A Comparative Study on Deep Learning Models for COVID-19 Forecast.

The COVID-19 pandemic has led to a global health crisis with significant morbidity, mortality, and socioeconomic disruptions. Understanding and predicting the dynamics of COVID-19 are crucial for public health interventions, resource allocation, and policy decisions. By developing accurate models, informed public health strategies can be devised, resource allocation can be optimized, and virus transmission can be reduced. Various mathematical and computational models have been developed to estimate transmission dynamics and forecast the pandemic's trajectories. However, the evolving nature of COVID-19 demands innovative approaches to enhance prediction accuracy. The machine learning technique, particularly the deep neural networks (DNNs), offers promising solutions by leveraging diverse data sources to improve prevalence predictions. In this study, three typical DNNs, including the Long Short-Term Memory (LSTM) network, Physics-informed Neural Network (PINN), and Deep Operator Network (DeepONet), are employed to model and forecast COVID-19 spread. The training and testing data used in this work are the global COVID-19 cases in the year of 2021 from the Center for Systems Science and Engineering (CSSE) at Johns Hopkins University. A seven-day moving average as well as the normalization techniques are employed to stabilize the training of deep learning models. We systematically investigate the effect of the number of training data on the predicted accuracy as well as the capability of long-term forecast in each model. Based on the relative L2 errors between the predictions from deep learning models and the reference solutions, the DeepONet, which is capable of learning hidden physics given the training data, outperforms the other two approaches in all test cases, making it a reliable tool for accurate forecasting the dynamics of COVID-19.

Physics-informed neural networks for predicting gas flow dynamics and unknown parameters in diesel engines.

This paper presents a physics-informed neural network (PINN) approach for monitoring the health of diesel engines. The aim is to evaluate the engine dynamics, identify unknown parameters in a "mean value" model, and anticipate maintenance requirements. The PINN model is applied to diesel engines with a variable-geometry turbocharger and exhaust gas recirculation, using measurement data of selected state variables. The results demonstrate the ability of the PINN model to predict simultaneously both unknown parameters and dynamics accurately with both clean and noisy data, and the importance of the self-adaptive weight in the loss function for faster convergence. The input data for these simulations are derived from actual engine running conditions, while the outputs are simulated data, making this a practical case study of PINN's ability to predict real-world dynamical systems. The mean value model of the diesel engine incorporates empirical formulae to represent certain states, but these formulae may not be generalizable to other engines. To address this, the study considers the use of deep neural networks (DNNs) in addition to the PINN model. The DNNs are trained using laboratory test data and are used to model the engine-specific empirical formulae in the mean value model, allowing for a more flexible and adaptive representation of the engine's states. In other words, the mean value model uses both the PINN model and the DNNs to represent the engine's states, with the PINN providing a physics-based understanding of the engine's overall dynamics and the DNNs offering a more engine-specific and adaptive representation of the empirical formulae. By combining these two approaches, the study aims to offer a comprehensive and versatile approach to monitoring the health and performance of diesel engines.

Eef2k is not required for fertility in male mice.

BACKGROUND: Eukaryotic elongation factor-2 kinase (Eef2k) is a protein kinase associated with the calmodulin-induced signaling pathway and an atypical alpha-kinase family member. Eef2k-mediated phosphorylation of eukaryotic translation elongation factor 2 (Eef2) can inhibit the functionality of this protein, altering protein translation. Prior work suggests Eef2k to be overexpressed in breast, pancreatic, brain, and lung cancers wherein it may control key processes associated with apoptosis, autophagy, and cell cycle progression. The functional importance of Eef2k in the testes of male mice, however, has yet to be clarified. METHODS: A CRISPR/Cas9 approach was used to generate male Eef2k-knockout mice, which were evaluated for phenotypic changes in epididymal or testicular tissues through histological and immunofluorescent staining assays. In addition, TUNEL staining was conducted to assess the apoptotic death of cells in the testis. Fertility, sperm counts, and sperm motility were further assessed. RESULTS: Male Eef2k-knockout mice were successfully generated, and exhibited normal fertility and development. No apparent differences were observed with respect to spermatogenesis, sperm counts, or germ cell apoptosis when comparing male Eef2k -/- and Eef2k +/+ mice. CONCLUSIONS: Male Eef2k-knockout mice remained fertile and were free of any evident developmental or spermatogenic abnormalities, suggesting Eef2k to be dispensable in the context of male fertility.

Discrete effects on boundary conditions of the lattice Boltzmann method for fluid flows with curved no-slip walls.

The lattice Boltzmann method (LBM) has been formulated as a powerful numerical tool to simulate incompressible fluid flows. However, it is still a critical issue for the LBM to overcome the discrete effects on boundary conditions successfully for curved no-slip walls. In this paper, we focus on the discrete effects of curved boundary conditions within the framework of the multiple-relaxation-time (MRT) model. We analyze in detail a single-node curved boundary condition [Zhao et al., Multiscale Model. Simul. 17, 854 (2019)10.1137/18M1201986] for predicting the Poiseuille flow and derive the numerical slip at the boundary dependent on a free parameter as well as the distance ratio and the relaxation times. An approach by virtue of the free parameter is then proposed to eliminate the slip velocity while with uniform relaxation parameters. The theoretical analysis also indicates that for previous curved boundary schemes only with the distance ratio and the halfway bounce-back (HBB) boundary scheme, the numerical slip cannot be removed with uniform relaxation times virtually. We further carried out some simulations to validate our theoretical derivations, and the numerical results for the case of straight and curved boundaries confirm our theoretical analysis. Finally, for fluid flows with curved boundary geometries, resorting to more degrees of freedom from the boundary scheme may have more potential to eliminate the discrete effect at the boundary with uniform relaxation times.

Boundary scheme for linear heterogeneous surface reactions in the lattice Boltzmann method.

A boundary scheme is proposed to treat the linear heterogeneous surface reaction (i.e., general Robin boundary condition) for the lattice Boltzmann method (LBM) in this study. The basic idea of the present scheme is to compute the scalar variable gradient in the boundary condition with the moment of the nonequilibrium distribution functions. The unknown distribution functions on the boundary can then be obtained on the basis of the given boundary condition. For a straight wall, where the lattice nodes are located on the boundaries, both the scalar variable and its gradient can be expressed in terms of the distribution functions at the boundary node, and the scheme is purely local. The scheme is also extended to problems with a curved wall, in which a linear extrapolation is employed to realize the exact boundary position. A common feature of the two schemes lies in the easy treatment of the heterogenous reactions in comparison with existing methods. A number of simulations are performed to test the accuracy of the schemes. The results show that for flat walls the scheme can achieve second-order accuracy, while for curved walls the order of the accuracy is between 1.0 and 1.5.

Multiple-relaxation-time lattice Boltzmann model for incompressible miscible flow with large viscosity ratio and high Péclet number.

A lattice Boltzmann model with a multiple-relaxation-time (MRT) collision operator is proposed for incompressible miscible flow with a large viscosity ratio as well as a high Péclet number in this paper. The equilibria in the present model are motivated by the lattice kinetic scheme previously developed by Inamuro et al. [Philos. Trans. R. Soc. London, Ser. A 360, 477 (2002)]. The fluid viscosity and diffusion coefficient depend on both the corresponding relaxation times and additional adjustable parameters in this model. As a result, the corresponding relaxation times can be adjusted in proper ranges to enhance the performance of the model. Numerical validations of the Poiseuille flow and a diffusion-reaction problem demonstrate that the proposed model has second-order accuracy in space. Thereafter, the model is used to simulate flow through a porous medium, and the results show that the proposed model has the advantage to obtain a viscosity-independent permeability, which makes it a robust method for simulating flow in porous media. Finally, a set of simulations are conducted on the viscous miscible displacement between two parallel plates. The results reveal that the present model can be used to simulate, to a high level of accuracy, flows with large viscosity ratios and/or high Péclet numbers. Moreover, the present model is shown to provide superior stability in the limit of high kinematic viscosity. In summary, the numerical results indicate that the present lattice Boltzmann model is an ideal numerical tool for simulating flow with a large viscosity ratio and/or a high Péclet number.

[Impact of transurethral resection of the prostate on erectile function: a report of 64 cases].

OBJECTIVE: To analyze the impact of transurethral resection of the prostate (TURP) on erectile function and the factors influencing postoperative erectile function. METHODS: Altogether 64 male patients aged 53 -75 (mean 66.5) years underwent TURP for prostatic hyperplasia. Before and 3 months after surgery, we observed the nocturnal penile tumescence of the patients and analyzed their scores on the 5-item version of the International Index of Erectile Function (IIEF-5) and the Self-Rating Anxiety Scale (SAS). RESULTS: Intraoperative prostatic capsule perforation and postoperative stress were significantly related to postoperative erectile dysfunction (P < 0.05). The mean score of IIEF-5 was significantly decreased (P < 0.01) while that of SAS remarkably increased (P < 0.01) after TURP as compared with those before surgery. The frequency of nocturnal penile tumescence was reduced at 3 months after surgery, but with no statistically significant difference. CONCLUSION: Intraoperative prostatic capsule perforation and postoperative stress obviously affect postoperative erectile function.