In this paper, we propose a new automatic hyperparameter selection approach for determining the optimal network configuration (network structure and hyperparameters) for deep neural networks using ...particle swarm optimization (PSO) in combination with a steepest gradient descent algorithm. In the proposed approach, network configurations were coded as a set of real-number m-dimensional vectors as the individuals of the PSO algorithm in the search procedure. During the search procedure, the PSO algorithm is employed to search for optimal network configurations via the particles moving in a finite search space, and the steepest gradient descent algorithm is used to train the DNN classifier with a few training epochs (to find a local optimal solution) during the population evaluation of PSO. After the optimization scheme, the steepest gradient descent algorithm is performed with more epochs and the final solutions (pbest and gbest) of the PSO algorithm to train a final ensemble model and individual DNN classifiers, respectively. The local search ability of the steepest gradient descent algorithm and the global search capabilities of the PSO algorithm are exploited to determine an optimal solution that is close to the global optimum. We constructed several experiments on hand-written characters and biological activity prediction datasets to show that the DNN classifiers trained by the network configurations expressed by the final solutions of the PSO algorithm, employed to construct an ensemble model and individual classifier, outperform the random approach in terms of the generalization performance. Therefore, the proposed approach can be regarded an alternative tool for automatic network structure and parameter selection for deep neural networks.
Developing a robust catalyst for the oxygen evolution reaction is the major challenge in the field of renewable energy. The difficulty comes from not only the low intrinsic activity, but also the ...structural uncertainty of catalysts under the operating conditions. Therefore, finding the relationship between structural evolution and the OER activity is urgently required. At present, first‐principles simulations have become a powerful tool to understand the mechanism of the OER at the atomic level. In this review, TiO2, MnOx, and CoS2 are used as examples to demonstrate how first‐principles calculations can predict the morphology of nanoparticles, explore the pathway of electrochemically induced phase transition, and resolve the structure of a heterojunction. With these new theoretical techniques, the structure–activity relationship of the OER for a complex catalytic system can be determined without experimental inputs. Such a bottom‐up strategy holds great promise to reveal the active site and mechanism of a complex catalytic system from first‐principles calculations.
Predicting change: First‐principles simulations are a valuable tool to reveal structural evolutions in reactive environments. Herein, TiO2, MnOx, and CoS2 are used as examples to show that first‐principles calculations can predict the morphology of nanoparticles, explore the pathway of phase transition in electrochemical conditions, and resolve the structure of interfaces.
Heart disease is a malignant threat to human health. Electrocardiogram (ECG) tests are used to help diagnose heart disease by recording the heart's activity. However, automated medical-aided ...diagnosis with computers usually requires a large volume of labeled clinical data without patients' privacy to train the model, which is an empirical problem that still needs to be solved. To address this problem, we propose a generative adversarial network (GAN), which is composed of a bidirectional long short-term memory(LSTM) and convolutional neural network(CNN), referred as BiLSTM-CNN,to generate synthetic ECG data that agree with existing clinical data so that the features of patients with heart disease can be retained. The model includes a generator and a discriminator, where the generator employs the two layers of the BiLSTM networks and the discriminator is based on convolutional neural networks. The 48 ECG records from individuals of the MIT-BIH database were used to train the model. We compared the performance of our model with two other generative models, the recurrent neural network autoencoder(RNN-AE) and the recurrent neural network variational autoencoder (RNN-VAE). The results showed that the loss function of our model converged to zero the fastest. We also evaluated the loss of the discriminator of GANs with different combinations of generator and discriminator. The results indicated that BiLSTM-CNN GAN could generate ECG data with high morphological similarity to real ECG recordings.
Frustrated quantum magnets not only provide exotic ground states and unusual magnetic structures, but also support unconventional excitations in many cases. Using a physically relevant spin model for ...a breathing pyrochlore lattice, we discuss the presence of topological linear band crossings of magnons in antiferromagnets. These are the analogues of Weyl fermions in electronic systems, which we dub Weyl magnons. The bulk Weyl magnon implies the presence of chiral magnon surface states forming arcs at finite energy. We argue that such antiferromagnets present a unique example, in which Weyl points can be manipulated in situ in the laboratory by applied fields. We discuss their appearance specifically in the breathing pyrochlore lattice, and give some general discussion of conditions to find Weyl magnons, and how they may be probed experimentally. Our work may inspire a re-examination of the magnetic excitations in many magnetically ordered systems.
COVID-19 was declared a pandemic on March 11 by WHO, due to its great threat to global public health. The coronavirus main protease (M
, also called 3CLpro) is essential for processing and maturation ...of the viral polyprotein, therefore recognized as an attractive drug target. Here we show that a clinically approved anti-HCV drug, Boceprevir, and a pre-clinical inhibitor against feline infectious peritonitis (corona) virus (FIPV), GC376, both efficaciously inhibit SARS-CoV-2 in Vero cells by targeting M
. Moreover, combined application of GC376 with Remdesivir, a nucleotide analogue that inhibits viral RNA dependent RNA polymerase (RdRp), results in sterilizing additive effect. Further structural analysis reveals binding of both inhibitors to the catalytically active side of SARS-CoV-2 protease M
as main mechanism of inhibition. Our findings may provide critical information for the optimization and design of more potent inhibitors against the emerging SARS-CoV-2 virus.
PDZ domains are highly abundant protein-protein interaction modules and are often found in multidomain scaffold proteins. PDZ-domain-containing scaffold proteins regulate multiple biological ...processes, including trafficking and clustering receptors and ion channels at defined membrane regions, organizing and targeting signalling complexes at specific cellular compartments, interfacing cytoskeletal structures with membranes, and maintaining various cellular structures. PDZ domains, each with ~90-amino-acid residues folding into a highly similar structure, are best known to bind to short C-terminal tail peptides of their target proteins. A series of recent studies have revealed that, in addition to the canonical target-binding mode, many PDZ-target interactions involve amino acid residues beyond the regular PDZ domain fold, which we refer to as extensions. Such extension sequences often form an integral structural and functional unit with the attached PDZ domain, which is defined as a PDZ supramodule. Correspondingly, PDZ-domain-binding sequences from target proteins are frequently found to require extension sequences beyond canonical short C-terminal tail peptides. Formation of PDZ supramodules not only affords necessary binding specificities and affinities demanded by physiological functions of PDZ domain targets, but also provides regulatory switches to be built in the PDZ-target interactions. At the 20th anniversary of the discovery of PDZ domain proteins, we try to summarize structural features and target-binding properties of such PDZ supramodules emerging from studies in recent years.
To reduce undesired iron leaching in Fenton reaction and to realize reusability of catalyst, chitosan-coated activated carbon cloth support loaded with nano zero-valent iron (ACC–CH–nZVI) was applied ...as a heterogeneous Fenton catalyst to treat melanoidin wastewater. Chitosan coating on ACC by chemical crosslinking results in 6% chitosan on ACC subsequently loading 3.5% iron. At optimum conditions, ACC–CH–nZVI leads to 88.4% and 76.2% of color and chemical oxygen demand (COD) removal, respectively, upon treating synthetic melanoidin wastewater of 8000 mg/l COD. The corresponding weight ratio of consumed H2O2 to COD is 1.02, far below the stoichiometric ratio 2.125, indicating the economic value of this catalyst. Reusability of ACC–CH–nZVI is demonstrated for five cycles of treatment with minimal iron leaching (<2%). The high removal efficiency and very low levels of iron leaching suggests that ACC–CH–nZVI is a highly efficient and cost-effective catalyst for Fenton-like oxidation of non-biodegradable organic wastes in water.
Display omitted
•Activated carbon cloth (ACC) and chitosan (CH) supported nano zero-valent iron (nZVI).•ACC–CH–nZVI catalyst leads to high color and COD removal from melanoidin wastewater.•The highest removal efficiency was obtained at pH 3 with low consumption of H2O2.•ACC–CH–nZVI catalyst exhibits excellent reusability.•Very low (<2%) iron leaching for ACC–CH–nZVI catalyst in five cycles of treatment.
Abstract
Iron fluorescence emission lines from X-ray binaries and active galactic nuclei are important diagnostic tools for studying the physical processes near the event horizon of both the ...stellar-mass black holes in X-ray binaries and the supermassive black holes in active galactic nuclei. In this work, we investigate the line profile of the relativistic broad iron lines from the cool accretion disk of a black hole due to the asymmetric illumination of a moving corona, which moves away from the disk with a relativistic velocity. Both the off-axis location and the radial velocity of the moving corona are considered. Our results clearly show that the illumination and the line profile are dependent on the position and velocity of the corona, since the disk region below the corona receives more flux, which is the most important factor affecting the line profiles. As expected, if the corona is close to the receding part of the rotating disk, the red peak is enhanced, while the blue peak is weakened in the broad line profile, and the central energy of the emission line is low. Conversely, if the corona is close to the approaching part of the disk, the blue peak is strong and the central energy of the emission line is high, even higher than the intrinsic energy of the emission line. Due to the beaming effect of the moving corona, the corona with high velocity illuminates the outer region of the disk, which leads to the red peak disappearing and there being only one blue peak in the profile of the emission line.
Phase transformation of electrode materials widely occurs in electrocatalytic reactions. Metal oxides are promising electrocatalysts for the oxygen evolution reaction (OER); their phase ...transformation is a key step for the multi-electron OER, and requires extra overpotential. However, little attention has been paid to accelerating and enhancing the phase transformation. Here, we report for the first time that single-atom Pt incorporated into the bulk crystalline phase of porous NiO nanocubes (0.5 wt% Pt/NiO) can greatly promote the active phase (NiOOH) evolution. The Pt doping was achieved by a scalable nanocasting approach using SiO
as the hard template. In comparison with Pt/NiO samples with PtO
nanoparticles segregated at the NiO surface (1 wt% Pt), as well as atomistic Pt atoms solely bound at the surface by atomic layer deposition, the bulk Pt doping shows the strongest power in facilitating active phase transformation, which leads to improved OER activity with reduced overpotential and Tafel slope. Experiential data revealed that the charge-transfer from Pt to Ni through O leads to a local weaker Ni-O bond. First principles calculations confirmed that rather than acting as an active site for the OER, monatomic Pt effectively increases the phase transformation rate by reducing the migration barrier of nearby Ni atoms. Our discoveries reveal the relationships of the heteroatom doped structure and phase transformation behavior during the electrochemical process and offer a new route for designing high-performance electrocatalysts.