Divergence Loss Research Articles

A branching process for homology distribution-based inference of polyploidy, speciation and loss

BackgroundThe statistical distribution of the similarity or difference between pairs of paralogous genes, created by whole genome doubling, or between pairs of orthologous genes in two related species is an important source of information about genomic evolution, especially in plants.MethodsWe derive the mixture of distributions of sequence similarity for duplicate gene pairs generated by repeated episodes of whole gene doubling. This involves integrating sequence divergence and gene pair loss through fractionation, using a branching process and a mutational model. We account not only for the timing of these events in terms of local modes, but also the amplitude and variance of the component distributions. This model is then extended to orthologous gene pairs.ResultsWe apply the model and inference procedures to the evolution of the Solanaceae, focusing on the genomes of economically important crops. We assess how consistent or variable fractionation rates are from species to species and over time.

Relative Pairwise Relationship Constrained Non-Negative Matrix Factorisation

Non-negative Matrix Factorisation (NMF) has been extensively used in machine learning and data analytics applications. Most existing variations of NMF only consider how each row/column vector of factorised matrices should be shaped, and ignore the relationship among pairwise rows or columns. In many cases, such pairwise relationship enables better factorisation, for example, image clustering and recommender systems. In this paper, we propose an algorithm named, Relative Pairwise Relationship constrained Non-negative Matrix Factorisation (RPR-NMF), which places constraints over relative pairwise distances amongst features by imposing penalties in a triplet form. Two distance measures, squared Euclidean distance and Symmetric divergence, are used, and exponential and hinge loss penalties are adopted for the two measures respectively. It is well known that the so-called "multiplicative update rules" result in a much faster convergence than gradient descend for matrix factorisation. However, applying such update rules to RPR-NMF and also proving its convergence is not straightforward. Thus, we use reasonable approximations to relax the complexity brought by the penalties, which are practically verified. Experiments on both synthetic datasets and real datasets demonstrate that our algorithms have advantages on gaining close approximation, satisfying a high proportion of expected constraints, and achieving superior performance compared with other algorithms.

Advances and Applications in the Quest for Orthologs.

Gene families evolve by the processes of speciation (creating orthologs), gene duplication (paralogs), and horizontal gene transfer (xenologs), in addition to sequence divergence and gene loss. Orthologs in particular play an essential role in comparative genomics and phylogenomic analyses. With the continued sequencing of organisms across the tree of life, the data are available to reconstruct the unique evolutionary histories of tens of thousands of gene families. Accurate reconstruction of these histories, however, is a challenging computational problem, and the focus of the Quest for Orthologs Consortium. We review the recent advances and outstanding challenges in this field, as revealed at a symposium and meeting held at the University of Southern California in 2017. Key advances have been made both at the level of orthology algorithm development and with respect to coordination across the community of algorithm developers and orthology end-users. Applications spanned a broad range, including gene function prediction, phylostratigraphy, genome evolution, and phylogenomics. The meetings highlighted the increasing use of meta-analyses integrating results from multiple different algorithms, and discussed ongoing challenges in orthology inference as well as the next steps toward improvement and integration of orthology resources.

A simple beam expanded for indoor optical wireless communication with short transmission distance

An optical wireless communication with laser light source is proposed and experimentally evaluated. We have used an arc curvature design at end-face of a piece of commercial multimode fiber (MMF) which acting as an emitting head to expand the beam diameter. The behavior of light field distributions with different emitted head were also measured and discussed. Optical amplification was carried out to compensate for the beam divergence loss in free space. The Bit error rate (BER) and launched power, measured in the distance of 2 m from ceiling to desk, are 400 Mb/s data rate and 10 dBm, respectively. The proposed method for laser expanded could be widely used in optical wireless communication system for short transmission distance.

Pitman closeness domination in predictive density estimation for two-ordered normal means under $$\alpha$$-divergence loss

We consider Pitman closeness domination in predictive density estimation problems when the underlying loss metric is $$\alpha$$ -divergence, $$\{D(\alpha )\}$$ , a loss introduced by Csiszar (Stud Sci Math Hung 2:299–318, 1967). The underlying distributions considered are normal location-scale models, including the distribution of the observables, the distribution of the variable, whose density is to be predicted, and the estimated predictive density which will be taken to be of the plug-in type. The scales may be known or unknown. Chang and Strawderman (J Multivar Anal 128:1–9, 2014) have derived a general expression for the $$\alpha$$ -divergence loss in this setup, and have shown that it is a concave monotone function of quadratic loss, and also a function of the variances (predicand, and plug-in). We demonstrate $$\{D(\alpha )\}$$ Pitman closeness domination of certain plug-in predictive densities over others for the entire class of metrics simultaneously when modified Pitman closeness domination holds in the related problem of estimating the mean. We also establish $$\{D(\alpha )\}$$ Pitman closeness results for certain generalized Bayesian (best invariant) predictive density estimators. Examples of $$\{D(\alpha )\}$$ Pitman closeness domination presented relate to the problem of estimating the predictive density of the variable with the larger mean. We also consider the case of two-ordered normal means with a known covariance matrix.

On the Loss Robustness of Least-Square Estimators

The article revisits univariate and multivariate linear regression models. It is shown that least-square estimators (LSEs) are minimum risk estimators in general class of linear unbiased estimators under some general divergence loss. This amounts to the loss robustness of LSEs.

Every beach an island—deep population divergence and possible loss of genetic diversity in Tylos granulatus, a sandy shore isopod

MEPS Marine Ecology Progress Series Contact the journal Facebook Twitter RSS Mailing List Subscribe to our mailing list via Mailchimp HomeLatest VolumeAbout the JournalEditorsTheme Sections MEPS 614:111-123 (2019) - DOI: https://doi.org/10.3354/meps12882 Every beach an island—deep population divergence and possible loss of genetic diversity in Tylos granulatus, a sandy shore isopod Nozibusiso A. Mbongwa1, Cang Hui2, Andrea Pulfrich3, Sophie von der Heyden1,* 1Department of Botany and Zoology, University of Stellenbosch, Private Bag X1, Matieland 7602, South Africa 2Department of Mathematical Sciences, University of Stellenbosch, and African Institute for Mathematical Sciences, Matieland 7602, South Africa 3Pisces Environmental Services (Pty) Ltd, PO Box 302, McGregor 6708, South Africa *Corresponding author: svdh@sun.ac.za ABSTRACT: Biogeographic and phylogeographic patterns of sandy beach species are poorly understood, although these ecosystems are heavily impacted by anthropogenic pressures and are of elevated conservation concern. To contribute towards filling the knowledge gap on sandy beaches, we made use of phylogeographic approaches to determine levels of genetic structuring and diversity for Tylos granulatus, a large isopod with direct development distributed in South Africa and Namibia. Individuals (n = 214) were sampled from 9 locations encompassing the entire distribution range, and sequence data were generated for mtDNA cytochrome c oxidase subunit I (COI) and 16S. Results revealed high levels of population structuring between populations (ΦST = 0.11-0.96, p < 0.05), 2 deeply divergent lineages of T. granulatus and a new phylogeographic break in southern Africa. Northern populations are genetically more diverse, suggesting more stable evolutionary history compared to those in the south. Importantly, the patterns of divergence suggest unique evolutionary signals over short spatial scales (<80 km), with nearly all T. granulatus populations effectively isolated from each other. This, in combination with increasing anthropogenic disturbance throughout their range, leaves this species extremely vulnerable to local, and potentially regional, loss and extinction of genetic diversity. We suggest that T. granulatus serves as a valuable bioindicator, which, if recognised for protection, will broadly capture biological and evolutionary patterns of other sandy beach macrofauna in the region. Our work adds to a growing field of sandy beach science, and contributes towards a better understanding of the significance, vulnerability and complexity of sandy beach ecosystems for conservation and management aims. KEY WORDS: Anthropogenic pressures · Marine isopod · Sandy beach ecosystems · Population differentiation · Evolutionary lineages · Conservation genetics · Population decline Full text in pdf format PreviousNextCite this article as: Mbongwa NA, Hui C, Pulfrich A, von der Heyden S (2019) Every beach an island—deep population divergence and possible loss of genetic diversity in Tylos granulatus, a sandy shore isopod. Mar Ecol Prog Ser 614:111-123. https://doi.org/10.3354/meps12882 Export citation RSS - Facebook - Tweet - linkedIn Cited by Published in MEPS Vol. 614. Online publication date: April 04, 2019 Print ISSN: 0171-8630; Online ISSN: 1616-1599 Copyright © 2019 Inter-Research.

Multi-representational learning for Offline Signature Verification using Multi-Loss Snapshot Ensemble of CNNs

Offline Signature Verification (OSV) is a challenging pattern recognition task, especially in the presence of skilled forgeries that are not available during training. This study aims to tackle its challenges and meet the substantial need for generalization for OSV by examining different loss functions for Convolutional Neural Network (CNN). We adopt our new approach to OSV by asking two questions: 1. which classification loss provides more generalization for feature learning in OSV?, and 2. How integration of different losses into a unified multi-loss function lead to an improved learning framework?These questions are studied based on analysis of three loss functions, including cross entropy, Cauchy-Schwarz divergence, and hinge loss. According to complementary features of these losses, we combine them into a dynamic multi-loss function and propose a novel ensemble framework for simultaneous use of them in CNN. Our proposed Multi-Loss Snapshot Ensemble (MLSE) consists of several sequential trials. In each trial, a dominant loss function is selected from the multi-loss set, and the remaining losses act as a regularizer. Different trials learn diverse representations for each input based on signature identification task. This multi-representation set is then employed for the verification task. An ensemble of SVMs is trained on these representations, and their decisions are finally combined according to the selection of most generalizable SVM for each user.We conducted two sets of experiments based on two different protocols of OSV, i.e., writer-dependent and writer-independent on three signature datasets: GPDS-Synthetic, MCYT, and UT-SIG. Based on the writer-dependent OSV protocol, On UT-SIG, we achieved 6.17% Equal Error Rate (EER) which showed substantial improvement over the best EER in the literature, 9.61%. Our method surpassed state-of-the-arts by 2.5% on GPDS-Synthetic, achieving 6.13%. Our result on MCYT was also comparable to the best previous results. The second set of experiments examined the robustness of our proposed method to the arrival of new users enrolled in the OSV system based on the writer-independent protocol. The results also confirmed that our proposed system efficiently performed the verification of new users enrolled in the OSV system.

Head pose estimation using deep multitask learning

Head pose estimation (HPE) plays a vital role in the field of human–computer interaction and remains a challenging task due to individual differences. To mitigate this issue, we propose a simple yet effective deep multitask learning framework for joint HPE and face verification (FV), where FV acts as an auxiliary task to boost the performance of HPE. The framework comprises a backbone net, data separate module, and two branches for HPE and FV, respectively. Considering other regions beyond those of the face provides useful information for HPE while FV should focus only on face regions, two kinds of regions, head and face, which share common feature representations in the backbone net, are separated by the data separate module and then fed into the appropriate branches. Kullback–Leibler divergence loss and L2-constrained softmax loss are connected to the end of the HPE branch and FV branch, respectively, to optimize the architecture. The proposed method is validated on three publicly available datasets: Pointing04, CAS-PEAL-R1, and CMU multi-PIE. The experimental results demonstrate that our method surpasses the state of the art, showing up to a 16.38% improvement on the famous benchmark, Pointing04. The best accuracies we report on these datasets are 89.53%, 99.74%, and 99.72%, respectively.

Head pose estimation with soft labels using regularized convolutional neural network

Head pose estimation has many wide applications such as driver monitoring, attention recognition and multi-view facial analysis. Most of the previous works routinely utilize detected face regions to further estimate head pose with hard labels, which limits to explore more discriminative texture information and tends to over-fit. In this paper, we present a novel framework to alleviate this problem, which takes entire images as input and constructs soft labels using a Gaussian distribution function as supervision information, and then introduces a regularized convolutional neural network architecture that is optimized by two types of similarity measure functions: Kullback–Leibler divergence loss and Jeffreys divergence loss. The regularized architecture includes four modules: one backbone net for learning common features, two parallel branches named sub-net1 and sub-net2 for learning complementary features and one feature fusion module, namely, fused net. The architecture is trained in an alternately training fashion, making the learned model more robust and stable. Extensive experiments have been carried out on three public datasets: Pointing04, CAS-PEAL-R1 and CMU Multi-PIE. The results show that our method achieves a significant improvement in performance compared to the state of the art. The best accuracy on the three datasets we achieve are 85.77%, 99.19% and 99.88%, respectively.

Анализ дивергентной потери устойчивости стационарных режимов движения двухзвенного экипажа

Анализ дивергентной потери устойчивости стационарных режимов движения двухзвенного экипажа

Some Variants of Constrained Estimation in Finite Population Sampling

SummaryThis is a review article that unifies several important examples using constrained optimisation techniques. The basic tools are three simple mathematical optimisation results subject to certain constraints. Applications include calibration, benchmarking in small area estimation and imputation. A final illustration is constrained optimisation under a general divergence loss.

Hierarchical Empirical Bayes Estimation of Two Sample Means Under Divergence Loss

We consider the problem of simultaneous estimation of two population means when one suspects that the two means are nearly equal. It is shown that the hierarchical empirical Bayes estimators which shrink the sample means towards the suspected hypothesis dominate the sample mean vectors in simultaneous estimation under the divergence loss function.

Secondary electron emission detectors for neutral beam characterization on C-2W.

Heating, current drive, and partial fueling from neutral beam injection are essential to sustainment of C-2W field-reversed configuration plasmas. C-2W has eight 2.1 MW neutral beams (16.8 MW of total electrical power), capable of providing a beam of 15 keV hydrogen neutrals for 30 ms. To maximize the effectiveness of neutral beam injection, duct losses must be minimized by maintaining beam alignment and optimizing beam current for minimum divergence. Each beam terminates on a vertical and horizontal array of secondary electron emission detectors (nine in the vertical, seven in the horizontal, and sharing one in the middle). The molybdenum detectors are spatially separated to characterize the beam size and alignment. With knowledge of the geometry of the vacuum ducts and horizontal and vertical beam profiles from test stand measurements, the focal length, divergence, and power loss were calculated. Through characterization, the set of neutral beams are optimized to inject up to 12 MW of power into the confinement vessel throughout the plasma discharge.

Optimal Shrinkage of Eigenvalues in the Spiked Covariance Model.

We show that in a common high-dimensional covariance model, the choice of loss function has a profound effect on optimal estimation. In an asymptotic framework based on the Spiked Covariance model and use of orthogonally invariant estimators, we show that optimal estimation of the population covariance matrix boils down to design of an optimal shrinker η that acts elementwise on the sample eigenvalues. Indeed, to each loss function there corresponds a unique admissible eigenvalue shrinker η* dominating all other shrinkers. The shape of the optimal shrinker is determined by the choice of loss function and, crucially, by inconsistency of both eigenvalues and eigenvectors of the sample covariance matrix. Details of these phenomena and closed form formulas for the optimal eigenvalue shrinkers are worked out for a menagerie of 26 loss functions for covariance estimation found in the literature, including the Stein, Entropy, Divergence, Fréchet, Bhattacharya/Matusita, Frobenius Norm, Operator Norm, Nuclear Norm and Condition Number losses.

Investigation of road train divergent stability loss when moving along a program trajectory

The object of research is transport driven multilink wheel systems. In the development of the realized possibility of controlled movement of a semi-trailer truck along a program trajectory, the possibility of constructing a bifurcation set by a velocity parameter is considered. The velocity value is calculated for each discrete value of the calculated real trajectory. The trajectory can be specified in an explicit, implicit, parametric form or by the law of variation of the curvature radius. The study of this parameter is one of the most problematic places for analyzing the stability of the movement of a road train. Changes in this parameter at certain values, called bifurcation, lead to changes in the qualitative structure of the solutions of the system of differential equations and, as a result, divergent stability of the road train. For such an investigation of the phenomenon, the method of continuation by parameter and the first Lyapunov method are applied. During the research, many bifurcation velocity values are obtained. This is due to the fact that the proposed approach has a number of features, in particular, an iteration is performed for all control parameters of the program trajectory, and for each such value, the velocity has been iterated until its bifurcation value is reached. At each iteration, the roots of the characteristic equation are checked for the presence of at least one root with a positive real part, which corresponds to the bifurcation value of the parameter of the velocity of a road train according to Lyapunov. Due to this, it is possible to obtain this set by an exclusively analytical method using computer calculations, without resorting to the use of graphic-analytical methods. Obtaining these bifurcation sets can practically be used both to limit the velocity of a road train and to warn of its excess. Compared with similar known methods, this provides such advantages as a significant acceleration of the construction of this set and, as a result, its use in real time.

Effect of low-frequency oscillation on performance of Hall thrusters

In this paper, a direct connection between the discharge current amplitude and the thruster performance is established by varying solely the capacitance of the filter unit of the Hall thrusters. To be precise, the variation characteristics of ion current, propellant utilization efficiency, and divergence angle of plume at different low-frequency oscillation amplitudes are measured. The findings demonstrate that in the case of the propellant in the discharge channel just meets or falls below the full ionization condition, the increase of low-frequency oscillation amplitude can significantly enhance the ionization degree of the neutral gas in the channel and increase the thrust and anode efficiency of thruster. On the contrary, the increase in the amplitude of low-frequency oscillation will lead to increase the loss of plume divergence, therefore the thrust and anode efficiency of thruster decrease.

Feasibility Analysis of 1-D Carbon Material in Application of Negative Particle Beam Diagnostics

An important feature of the China Fusion Engineering Test Reactor (CFETR) project is the additional heating obtained from the injection of neutral beams based on accelerated negative ions. For the neutral beams based on negative ions, the most important measurements are beam uniformity, beamlet divergence, and stripping losses. According to the CFETR requirement, the maximum allowed beam divergence angle and beam nonuniformity are 6 mrads and ±10%, respectively. As one-dimensional (1-D) carbon tiles have large ratio between perpendicular conductivity and parallel conductivity and high stability, they can be used for beam uniformity and beamlet divergence measurement. This paper investigates the influence on the response of 1-D carbon tile having the thermal characteristics and features of some dedicated diagnostics. Simulations show that it will be possible to verify experimentally whether the beam meets the requirement about the maximum allowed value. This work lays a foundation for design and application of high-precision beam diagnostic targets.

Particle simulation of grid system for krypton ion thrusters

The transport processes of plasmas in grid systems of krypton (Kr) ion thrusters at different acceleration voltages were simulated with a 3D-PIC model, and the result was compared with xenon (Xe) ion thrusters. The variation of the screen grid transparency, the accelerator grid current ratio and the divergence loss were explored. It is found that the screen grid transparency increases with the acceleration voltage and decreases with the beam current, while the accelerator grid current ratio and divergence loss decrease first and then increase with the beam current. This result is the same with Xe ion thrusters. Simulation results also show that Kr ion thrusters have more advantages than Xe ion thrusters, such as higher screen grid transparency, smaller accelerator grid current ratio, larger cut-off current threshold, and better divergence loss characteristic. These advantages mean that Kr ion thrusters have the ability of operating in a wide range of current. Through comprehensive analyses, it can be concluded that using Kr as propellant is very suitable for a multi-mode ion thruster design.

Reflection assisted beam propagation model for obstructed line-of-sight FSO links

A non-line-of-sight (NLOS) infra-red reflection based beam propagation model is proposed as a supplement to conventional terrestrial free space optical (FSO) communication system. This ray propagation model lets tactically positioned optical reflectors to smartly exploit the aggregated advantages of mirror characteristics to bridge the existent communication gap between two FSO nodes due to inclined or obstructed line-of-sight view. Additionally, a numerical framework of the proposed system is presented that analytically explores the optical losses induced by harmonic distortions and the resultant beam divergence at the receiver. The impact of the different reflectors on the traversing beam is then investigated through an experimental FSO test-bed set in an outdoor environment in terms of phase shifts, divergence loss, noise margin and maximum achievable link length. Matlab based simulations, based on the experimental outcomes, envisages that concave reflectors can effectively compensate the turbulence induced signal fading and restrict the beam divergence loss; thereby, improving the maximum achievable NLOS FSO link length.