Union2 Dataset Research Articles

An Improved YOLOv5-Based Underwater Object-Detection Framework

To date, general-purpose object-detection methods have achieved a great deal. However, challenges such as degraded image quality, complex backgrounds, and the detection of marine organisms at different scales arise when identifying underwater organisms. To solve such problems and further improve the accuracy of relevant models, this study proposes a marine biological object-detection architecture based on an improved YOLOv5 framework. First, the backbone framework of Real-Time Models for object Detection (RTMDet) is introduced. The core module, Cross-Stage Partial Layer (CSPLayer), includes a large convolution kernel, which allows the detection network to precisely capture contextual information more comprehensively. Furthermore, a common convolution layer is added to the stem layer, to extract more valuable information from the images efficiently. Then, the BoT3 module with the multi-head self-attention (MHSA) mechanism is added into the neck module of YOLOv5, such that the detection network has a better effect in scenes with dense targets and the detection accuracy is further improved. The introduction of the BoT3 module represents a key innovation of this paper. Finally, union dataset augmentation (UDA) is performed on the training set using the Minimal Color Loss and Locally Adaptive Contrast Enhancement (MLLE) image augmentation method, and the result is used as the input to the improved YOLOv5 framework. Experiments on the underwater datasets URPC2019 and URPC2020 show that the proposed framework not only alleviates the interference of underwater image degradation, but also makes the mAP@0.5 reach 79.8% and 79.4% and improves the mAP@0.5 by 3.8% and 1.1%, respectively, when compared with the original YOLOv8 on URPC2019 and URPC2020, demonstrating that the proposed framework presents superior performance for the high-precision detection of marine organisms.

Remote assessments of human pressure on biodiversity may miss important human threats

AbstractMonitoring human pressure on biodiversity within protected areas (PAs) remains a challenging task due to the vast number of PAs worldwide (>250,000) and the fact that we lack basic information for most of them. Consequently, researchers are increasingly relying on remote sensing techniques to overcome this challenge. However, remote assessments of human pressure on biodiversity may miss crucial threats, such as hunting and invasive species, which are often best documented through in situ field surveys. Here, we use a unique European Union dataset documenting human threats within 8210 PAs through field surveys to assess the relationship between those threats and three commonly used indices of human pressure that are largely derived using remote‐sensing methods—the Human Footprint and Human Modification indices and the LandScan Global Population Database. We find that the indices are not always related to the threats recorded within PAs through field surveys. The indices seem to best capture threats associated with urbanization, agriculture, and pollution but not necessarily others. Although remote assessments of human pressure on biodiversity are crucial for conservation, researchers and practitioners must be aware of their limitations and must complement such assessments with information collected in the field whenever possible.

Privacy-Preserving Distributed Clustering for Electrical Load Profiling

Electrical load profiling supports retailers in identifying consumer categories for customizing tariff design. However, each retailer only has access to the data of the customers it serves. Centralized joint clustering on retailers' union load dataset either enables the identification of more types of users that allows to design more customized retail plans, or informs whether each retailer already has a sufficiently broad customer base. However, the centralized clustering requires access to the confidential data of retailers. This may cause privacy issues among retailers, because retailers can not or do not want to share their confidential information with others. To tackle this issue, we propose a privacy-preserving distributed clustering framework by developing a privacy-preserving accelerated average consensus (PP-AAC) algorithm. Using the proposed framework, we modify several commonly used clustering methods, including k-means, fuzzy C-means, and Gaussian mixture model, to provide privacy-preserving distributed clustering methods. In this way, the clustering on retailers' union dataset can be achieved only by local calculations and information sharing between neighboring retailers without sacrificing privacy. The correctness, privacy-preserving property, time-saving feature, and robustness to random communication failures of the proposed methods are verified using a real-world Irish residential dataset.

Are Type Ia supernova powerful tool to detect anisotropic expansion of the Universe?

In this paper, we present a detailed analysis of the dark energy dipole using Union2, Pantheon and GRB dataset in Chameleon and Teleparallel dark energy models, in comparison with Lambda CDM. Both models are extensively studied in recent years and our result shows that with Union2 and Pantheon data, the preferred direction of the anisotropy in both models are very close to each other as well as with those obtained in some studies for Lambda CDM. However, when the models fitted with a combination of Union 2 and GRB, the statistical analysis slightly favors the Chameleon cosmology over Teleparallel gravity, with the maximum anisotropic direction of (l = 330^{+30}_{-28}, b = -15^{+23}_{-25}) in galactic coordinate system, comparable with alpha -dipole result in Keck-VLT data and LCDM.

Assessing eHealth skills across Europeans

The Internet has made health information more assessable to the general public. However, some serious concerns remain about the quality and reliability of that information, and the ability of the population to accurately interpret these data. An important aspect interpreting healthcare information are eHealth skills. To date, the available evidence is usually based on national analyses of composite measures of eHealth skills. To contribute to the debate our paper analyses particular eHealth skills based on data from a pan-European Union data set. In particular, with microdata from a representative sample, five skills are analysed, viz. searching, locating, understanding, evaluating and using online health information. We use a discrete choice model to identify statistical associations between respondent socio-demographic characteristics and skills. Our analysis reveals a complex pattern of eHealth skills is present across socio-demographic groups, with only self-reported health status and Internet experience influential for all skills. This finding suggests that targeted training actions are necessary to improve eHealth kills, with the seniors and the less educated the groups most in need of training.

A Comparative Look into Public IXP Datasets

Internet eXchange Points (IXPs) are core components of the Internet infrastructure where Internet Service Providers (ISPs) meet and exchange traffic. During the last few years, the number and size of IXPs have increased rapidly, driving the flattening and shortening of Internet paths. However, understanding the present status of the IXP ecosystem and its potential role in shaping the future Internet requires rigorous data about IXPs, their presence, status, participants, etc. In this work, we do the first cross-comparison of three well-known publicly available IXP databases, namely of PeeringDB, Euro-IX, and PCH. A key challenge we address is linking IXP identifiers across databases maintained by different organizations. We find different AS-centric versus IXP-centric views provided by the databases as a result of their data collection approaches. In addition, we highlight differences and similarities w.r.t. IXP participants, geographical coverage, and co-location facilities. As a side-product of our linkage heuristics, we make publicly available the union of the three databases, which includes 40.2% more IXPs and 66.3% more IXP participants than the commonly-used PeeringDB. We also publish our analysis code to foster reproducibility of our experiments and shed preliminary insights into the accuracy of the union dataset.

Cosmological constraints on the variable modified Chaplygin gas model by using MCMC approach

We investigate the cosmological constraints on the variable modified Chaplygin gas (VMCG) model from the latest observational data: Union2 dataset of Type Ia supernovae (SNIa), the observational Hubble data (OHD), the baryon acoustic oscillations (BAO) and the cosmic microwave background (CMB) data. By using the Markov chain Monte Carlo (MCMC) method, we obtain the mean values of parameters in the flat model: [Formula: see text], [Formula: see text], [Formula: see text], [Formula: see text], [Formula: see text], [Formula: see text] and [Formula: see text]. Furthermore, we investigate the thermodynamical properties of VMCG model at apparent horizon, event horizon and particle horizon respectively.

Anisotropic expansion and SNIa: An open issue

We review the appropriateness of using SNIa observations to detect potential signatures of anisotropic expansion in the Universe. We focus on Union2 and SNLS3 SNIa datasets and use the hemispherical comparison method to detect possible anisotropic features. Unlike some previous works where non-diagonal elements of the covariance matrix were neglected, we use the full covariance matrix of the SNIa data, thus obtaining more realistic and not underestimated errors. As a matter of fact, the significance of previously claimed detections of a preferred direction in the Union2 dataset completely disappears once we include the effects of using the full covariance matrix. Moreover, we also find that such a preferred direction is aligned with the orthogonal direction of the SDSS observational plane and this suggests a clear indication that the SDSS subsample of the Union2 dataset introduces a significant bias, making the detected preferred direction unphysical. We thus find that current SNIa surveys are inappropriate to test anisotropic features due to their highly non-homogeneous angular distribution in the sky. In addition, after removal of the highest inhomogeneous sub-samples, the number of SNIa is too low. Finally, we take advantage of the particular distribution of SNLS SNIa sub-sample in the SNLS3 data set, in which the observations were taken along four different directions. We fit each direction independently and find consistent results at the 1σ level. Although the likelihoods peak at relatively different values of Ωm, the low number of data along each direction gives rise to large errors so that the likelihoods are sufficiently broad as to overlap within 1σ.

Comparison between hemisphere comparison method and dipole-fitting method in tracing the anisotropic expansion of the Universe use the Union2 data set

Type-Ia supernovae (SNe Ia) are often used as the standard candles to probe the anisotropic expansion of the Universe. In this paper, we make a comprehensive comparison between the hemisphere comparison (HC) method and dipole-fitting (DF) method in searching for the cosmological preferred direction using the Union2 dataset, a compilation of 557 well-calibrated SNe Ia. We find that the directions of the faintest SNe Ia derived from these two methods are approximately opposite. Monte Carlo simulations show that the results of the HC method strongly depend on the distribution of the data points in the sky. The coincidence that the HC method and DF method give two completely opposite directions may be due to the extremely nonuniform distribution of the Union2 dataset.

Constraining the anisotropy of the universe from supernovae and gamma-ray bursts

Recently, an anisotropic cosmological model was proposed. An arbitrary one-form, which picks out a privileged axis in the universe, was added to the Friedmann–Robertson–Walker (FRW) line element. The distance-redshift relation was modified such that it is direction-dependent. In this paper, we use the Union2 dataset and 59 high-redshift gamma-ray bursts (GRBs) to give constraints on the anisotropy of the universe. The results show that the magnitude of anisotropy is about D = -0.044±0.018, and the privileged axis points toward the direction (l0, b0) = (306.1°±18.7°, -18.2°±11.2°) in the galactic coordinate system. The anisotropy is small and the isotropic cosmological model is an excellent approximation.

Modified holographic Ricci dark energy coupled to interacting dark matter and a non-interacting baryonic component

We examine a Friedmann-Robertson-Walker universe filled with interacting dark matter, modified holographic Ricci dark energy (MHRDE), and a decoupled baryonic component.The estimations of the cosmic parameters with Hubble data lead to an age of the universe of $13.17 {\rm Gyr}$ and show that the MHRDE is free from the cosmic-age problem at low redshift ($0 \leq z \leq 2$) in contrast to holographic Ricci dark energy (HRDE) case. We constrain the parameters with the Union2 data set and contrast with the Hubble data. We also study the behavior of dark energy at early times by taking into account the severe bounds found at recombination era and/or at big bang nucleosynthesis. The inclusion of a non interacting baryonic matter forces that the amount of dark energy at $z_{t} \sim {\cal O}(1)$ changes abruptly implying that $\Omega_x (z\simeq 1100) =0.03$, so the bounds reported by the forecast of Planck and CMBPol experiments are more favored for the MHRDE model than in the case of HRDE cutoff. For the former model, we also obtain that at high redshift the fraction of dark energy varies from 0.006 to 0.002, then the amount of $\Omega_x $ at the big bang nucleosynthesis era does not disturb the observed Helium abundance in the universe provided that the bound $\Omega_{x}(z\simeq 10^{10})<0.21$ is hold.

An accelerated universe from Brans-Dicke theory in the Einstein frame

The current acceleration of the universe is not consistent with the theory of general relativity (GR). The substitutes of GR have been studied widely. One of them is the Brans-Dicke (BD) theory. Thereinto, most studies of BD theory are in the frame of Jordan representation. In this paper we investigate the accelerating cosmology from BD theory in the other frame, i.e. in the Einstein frame, where it indicates the interaction between the matter and the BD scalar field. We apply the Markov Chain Monte Carlo method to study the cosmic constraints on BD theory according to the current experiment data from the kinematical observations: the Union2 dataset of type supernovae Ia (SNIa) and the observational Hubble data (OHD), and the dynamical observations: the baryon acoustic oscillation (BAO), the cosmic microwave background (CMB) data, and the cluster X-ray gas mass fraction. It is shown that an expanded universe from deceleration to acceleration is given in this theory. And the constraint results on BD model parameters are β = −0.0067(±0.2268) −0.2372−0.4407 +0.2395+0.4168 , γ = 0.0237(±0.0342) −0.0329−0.0647 +0.0326+0.0715 , e = −0.0063(±0.0073) −0.0073−0.0130 +0.0072+0.0158 ; and the constraints values of some correlative cosmological parameters are the dimensionless matter density Ω 0m = 0.2853(±0.0132) −0.0133−0.0247 +0.0132+0.0262 , the cosmic age = 13.5699(±0.2398) −0.2374−0.4568 +0.2371+0.4806 Gyr, and the current Hubble parameter H 0 = 70.2623(±1.2695) −1.2675−2.5150 +1.2749+2.4532 .

CONSTRAINTS ON LEMAÎTRE-TOLMAN-BONDI MODELS FROM OBSERVATIONAL HUBBLE PARAMETER DATA

We use the observational Hubble parameter data (OHD), both the latest observational dataset (Stern et al. 2010, referred to as SJVKS) and the simulated datasets, to constrain Lema\^{\i}tre-Tolman-Bondi (LTB) void models. The necessity of the consistency check on OHD itself in the LTB cosmology is stressed. Three voids are chosen as test models and are constrained using the Union2 dataset of SN Ia as well as OHD. Despite their different parametrization, the results from our test models show some indicating similarities, e.g., the best-fit voids obtained from OHD are all considerably broader than those from SN Ia. Due to the small size of the SJVKS dataset, the constraints are not conclusive. The constraining power of the future OHD observations are therefore investigated, through a Figure of Merit (FoM) analysis based on the Monte Carlo simulated data. We found that, in the case that the future OHD become more consistent with SN Ia, the results from the test models are almost unanimous: 1) as many as 32 OHD data points at the SJVKS-like uncertainty level are needed to give a higher FoM than the Union2 dataset of SN Ia; 2) precise observation helps reduce this required number; 3) increasing the survey depth does not always increase the FoM. On the other hand, if the future OHD and the Union2 dataset keep favor different voids, in a similar manner as they do at present, the 1{\sigma} confidence regions obtained from the two probes should finally separate. We test this conjecture and found that, the minimum observational requirement (the size of the dataset, the uncertainty level and the survey depth) for this inconsistency to emerge depends strongly on the void model.

Cosmological constraints on the generalized holographic dark energy

We use the Markov Chain Monte Carlo method to investigate global constraints on the generalized holographic (GH) dark energy with flat and non-flat universe from the current observed data: the Union2 dataset of type supernovae Ia (SNIa), high-redshift Gamma-Ray Bursts (GRBs), the observational Hubble data (OHD), the cluster X-ray gas mass fraction, the baryon acoustic oscillation (BAO), and the cosmic microwave background (CMB) data. The most stringent constraints on the GH model parameter are obtained. In addition, it is found that the equation of state for this generalized holographic dark energy can cross over the phantom boundary w de =−1.

Does accelerating universe indicate Brans-Dicke theory?

The evolution of universe in Brans-Dicke (BD) theory is discussed in this paper. Considering a parameterized scenario for BD scalar field $\phi=\phi_{0}a^{\alpha}$ which plays the role of gravitational "constant" $G$, we apply the Markov Chain Monte Carlo method to investigate a global constraints on BD theory with a self-interacting potential according to the current observational data: Union2 dataset of type supernovae Ia (SNIa), high-redshift Gamma-Ray Bursts (GRBs) data, observational Hubble data (OHD), the cluster X-ray gas mass fraction, the baryon acoustic oscillation (BAO), and the cosmic microwave background (CMB) data. It is shown that an expanded universe from deceleration to acceleration is given in this theory, and the constraint results of dimensionless matter density $\Omega_{0m}$ and parameter $\alpha$ are, $\Omega_{0m}=0.286^{+0.037+0.050}_{-0.039-0.047}$ and $\alpha=0.0046^{+0.0149+0.0171}_{-0.0171-0.0206}$ which is consistent with the result of current experiment exploration, $\mid\alpha\mid \leq 0.132124$. In addition, we use the geometrical diagnostic method, jerk parameter $j$, to distinguish the BD theory and cosmological constant model in Einstein's theory of general relativity.

Fitting Cosmological Data to the Function q(z) from GR Theory: Modified Chaplygin Gas

In the Friedmann cosmology the deceleration of the expansion $q$ plays a fundamental role. We derive the deceleration as a function of redshift $q(z)$ in two scenarios: $\Lambda$CDM model and modified Chaplygin gas ($MCG$) model. The function for the $MCG$ model is then fitted to the cosmological data in order to obtain the cosmological parameters that minimize $\chi^2$. We use the Fisher matrix to construct the covariance matrix of our parameters and reconstruct the q(z) function. We use Supernovae Ia, WMAP5 and BAO measurements to obtain the observational constraints. We determined the present acceleration as $q_0=-0.60 \pm 0.12$ for the $MCG$ model using the Constitution dataset of SNeIa and BAO, and $q_0=-0.63 \pm 0.17$ for the Union dataset and BAO. The transition redshift from deceleration to acceleration was found to be around $0.6$ for both datasets. We have also determined the dark energy parameter for the $MCG$ model: $\Omega_{X0}=0.834 \pm 0.028$ for the Constitution dataset and $\Omega_{X0}=0.854 \pm 0.036$ using the Union dataset.

Measuring the cosmological bulk flow using the peculiar velocities of supernovae

We study large-scale coherent motion in our universe using the existing Type IA supernovae data. If the recently observed bulk flow is real, then some imprint must be left on supernovae motion. We perform a Bayesian Monte Carlo Markov Chain analysis in various redshift bins and find a sharp contrast between the z < 0.05 and z > 0.05 data. The z < 0.05 data are consistent with the bulk flow in the direction (l,b) = (290+39−31°,20+32−32°) with a magnitude of vbulk = 188+119−103km/s at 68% confidence. The significance of detection (compared to the null hypothesis) is 95%. In contrast, z > 0.05 data (which contains 425 of the 557 supernovae in the Union2 data set) show no evidence for the bulk flow. While the direction of the bulk flow agrees very well with previous studies, the magnitude is significantly smaller. For example, the Kashlinsky, et al.'s original bulk flow result of vbulk > 600km/s is inconsistent with our analysis at greater than 99.7% confidence level. Furthermore, our best-fit bulk flow velocity is consistent with the expectation for the ΛCDM model, which lies inside the 68% confidence limit.

Exploring the latest Union2 type Ia supernovae dataset by using model-independent parametrization methods

We explore the cosmological consequences of the recently released Union2 sample of 557 Type Ia supernovae (SNIa). Combining this latest SNIa dataset with the Cosmic microwave background (CMB) anisotropy data from the Wilkinson Microwave Anisotropy Probe 7 year (WMAP7) observations and the baryon acoustic oscillation (BAO) results from the Sloan Digital Sky Survey (SDSS) Data Release 7 (DR7), we measure the dark energy density function $f(z)\equiv \rho_{de}(z)/\rho_{de}(0)$ as a free function of redshift. Two model-independent parametrization methods (the binned parametrization and the polynomial interpolation parametrization) are used in this paper. By using the $\chi^2$ statistic and the Bayesian information criterion, we find that the current observational data are still too limited to distinguish which parametrization method is better, and a simple model has advantage in fitting observational data than a complicated model. Moreover, it is found that all these parametrizations demonstrate that the Union2 dataset is still consistent with a cosmological constant at 1$\sigma$ confidence level. Therefore, the Union2 dataset is different from the Constitution SNIa dataset, which more favors a dynamical dark energy.

Non-minimally coupled [formula omitted] cosmology

We investigate the consequences of non-minimal gravitational coupling to matter and study how it differs from the case of minimal coupling by choosing certain simple forms for the nature of coupling. The values of the parameters are specified at z=0 (present epoch) and the equations are evolved backwards to calculate the evolution of cosmological parameters. We find that the Hubble parameter evolves more slowly in non-minimal coupling case as compared to the minimal coupling case. In both the cases, the universe accelerates around present time, and enters the decelerating regime in the past. Using the latest Union2 dataset for supernova Type Ia observations as well as the data for baryon acoustic oscillation (BAO) from SDSS observations, we constraint the parameters of Linder exponential model in the two different approaches. We find that there is an upper bound on model parameter in minimal coupling. But for non-minimal coupling case, there is range of allowed values for the model parameter.

Searching for a cosmological preferred axis: Union2 data analysis and comparison with other probes

We review, compare and extend recent studies searching for evidence for a preferred cosmological axis. We start from the Union2 SnIa dataset and use the hemisphere comparison method to search for a preferred axis in the data. We find that the hemisphere of maximum accelerating expansion rate is in the direction (l,b) = (309°+23°−3°,18°+11°−10°) (Ω0m = 0.19) while the hemisphere of minimum acceleration is in the opposite direction (l,b) = (129°+23°−3°,−18°+10°−11°) (Ω0m = 0.30). The level of anisotropy is described by the normalized difference of the best fit values of Ω0m between the two hemispheres in the context of ΛCDM fits. We find a maximum anisotropy level in the Union2 data of ΔΩ0m max / Ω0m = 0.43±0.06. Such a level does not necessarily correspond to statistically significant anisotropy because it is reproduced by about 30% of simulated isotropic data mimicking the best fit Union2 dataset. However, when combined with the axes directions of other cosmological observations (bulk velocity flow axis, three axes of CMB low multipole moments and quasar optical polarization alignment axis), the statistical evidence for a cosmological anisotropy increases dramatically. We estimate the probability that the above independent six axes directions would be so close in the sky to be less than 1%. Thus either the relative coincidence of these six axes is a very large statistical fluctuation or there is an underlying physical or systematic reason that leads to their correlation.