mexico City. With Pearson correlation, we compared the number of infectious episodes during their pregnancy that was associated with mucosal compartments (skin, respiratory and gastrointestinal tracts) and colostrum IgA subclasses. RESULTS: We show a correlation between increased colostrum IgA1 levels and the number of infectious episodes at respiratory tract and the skin. In contrast, infections at the gastrointestinal tract correlated with increased IgA2 amounts. CONCLUSIONS: Infections present during pregnancy at certain mucosal site increase specific IgA subclasses levels in human colostrum. These results will help in understanding infections and immunizations effects on maternal IgA at the mammary gland, and their impact on the development and protection of the newborn.

Localization and transport of indole-3-acetic acid during somatic embryogenesis in Coffea canephora.

Auxin and polar auxin transport have been implicated in controlling zygotic embryo development, but less is known about their role in the development of somatic embryos. The aim of this study was to determine if indole-3-acetic acid (IAA) and the PIN1 transporter participate in the induction of somatic embryogenesis (SE) and the development of somatic embryos. The results show that IAA levels gradually increase during pre-treatment and accumulate in the chloroplast. During pre-treatment and the globular stage of SE in C. canephora, auxin is distributed uniformly in all of the cells of the somatic embryo. During the subsequent stages of development, auxins are mobilized to the cells that will form the cotyledons and the root meristem. The location of the PIN transporters shifts from the plasmalemma of the protoderm cells during the globular stage to the plasmalemma of the cells that will give rise to the cotyledons and the vascular tissue in the late stages of somatic embryogenesis. The incubation of the explants in the presence of 2,3,5-triiodobenzoic acid (TIBA) produced aberrant somatic embryos, suggesting that PIN1 mediates the transport of IAA.