Coronavirus disease (COVID-19) is rapidly spreading worldwide. Recent studies show that radiological images contain accurate data for detecting the coronavirus. This paper proposes a pre-trained convolutional neural network (VGG16) with Capsule Neural Networks (CapsNet) to detect COVID-19 with unbalanced data sets. The CapsNet is proposed due to its ability to define features such as perspective, orientation, and size. Synthetic Minority Over-sampling Technique (SMOTE) was employed to ensure that new samples were generated close to the sample center, avoiding the production of outliers or changes in data distribution. As the results may change by changing capsule network parameters (Capsule dimensionality and routing number), the Gaussian optimization method has been used to optimize these parameters. Four experiments have been done, (1) CapsNet with the unbalanced data sets, (2) CapsNet with balanced data sets based on class weight, (3) CapsNet with balanced data sets based on SMOTE, and (4) CapsNet hyperparameters optimization with balanced data sets based on SMOTE. The performance has improved and achieved an accuracy rate of 96.58% and an F1- score of 97.08%, a competitive optimized model compared to other related models.
- Keywords
- COVID-19, Capsule Neural Networks, Convolution neural networks, Coronavirus, Gaussian optimization method, VGG16,
- Publication type
- Journal Article MeSH
A consistent method for optimizing Gaussian primitives for Rydberg and multiply excited helium states is designed. A novel series for the "exponentially tempered Gaussians" is introduced, which is markedly more efficient than the commonly used series of even tempered Gaussians. The optimization is made computationally feasible due to an approximate calculation of excited states using the effective one-electron Hamiltonian that is defined as Fockian from which the redundant Coulomb and exchange terms are dropped. Finally, ExTG5G and ExTG7F Gaussian basis sets are proposed. They enable calculations of the helium spectrum all the way from the ground state up to the (5, 4)(5) (1)S(e) and (6, 5)(7) (1)S(e) doubly excited resonances, respectively, mostly in the spectroscopic accuracy of 1 cm(-1).
- Publication type
- Journal Article MeSH
Numerical simulation of continuous variable quantum state preparation is a necessary tool for optimization of existing quantum information processing protocols. A powerful instrument for such simulation is the numerical computation in the Fock state representation. It unavoidably uses an approximation of the infinite-dimensional Fock space by finite complex vector spaces implementable with classical digital computers. In this approximation we analyze the accuracy of several currently available methods for computation of the truncated coherent displacement operator. To overcome their limitations we propose an alternative with improved accuracy based on the standard matrix exponential. We then employ the method in analysis of non-Gaussian state preparation scheme based on coherent displacement of a two mode squeezed vacuum with subsequent photon counting measurement. We compare different detection mechanisms, including avalanche photodiodes, their cascades, and photon number resolving detectors in the context of engineering non-linearly squeezed cubic states and construction of qubit-like superpositions between vacuum and single photon states.
- Publication type
- Journal Article MeSH
For accurate ab initio description of Rydberg excited states, this study suggests generating appropriate diffuse basis functions by cheap variational optimization of virtual orbitals of the corresponding ion core. By following this approach, dozens of converged correlated lithium Rydberg states, namely, all the states up to 24 2S, 25 2P, 14 2D, 16 2F and 16 2G, not yet achieved via other ab initio approaches, could be obtained at the EOM-CCSD level of theory with compact and mostly state-selective contracted Gaussian basis sets. Despite its small size and Gaussian character, the optimized basis leads to highly accurate excitation energies that differ merely in the order of meV from the reference state-of-the-art explicitly correlated Gaussian method and even surpass Full-CI results on the Slater basis by an order of magnitude.
- Publication type
- Journal Article MeSH
Tensor network states and specifically matrix-product states have proven to be a powerful tool for simulating ground states of strongly correlated spin models. Recently, they have also been applied to interacting fermionic problems, specifically in the context of quantum chemistry. A new freedom arising in such nonlocal fermionic systems is the choice of orbitals, it being far from clear what choice of fermionic orbitals to make. In this Letter, we propose a way to overcome this challenge. We suggest a method intertwining the optimization over matrix product states with suitable fermionic Gaussian mode transformations. The described algorithm generalizes basis changes in the spirit of the Hartree-Fock method to matrix-product states, and provides a black box tool for basis optimization in tensor network methods.
- Publication type
- Journal Article MeSH
We study the optimality conditions of information transfer in systems with memory in the low signal-to-noise ratio regime of vanishing input amplitude. We find that the optimal mutual information is represented by a maximum variance of the signal time course, with correlation structure determined by the Fisher information matrix. We provide illustration of the method on a simple biologically inspired model of electrosensory neuron. Our general results apply also to the study of information transfer in single neurons subject to weak stimulation, with implications to the problem of coding efficiency in biological systems.
- MeSH
- Biophysics methods MeSH
- Models, Biological MeSH
- Electrochemistry methods MeSH
- Humans MeSH
- Neurons metabolism pathology physiology MeSH
- Normal Distribution MeSH
- Models, Statistical MeSH
- Stochastic Processes MeSH
- Animals MeSH
- Check Tag
- Humans MeSH
- Animals MeSH
- Publication type
- Journal Article MeSH
- Research Support, Non-U.S. Gov't MeSH
INTRODUCTION: The aim of this study was to determine the optimal image matrix and half-width of the Gaussian filter after iterative reconstruction of the PET image with point-spread function (PSF) and time-of-flight (TOF) correction, based on measuring the recovery coefficient (RC) curves. The measured RC curves were compared to those from an older system which does not use PSF and TOF corrections. MATERIALS AND METHODS: The measurements were carried out on a NEMA IEC Body Phantom. We measured the RC curves based on SUVmax and SUVA50 in source spheres with different diameters. The change in noise level for different reconstruction parameter settings and the relation between RC curves and the administered activity were also evaluated. RESULTS: With an increasing size of image matrix and reduction in the half-width of the post-reconstruction Gaussian filter, there was a significant increase in image noise and overestimation of the SUV. The local increase in SUV, observed for certain filtrations and objects with a diameter below 13mm, was caused by PSF correction. The decrease in administered activity, while maintaining the same conditions of acquisition and reconstruction, also led to overestimation of readings of the SUV and additionally to deterioration in reproducibility. CONCLUSION: This study proposes a suitable size for the image matrix and filtering for displaying PET and SUV measurements. The benefits were demonstrated as improved image parameters for the newer instrument, these even being found using relatively strong filtration of the reconstructed images.
- MeSH
- Phantoms, Imaging * MeSH
- Humans MeSH
- Image Processing, Computer-Assisted * MeSH
- Positron-Emission Tomography * MeSH
- Reproducibility of Results MeSH
- Check Tag
- Humans MeSH
- Publication type
- Journal Article MeSH
The impact of the classification method and features selection for the speech emotion recognition accuracy is discussed in this paper. Selecting the correct parameters in combination with the classifier is an important part of reducing the complexity of system computing. This step is necessary especially for systems that will be deployed in real-time applications. The reason for the development and improvement of speech emotion recognition systems is wide usability in nowadays automatic voice controlled systems. Berlin database of emotional recordings was used in this experiment. Classification accuracy of artificial neural networks, k-nearest neighbours, and Gaussian mixture model is measured considering the selection of prosodic, spectral, and voice quality features. The purpose was to find an optimal combination of methods and group of features for stress detection in human speech. The research contribution lies in the design of the speech emotion recognition system due to its accuracy and efficiency.
- MeSH
- Algorithms * MeSH
- Databases, Factual MeSH
- Emotions physiology MeSH
- Voice Quality MeSH
- Humans MeSH
- Neural Networks, Computer MeSH
- Signal Processing, Computer-Assisted instrumentation MeSH
- Speech physiology MeSH
- ROC Curve MeSH
- Pattern Recognition, Automated * MeSH
- Pattern Recognition, Physiological physiology MeSH
- Check Tag
- Humans MeSH
- Publication type
- Journal Article MeSH
In this study, we consider a method for investigating the stochastic response of a nonlinear dynamical system affected by a random seismic process. We present the solution of the probability density of a single/multiple-degree of freedom (SDOF/MDOF) system with several statically stable equilibrium states and with possible jumps of the snap-through type. The system is a Hamiltonian system with weak damping excited by a system of non-stationary Gaussian white noise. The solution based on the Gibbs principle of the maximum entropy of probability could potentially be implemented in various branches of engineering. The search for the extreme of the Gibbs entropy functional is formulated as a constrained optimization problem. The secondary constraints follow from the Fokker-Planck equation (FPE) for the system considered or from the system of ordinary differential equations for the stochastic moments of the response derived from the relevant FPE. In terms of the application type, this strategy is most suitable for SDOF/MDOF systems containing polynomial type nonlinearities. Thus, the solution links up with the customary formulation of the finite elements discretization for strongly nonlinear continuous systems.
The structure of dithienobicyclo[4.4.1]undeca-3,8-diene-11-one ethylene glycol ketal (database code RESVAN) was determined using the wave function theory (WFT) as well as density functional theory (DFT) methods combined with various Gaussian AO basis sets. The apparently most accurate procedure, employing the CCSD(T)/complete basis set (CBS), provides an S-S distance and an angle between the two thiophene rings which differ considerably from experimental values. The best agreement with the experimental data among all WFT methods was surprisingly obtained at the MP3/aug-cc-pVDZ and MP3/CBS(B) levels (the correction term to CBS was obtained by the aug-cc-pVDZ basis set). The very good results obtained by the CCSD(T)/6-31G* method are clearly a consequence of fortunate error compensation. MP2 calculations, even with a small basis set, overestimate the attraction between the thiophene rings, and the worst agreement with experimental data was found in full MP2/QZVP method optimizations (i.e., a strong distortion of the thiophene rings was observed). The SCS(MI)-MP2 and SCS-MP2 methods exhibit improvement over the MP2 procedure. All standard DFT approaches fail to predict reasonable S-S distances. The lack of intramolecular London dispersion energy results in too great distance between the thiophene rings. Much better agreement with experiment was obtained if advanced DFT methods, covering dispersion effects, were used. The best results were obtained at the TPSS-D/TZVP, M06-L/TZVP and B2PLYP-D/def2-TZVP levels. When a larger basis (LP in the case of TPSS functional) or more advanced versions of the new Truhlar functionals (M06-2X) was used, the agreement with experiment deteriorated. The accurate description of this molecule is highly functional/basis dependent and this dependence is hardly predictable. To estimate effects of neighboring molecules in the experimental crystal structure, an optimization in the electric field of the 26 closest RESVAN molecules was performed, which, however, leads to only moderate (<0.05 A) changes of the S-S distance.
- Publication type
- Journal Article MeSH
