K. Yi

In the 21 st century, the rapid development of human society has made people's demand for green energy more and more urgent.The high-energy-density hydrogen energy obtained by fully splitting water is not only environmentally friendly, but also is expected to solve the problems caused by the intermittent nature of new energy.However, the slow kinetics and large overpotential of the oxygen evolution reaction (OER) limit its application.The introduction of Te element is expected to bring new breakthroughs.With the least electronegativity among the chalcogens, the Te element has many special properties, such as multivalent states, strong covalentity, and high electrical conductivity, which make it a promising candidate in electrocatalytic OER.In this review, we introduce the peculiarities of Te element, summarize Te doping and the extraordinary performance of its compounds in OER, with emphasis on the scientific mechanism behind Te element promoting the OER kinetic process.Finally, challenges and development prospects of the applications of Te element in OER are presented.

Nanofluids have the potential to become the alternatives of conventional transformer oil for their exquisite electrical and thermal properties. Three kinds of nanoparticles with distinct conductivities, namely, nonconductive nanoparticle Al 2 O 3 , conductive nanoparticle Fe 3 O 4 , and semiconductive nanoparticle TiO 2 , with different concentrations from 5% to 40% w/v were selected and suspended into transformer oil to develop nanofluids. The lightening impulse breakdown strengths of the oil samples with and without nanoparticles were measured according to IEC standard methods. The positive impulse breakdown strength indicated that breakdown strength is first increased up to the maximum value at certain concentration and then starts decreasing. The results of negative impulse breakdown manifested that the breakdown voltages of nanofluids with different concentrations were less than the breakdown voltage of pure transformer oil. Different effect mechanisms of dielectric and conductive nanoparticles were also used to describe the difference among three prepared nanofluids.

The limited availability of annotated data often hinders real-world applications of machine learning. To efficiently learn from small quantities of multimodal data, we leverage the linguistic knowledge from a large pre-trained language model (PLM) and quickly adapt it to new domains of image captioning. To effectively utilize a pretrained model, it is critical to balance the visual input and prior linguistic knowledge from pretraining. We propose VisualGPT, which employs a novel self-resurrecting encoder-decoder attention mechanism to quickly adapt the PLM with a small amount of in-domain image-text data. The proposed self-resurrecting activation unit produces sparse activations that prevent accidental overwriting of linguistic knowledge. When trained on 0.1%, 0.5% and 1% of the respective training sets, VisualGPT surpasses the best baseline by up to 10.0% CIDEr on MS COCO [43] and 17.9% CIDEr on Conceptual Captions [63]. Furthermore, VisualGPT achieves the state-of-the-art result on IU X-ray [15], a medical report generation dataset. Our code is available at https://github.com/Vision-CAIR/VisualGPT.

Recent experiments have shown that some nanoparticles can influence the breakdown strength of transformer oil under lightning impulse voltage. To reveal the working mechanism, this paper presents an experimental study on the effect of TiO <sub xmlns:mml="http://www.w3.org/1998/Math/MathML" xmlns:xlink="http://www.w3.org/1999/xlink">2</sub> nanoparticles on the impulse breakdown strength and prebreakdown streamer propagation process in transformer oil-based nanofluid under both positive and negative lightning impulse voltage. The test results verify that the modification of nanoparticles on breakdown strength of transformer oil has a distinct polar effect: positive breakdown voltage of nanofluid is increased by up to 30.8%, whereas the negative one is decreased by 6.8%. Streamer shape, propagation length and velocity in both pure oil and nanofluid were investigated using the shadowgraph technique. It is revealed that the propagation characteristics of positive and negative streamers in nanofluid are markedly affected by the addition of TiO <sub xmlns:mml="http://www.w3.org/1998/Math/MathML" xmlns:xlink="http://www.w3.org/1999/xlink">2</sub> nanoparticles. The positive streamers in nanofluid form a bush-like structure with thicker and denser branches, developing much slower than tree-like streamers in pure oil. While negative streamers in nanofluid have a tree-like shape with much longer branches, propagating faster than the original bush-like streamer in pure oil. These differences in streamer propagation characteristics and breakdown strength in pure oil and nanofluid are closely related to the change of space charge distribution caused by shallow trap in nanofluid. More negative charges are formed through capturing fast electrons into slow electrons in shallow traps induced by the presence of TiO <sub xmlns:mml="http://www.w3.org/1998/Math/MathML" xmlns:xlink="http://www.w3.org/1999/xlink">2</sub> nanoparticles, which change the local electric field in front of the streamer tip. Thus, streamer propagation process in nanofluid is dramatically modified, leading to the change in breakdown strength.

Novel hole-transporting materials with high work function (WF) and transmittance are crucial for high-efficiency bifacial CdTe solar cells. Herein, the CuZnSe alloy films prepared by combinatorial sputtering exhibit tunable optical band gap (2.25–2.57 eV), WF (5.05–5.45 eV) and band structure with variable Zn content. A large WF of 5.45 eV for Cu0.46Zn0.74Se and small valence band offset of 0.13 eV at Cu0.46Zn0.74Se/CdTe interface have been determined by ultraviolet photoelectron spectroscopy (UPS) and X-ray photoelectron spectroscopy (XPS), respectively, which can promote the hole transport from CdTe absorber to back electrode. A maximum power conversion efficiency (PCE) of 15.3% was achieved for CdTe solar cells when using Cu0.46Zn0.74Se/Au as back contact. The device characterizations indicate that Cu0.46Zn0.74Se has the excellent effects on reducing back‐contact barrier and suppressing carrier recombination. Moreover, Cu0.46Zn0.74Se/ITO back contact yielded high back-illuminated efficiencies of 4.2% and 5.3% for the 3.2 μm and 1.7 μm CdTe devices, respectively, due to increased transmittance and effective carrier collection. This work provides a promising strategy to obtain optically transparent Cu-containing selenides with high work function by alloying, which have attractive application prospect for bifacial CdTe solar cells.

Rice-frog cultivation is a traditional farming system in China and has been reintroduced as an agricultural practice in China in recent years. The microbial community in paddy rhizospheric soils has attracted much attention because many microorganisms participate in functional processes in soils. In this study, Illumina MiSeq sequencing-based techniques were used to investigate soil microbial communities and functional gene patterns across samples obtained by conventional rice cultivation (CR) and rice-frog cultivation (RF). The results showed that RF significantly affected the microbial community composition and richness, which indicated that the rhizospheric microorganisms responded to the introduction of tiger frogs into the paddy fields. Operational taxonomic units (OTUs) from Sandaracinaceae, Anaerolineaceae, Candidatus Nitrososphaera, Candidatus Nitrosotalea, Candidatus and Nitrosoarchaeum and some unclassified OTUs from Euryarchaeota and Agaricomycetes were significantly enriched by RF. The abiotic parameters soil organic carbon (SOC), nitrate nitrogen (NO3--N) and available phosphorus (AP) changed under RF treatment and played essential roles in establishing the soil bacterial, archaeal and fungal compositions. Correlations between environmental factors and microbial communities were described using network analysis. SOC was strongly correlated with Anaerolineaceae, Methanosaeta and Scutellinia. NO3--N showed strong positive correlations with Opitutus, Geobacter, and Methanosaeta. NH4+-N was strongly positively associated with Sideroxydans, and TN was strongly positively correlated with Candidatus Nitrotoga. Compared to conventional CR, RF greatly enriched specific microbial taxa. These taxa may be involved in the decomposition of complex organic matter and the transformation of soil nutrients, thus promoting plant growth by improving nutrient cycling. The unique patterns of microbial taxonomic and functional composition in soil profiles suggested functional redundancy in these paddy soils. RF could significantly affect the bacterial, archaeal and fungal communities though changing SOC and AP levels.

Abstract Computation or deep learning-based functional protein generation methods address the urgent demand for novel biocatalysts, allowing for precise tailoring of functionalities to meet specific requirements. This emergence leads to the creation of highly efficient and specialized proteins with wide-ranging applications in scientific, technological, and biomedical domains. This study establishes a conditional protein diffusion model, namely CPDiffusion, to deliver diverse protein sequences with desired functions. While the model is free from extensive training data and the sampling process involves little guidance on the type of generated amino acids, CPDiffusion effectively secures essential highly conserved residues that are crucial for protein functionalities. We employed CPDiffusion and generated 27 artificially designed Argonaute proteins, programmable endonucleases applied for easy-to-implement and high-throughput screenings in gene editing and molecular diagnostics, that mutated approximately 200 − 400 amino acids with 40% sequence identities to those from nature. Experimental tests demonstrate the solubility of all 27 artificially-designed proteins (AP), with 24 of them displaying DNA cleavage activity. Remarkably, 74% of active APs exhibited superior activity compared to the template protein, and the most effective one showcased a remarkable nearly nine-fold enhancement of enzymatic activity. Moreover, 37% of APs exhibited enhanced thermostability. These findings emphasize CPDiffusion’s remarkable capability to generate long-sequence proteins in a single step while retaining or enhancing intricate functionality. This approach facilitates the design of intricate enzymes featuring multi-domain molecular structures through in silico generation and throughput, all accomplished without the need for supervision from labeled data.

Inverse protein folding is challenging due to its inherent one-to-many mapping characteristic, where numerous possible amino acid sequences can fold into a single, identical protein backbone. This task involves not only identifying viable sequences but also representing the sheer diversity of potential solutions. However, existing discriminative models, such as transformer-based auto-regressive models, struggle to encapsulate the diverse range of plausible solutions. In contrast, diffusion probabilistic models, as an emerging genre of generative approaches, offer the potential to generate a diverse set of sequence candidates for determined protein backbones. We propose a novel graph denoising diffusion model for inverse protein folding, where a given protein backbone guides the diffusion process on the corresponding amino acid residue types. The model infers the joint distribution of amino acids conditioned on the nodes' physiochemical properties and local environment. Moreover, we utilize amino acid replacement matrices for the diffusion forward process, encoding the biologically-meaningful prior knowledge of amino acids from their spatial and sequential neighbors as well as themselves, which reduces the sampling space of the generative process. Our model achieves state-of-the-art performance over a set of popular baseline methods in sequence recovery and exhibits great potential in generating diverse protein sequences for a determined protein backbone structure.

exotic multiquark states are reported using the data collected in pp collisions at √ = 13 TeV.

The Divergent Association Task (DAT), published in July 2021, is a psychological test designed to measure an individual's divergent thinking. The test requires participants to name ten nouns that exhibit maximum dissimilarity from each other. The semantic distance between these nouns is then calculated to indicate the person's level of divergent thinking. In this study, we explored the applicability of the DAT for elementary school students in Chinese contexts, given that it was not initially designed for this specific population and was available only in English. We recruited a total of 348 students who were asked to complete three creativity tasks: the DAT, the Alternative Uses Task (AUT), and the Bridge-the-Associative-Gap Task (BAG). We examined the associations between DAT and the scores of the AUT and BAG tests. Moreover, we tested the accuracy of the DAT using varying numbers of nouns and different natural language processing models to calculate the semantic distance between nouns. Our findings supported the suitability of using the DAT to measure divergent thinking in elementary school students within Chinese contexts. We also found that using only eight nouns, instead of ten, could achieve a relatively high accuracy in measuring divergent thinking based on the DAT method. The language model of Word2Vec performed better than the BERT (Bidirectional Encoder Representations from Transformers) and GPT (Generative Pre-trained Transformer) models when calculating semantic distances between nouns. This study has methodological and practical implications.

Federated Learning (FL) has garnered increasing attention due to its unique characteristic of allowing heterogeneous clients to process their private data locally and interact with a central server, while being respectful of privacy. A critical bottleneck in FL is the communication cost. A pivotal strategy to mitigate this burden is \emph{Local Training}, which involves running multiple local stochastic gradient descent iterations between communication phases. Our work is inspired by the innovative \emph{Scaffnew} algorithm, which has considerably advanced the reduction of communication complexity in FL. We introduce FedComLoc (Federated Compressed and Local Training), integrating practical and effective compression into \emph{Scaffnew} to further enhance communication efficiency. Extensive experiments, using the popular TopK compressor and quantization, demonstrate its prowess in substantially reducing communication overheads in heterogeneous settings.

Tabular data plays a crucial role in various domains but often suffers from missing values, thereby curtailing its potential utility. Traditional imputation techniques frequently yield suboptimal results and impose substantial computational burdens, leading to inaccuracies in subsequent modeling tasks. To address these challenges, we propose DiffImpute, a novel Denoising Diffusion Probabilistic Model (DDPM). Specifically, DiffImpute is trained on complete tabular datasets, ensuring that it can produce credible imputations for missing entries without undermining the authenticity of the existing data. Innovatively, it can be applied to various settings of Missing Completely At Random (MCAR) and Missing At Random (MAR). To effectively handle the tabular features in DDPM, we tailor four tabular denoising networks, spanning MLP, ResNet, Transformer, and U-Net. We also propose Harmonization to enhance coherence between observed and imputed data by infusing the data back and denoising them multiple times during the sampling stage. To enable efficient inference while maintaining imputation performance, we propose a refined non-Markovian sampling process that works along with Harmonization. Empirical evaluations on seven diverse datasets underscore the prowess of DiffImpute. Specifically, when paired with the Transformer as the denoising network, it consistently outperforms its competitors, boasting an average ranking of 1.7 and the most minimal standard deviation. In contrast, the next best method lags with a ranking of 2.8 and a standard deviation of 0.9. The code is available at https://github.com/Dendiiiii/DiffImpute.

Download This Paper Open PDF in Browser Add Paper to My Library Share: Permalink Using these links will ensure access to this page indefinitely Copy URL Copy DOI

Crafting pertinent policies to facilitate the high-level integration of culture and tourism has now become a vital agenda within the current discourse in China. However, relatively little is known about the actual implementation of various policies to achieve a high-level integration, especially how combinations of policy instruments are deployed in the process of realization. Based on the Policy Instrument Theory, this study uses fuzzy-set qualitative comparative analysis on a sample of 31 provincial administrative regions in China to investigate the influence of typical policy instruments on the integration level of tourism and culture. The results show that each single policy tool is not necessary for high-level integration of culture and tourism. On the contrary, only through an organic combination of different policy tools can affect the integration level. This study also summarizes five policy instrument configurations, which can be grouped into four driving modes of culture-tourism integration: the environment-driven supply-demand coordination mode, supply-driven demand-environment coordination mode, supply-driven mode, and supply-driven environment coordination mode. This study considerably provides critical theoretical and practical insights into the integration of culture and tourism from the perspective of governmental policies.

The discovery of numerous new charmonium-like structures since 2003 have revitalized interest in exotic meson spectroscopy. These structures do not fit easily into the conventional charmonium model, and proposals like four-quark states, hybrids and rescattering effects have been suggested as explanations. Since 2009, several new structures were reported in the J/ψϕ mass spectrum with the following characteristics: they are the first ones reported decaying into two heavy mesons which contain both a [Formula: see text] pair and a [Formula: see text] pair; and their masses are well beyond the open charm pair threshold. Conventional [Formula: see text] states with a mass beyond the J/ψϕ threshold are not expected to decay into this channel and the width is expected to be large, thus they are good candidates for exotic mesons. The focus of this paper is to review the recent developments on the structures in the J/ψϕ mass spectrum from CDF, Belle and LHCb.

In recent years, surface defect detection methods based on deep learning have been widely applied to steel plate surface defect detection. By locating and classifying defects on the surface of steel plates, production efficiency can be improved. However, there is still a conflict between speed and accuracy in the defect detection process. To address this issue, we propose a high-precision, low-latency surface defect detection algorithm called the GhostConv-ECA-YOLOv5 Network (GEA-Net). The GEA-Net model can predict defect categories without compromising classification and detection accuracy. Experimental results show that our proposed improved model has higher performance compared to other comparative models, achieving a 75.6% mAP on the NEU-DET dataset.

Food is closely related to national economy and people's livelihood. Rice is the largest grain crop in China, it is crucial to predict the loss rate of rice during processing to reduce food waste and ensure food security. This study first obtained the loss rate of rice processing through the recovery survey form of enterprises. Then, prediction was carried out using two common models: the BP neural network and multiple linear regression. Finally, the genetic algorithm was applied to optimize the BP neural network for further prediction and com-pared with the original models. The experimental results showed that the GA-BP model had higher prediction accuracy and smaller error compared to the first two models. It is valuable in reducing processing losses and maintaining food security.

Protein engineering faces challenges in finding optimal mutants from a massive pool of candidate mutants. In this study, we introduce a deep-learning-based data-efficient fitness prediction tool to steer protein engineering. Our methodology establishes a lightweight graph neural network scheme for protein structures, which efficiently analyzes the microenvironment of amino acids in wild-type proteins and reconstructs the distribution of the amino acid sequences that are more likely to pass natural selection. This distribution serves as a general guidance for scoring proteins toward arbitrary properties on any order of mutations. Our proposed solution undergoes extensive wet-lab experimental validation spanning diverse physicochemical properties of various proteins, including fluorescence intensity, antigen–antibody affinity, thermostability, and DNA cleavage activity. More than 40% of ProtLGN-designed single-site mutants outperform their wild-type counterparts across all studied proteins and targeted properties. More importantly, our model can bypass the negative epistatic effect to combine single mutation sites and form deep mutants with up to seven mutation sites in a single round, whose physicochemical properties are significantly improved. This observation provides compelling evidence of the structure-based model's potential to guide deep mutations in protein engineering. Overall, our approach emerges as a versatile tool for protein engineering, benefiting both the computational and bioengineering communities.

Isotactic polyacrylonitrile (PAN) with triad isotacticity of 0.53, which was determined by 13 C NMR, using dialkylmagnesium as an initiator, was successfully synthesized.Isothermal treatment of iso-PAN was conducted in air at 200, 220, 250 and 280 o C. Structural evolutions and chemical changes were studied with Fourier transformation infrared and wide-angle X-ray diffraction during stabilization.A new parameter CNF = I 2240cm -1 / (I 1595cm -1 +f*I 1595cm -1 ) was defined to evaluate residual nitrile groups.Crystallinity and crystal size were calculated with X-ray diffraction dates.The results indicated that the nitrile groups had partly converted into a ladder structure as stabilization proceeded.The rate of reaction increased with treatment temperature; crystallinity and crystal size decreased proportionally to pyrolysis temperature.The iso-conversional method coupled with the Kissinger and Flynn-Wall-Ozawa methods were used to determine kinetic parameters via differential scanning calorimetry analysis with different heating rates.The active energy of the reaction was 171.1 and 169.1 kJ/mol, calculated with the two methods respectively and implied the sensitivity of the reaction with temperature.

Nephrolithiasis is one of the most prevalent diseases of the urinary system. Approximately 80% of human kidney stones are composed of calcium oxalate (CaOx), and hypercalciuria is one of the most common metabolic disorders. Emerging evidence indicates that autophagy and inflammatory responses are related to the formation of CaOx nephrolithiasis. However, the roles of autophagy and inflammation in patients with hypercalciuria remain unclear. Ethyl pyruvate (EP) displays protective effects in experimental models of many illnesses. In this study, we investigated the protective effects of EP in vitro through its inhibition of autophagy and inflammatory responses after CaCl2-induced tubular epithelial cell injury.First, we cultured human tubular epithelial (HK-2) cells in the presence of various concentrations of CaCl2 (0, 0.1, 0.25, 0.5, 1.0, 1.5, and 2.0 mg/ml) for 12 h and EP (0, 1.0, 2.5, 5.0, and 10.0 mM) for 2 h to select the optimum concentration using the Cell Counting Kit-8 assay and lactate dehydrogenase (LDH) assay. Cells in culture were stimulated with CaCl2 (1.0 mg/ml, 12 h) with or without EP pretreatment (2.5 mM, 2 h). After the exposure, we detected the expression of inflammation-related proteins using an enzyme-linked immunosorbent assay and Western blot analysis. Finally, the levels of autophagy-related proteins were determined through Western blot analysis, and the number of GFP-LC3 dots and autophagic vacuoles was detected under confocal microscopy.With the use of the Cell Counting Kit-8 assay and the LDH assay, we identified the optimum concentration for CaCl2 (1.0 mg/ml) treatment and EP pretreatment (2.5 mM). Our research indicated that CaCl2 can induce autophagy and inflammatory responses in HK-2 cells. Furthermore, treatment with EP prior to CaCl2 stimulation attenuated HK-2 cell injury by inhibiting autophagy and inflammation.Our results provide evidence that EP attenuates CaCl2-induced injury of HK-2 cells by downregulating the expression of inflammation and autophagy proteins that may be associated with the inhibition of the high-mobility group box-1 (HMGB1)/toll-like receptor 4 (TLR4)/NF-κB pathway and the competitive interaction with Beclin-1 of HMGB1.

A fluid consisting of nano-sized particles is known as nanofluid. It manifests better cooling abilities compared to the host fluid. The aim of this study is to analyze the AC and impulse breakdown strength of mineral oil based SiO <sub xmlns:mml="http://www.w3.org/1998/Math/MathML" xmlns:xlink="http://www.w3.org/1999/xlink">2</sub> nanofluids (NFs). The volume fractions used was 20%. The influence of moisture on AC breakdown strength was investigated. Moreover, the positive and negative impulse breakdown strength was measured and compared with pure oil. The positive lighting impulse breakdown voltages indicated an enhancement as compared to pure oil where the measured negative impulse breakdown voltages of nanofluids were lower than the host oil. The modification mechanisms of nanoparticles on breakdown properties of mineral oil were also discussed.

The transparent cornea is the window of the eye, facilitating the entry of light rays and controlling focusing the movement of the light within the eye. The cornea is critical, contributing to 75% of the refractive power of the eye. Keratoconus is a progressive and multifactorial corneal degenerative disease affecting 1 in 2000 individuals worldwide. Currently, there is no cure for keratoconus other than corneal transplantation for advanced stage keratoconus or corneal cross-linking, which can only halt KC progression. The ability to accurately identify subtle KC or KC progression is of vital clinical significance. To date, there has been little consensus on a useful model to classify KC patients, which therefore inhibits the ability to predict disease progression accurately.In this paper, we utilised machine learning to analyse data from 124 KC patients, including topographical and clinical variables. Both supervised multilayer perceptron and unsupervised variational autoencoder models were used to classify KC patients with reference to the existing Amsler-Krumeich (A-K) classification system. Both methods result in high accuracy, with the unsupervised method showing better performance. The result showed that the unsupervised method with a selection of 29 variables could be a powerful tool to provide an automatic classification tool for clinicians. These outcomes provide a platform for additional analysis for the progression and treatment of keratoconus.

Mining different types of communities from web data have attracted a lot of research efforts in recent years. However, none of the existing community mining techniques has taken into account both the dynamic as well as heterogeneous nature of web data. In this paper, we propose to characterize and predict community members from the evolution of heterogeneous web data. We first propose a general framework for analyzing the evolution of heterogeneous networks. Then, the academic network, which is extracted from 1 million computer science papers, is used as an example to illustrate the framework. Finally, two example applications of the academic network are presented. Experimental results with a real and very large heterogeneous academic network show that our proposed framework can produce good results in terms of community member recommendation. Also, novel knowledge and insights can be gained by analyzing the community evolution pattern.

This article summarizes a brief development of exotic mesons with Vector-Vector (VV) final states starting from the $J/\psi\phi$ mass spectrum, as well as extensions to other final states for strong-interaction physics and physics beyond standard model. These VV final states are very suitable to be studied at CEPC through associate production and two-photon production, as well as a visionary low energy photon collider.

Nanoparticles have the potential to affect the insulation performance of transformer oil. The different types of nanoparticles (NPs) influence the breakdown strength of mineral oil. In this paper, mineral oil based nanofluids (NFs) have been prepared by dispersing different types of nanoparticles into transformer oil. The AC and positive lightening impulse breakdown strengths of the oil samples with and without nanoparticles were measured according to IEC standard methods. The results manifested that the addition of different types of nanoparticles to the base oil can modify the breakdown strength of the base oil. A possible mechanism of improvement of nanoparticles was also used to explain the difference among different nanofluids and base oil. The saturation charges for all nanoparticles was calculated to see whether it is useful to describe the breakdown strength results of different nanofluids but it was concluded it may not be used to explain the breakdown strength of different nanofluids solely but there are some other physical factor need to be considered as well.

In this paper, we aim to improve the data efficiency of image captioning. We propose VisualGPT, a data-efficient image captioning model that leverages the linguistic knowledge from a large pretrained language model (LM). A crucial challenge is to balance between the use of visual information in the image and prior linguistic knowledge acquired from pretraining.We designed a novel self-resurrecting encoder-decoder attention mechanism to quickly adapt the pretrained LM as the language decoder on a small amount of in-domain training data. The pro-posed self-resurrecting activation unit produces sparse activations but is not susceptible to zero gradients. When trained on 0.1%, 0.5% and 1% of MSCOCO and Conceptual Captions, the proposed model, VisualGPT, surpasses strong image captioning baselines. VisualGPT outperforms the best baseline model by up to 10.8% CIDEr on MS COCO and up to 5.4% CIDEr on Conceptual Captions.We also perform a series of ablation studies to quantify the utility of each system component. To the best of our knowledge, this is the first work that improves data efficiency of image captioning by utilizing LM pretrained on unimodal data. Our code is available at: this https URL.

Transformer oil-based nanofluids were prepared by suspending non-conductive Al2O3 nanoparticles to enhance the insulating properties of transformer oil. Breakdown voltages under positive and negative lightening impulse voltages were measured for all the prepared samples according to IEC standards. The results indicated that the inclusion of Al2O3 nanoparticles to the base oil can modify the breakdown strength of the base oil. The results showed that addition of alumina nanoparticles improve the mean lightening impulse breakdown voltages of nanofluids and were 1.09 times as compared to base transformer oil but not the negative lightening impulse breakdown strength. A possible mechanism of insulative nanoparticles was also used to explain the difference among nanofluids and base oil.

Cosmic microwave background radiation (CMB) is critical to the understanding of the early universe and precise estimation of cosmological constants. Due to the contamination of thermal dust noise in the galaxy, the CMB map that is an image on the two-dimensional sphere has missing observations, mainly concentrated on the equatorial region. The noise of the CMB map has a significant impact on the estimation precision for cosmological parameters. Inpainting the CMB map can effectively reduce the uncertainty of parametric estimation. In this paper, we propose a deep learning-based variational autoencoder - CosmoVAE, to restoring the missing observations of the CMB map. The input and output of CosmoVAE are square images. To generate training, validation, and test data sets, we segment the full-sky CMB map into many small images by Cartesian projection. CosmoVAE assigns physical quantities to the parameters of the VAE network by using Fourier coefficients, which are sampled by the angular power spectrum of the Gaussian random field as latent variables. CosmoVAE adopts a new loss function to improve the learning performance of the model, which consists of ℓ <sub xmlns:mml="http://www.w3.org/1998/Math/MathML" xmlns:xlink="http://www.w3.org/1999/xlink">1</sub> reconstruction loss, Kullback-Leibler divergence between the posterior distribution of encoder network and the prior distribution of latent variables, perceptual loss, and total-variation regularizer. The proposed model achieves state of the art performance for Planck Commander 2018 CMB map inpainting.

The search for a dark photon produced at e+e− colliders which subsequently decays into inelastic dark matter particles, is discussed. The heavier dark matter decays into a pair of visible charged particles and a lighter dark matter particle after traveling some distance. The visible decay products can be recorded by a dark matter detector made of emulsions and gas detectors, placed near the main e+e− detector. This setup can not only explore new parameter regions not reached before, but also re-open some regions thought to be excluded by previous experimental data. The physics potential for such a detector around BESIII and Belle II is presented.

Multiple rice bodies in the wrist is a rare disorder that requires surgery, and there are still many uncertainties regarding its diagnosis and treatment.We described a rare case of chronic idiopathic tenosynovitis with rice bodies of the wrist in a 71-year-old man and reviewed similar topics in the literature. A total of 43 articles and 61 cases were included in the literature review. Our case had a usual presentation: it was similar to those in the literature. The affected population was mainly older adults, with an average age of 59.43 (range, 3 to 90) years. The male-to-female ratio was 1.54:1 (37/24).Most of them showed limited swelling and pain, only 23.0% had carpal tunnel symptoms, and the average disease duration was 18.03 (0.5-60) mo. Wrist flexor tendon sheath involvement was the most common (95.1%, 58/61), and only 3 cases had extensor tendon sheath involvement.The main causes were tuberculosis (34.4%, 21/61), non-tuberculous mycobacteria (24.6%, 15/61), idiopathic tenosynovitis (31.1%, 19/61), and others (9.84%, 6/61). There were 10 patients with recurrences; in 6 of them, were due to non-tuberculous mycobacterial infections.We reported a case of wrist idiopathic tenosynovitis with rice body formation, and established a clinical management algorithm for wrist tenosynovitis with rice bodies, which can provide some reference for our clinical diagnosis and treatment. The symptoms of rice-body bursitis of the wrist are insidious, nonspecific, and difficult to identify. The aetiology is mainly idiopathic tenosynovitis and mycobacterial (tuberculosis or non-tuberculous) infections; the latter are difficult to treat and require long-duration systemic combination antibiotic therapies. Therefore, before a diagnosis of idiopathic tenosynovitis is made, we must exclude other causes, especially mycobacterial infections.

The main question we address in this paper is how to scale up visual recognition of unseen classes, also known as zero-shot learning, to tens of thousands of categories as in the ImageNet-21K benchmark. At this scale, especially with many fine-grained categories included in ImageNet-21K, it is critical to learn quality visual semantic representations that are discriminative enough to recognize unseen classes and distinguish them from seen ones. We propose a Hierarchical Graphical knowledge Representation framework for the confidence-based classification method, dubbed as HGR-Net. Our experimental results demonstrate that HGR-Net can grasp class inheritance relations by utilizing hierarchical conceptual knowledge. Our method significantly outperformed all existing techniques, boosting the performance by 7% compared to the runner-up approach on the ImageNet-21K benchmark. We show that HGR-Net is learning-efficient in few-shot scenarios. We also analyzed our method on smaller datasets like ImageNet-21K-P, 2-hops and 3-hops, demonstrating its generalization ability. Our benchmark and code are available at https://kaiyi.me/p/hgrnet.html .

Soil salinization has become a major environmental factor severely threatening global food security. The application of porous minerals could significantly ameliorate soil fertility and promote plant productivity under salt stress conditions. However, the effects of porous minerals on improving the salt resistance of grain crops in coastal saline soils is not fully studied. In this work, the shoot growth and grain yield of wheat plants grown in coastal saline fields, respectively amended with the four naturally available porous minerals, diatomite, montmorillonite, bentonite and zeolite, were assessed. The application of porous minerals, especially zeolite, significantly improved the biomass and grain yield of wheat plants under saline conditions, as demonstrated by the augmented plant fresh mass (14.8~61.2%) and increased seed size (3.8~58.8%) and number (1.4~57.5%). Soil property analyses exhibited that porous-mineral amendment decreased soil sodium content and sodium absorption ratio, and increased soil nutrients in both the rhizosphere and nonrhizosphere of wheat plants. Further quantitative-PCR and 16S high-throughput sequencing analysis revealed that porous-mineral application also remarkably increased the abundance of bacterial 16S rRNA (0.8~102.4%) and fungal 18S rRNA (89.2~209.6%), and altered the composition of the soil microbial community in the rhizosphere of wheat. Our findings suggest that zeolite could be used as an ideal salt soil amendment, and the changes in soil properties and microorganisms caused by the application of porous minerals like zeolite improved the salt resistance of wheat plants in coastal saline land, leading to increased shoot growth and seed production.

The biochar and organic fertilizer amendment have been used as an effective practice to increase soil fertility. Nevertheless, the mechanisms of microbial community response to organic fertilizer and biochar application on saline-alkali soil have not been clarified. This study investigated the effects at different concentrations of organic fertilizer and biochar on the microbial community of wheat rhizosphere soil under field experiment in the Yellow River Delta (China, YRD), using high-throughput sequencing technology. Biochar and organic fertilizer significantly influenced in most soil parameters (p < 0.05), apart from soil moisture content (M), pH, total nitrogen (TN) and soil total phosphorus (TP). Proteobacteria and Actinobacteriota were found in the rhizosphere soil as the main bacterial phyla, and the main fungal phyla were Ascomycota and Mortierellomycota. The soil bacterial and fungal communities under organic fertilizer were distinct from CK. Furthermore, redundancy analysis (RDA) directed that changes in bacterial communities were related to soil properties like pH, available phosphorus (AP), and total organic carbon (TOC), while pH, AP and TP, were crucial contributors in regulating fungal distribution. The correlation between soil parameters and bacteria or fungi varied with the application of biochar and organic fertilizers, and the interaction between the bacteria and fungi in organic fertilizer treatments formed more connections compared with biochar treatments. Our results indicated that biochar was superior to organic fertilizer under the contents set up in this study, and soil parameters increased with biochar and organic fertilizer application rate. The diversity and structure of soil bacteria and fungi differed with the application of biochar and organic fertilizer. The research provides a reference to rational application of organic fertilizer and biochar improvement in saline-alkali soil.

Abstract Protein engineering faces challenges in finding optimal mutants from the massive pool of candidate mutants. In this study, we introduce a deep learning-based data-efficient fitness prediction tool to steer protein engineering. Our methodology establishes a lightweight graph neural network scheme for protein structures, which efficiently analyzes the microenvironment of amino acids in wild-type proteins and reconstructs the distribution of the amino acid sequences that are more likely to pass natural selection. This distribution serves as a general guidance for scoring proteins toward arbitrary properties on any order of mutations. Our proposed solution undergoes extensive wet-lab experimental validation spanning diverse physicochemical properties of various proteins, including fluorescence intensity, antigen-antibody affinity, thermostability, and DNA cleavage activity. More than 40% of P rot LGN-designed single-site mutants outperform their wild-type counterparts across all studied proteins and targeted properties. More importantly, our model can bypass the negative epistatic effect to combine single mutation sites and form deep mutants with up to 7 mutation sites in a single round, whose physicochemical properties are significantly improved. This observation provides compelling evidence of the structure-based model’s potential to guide deep mutations in protein engineering. Overall, our approach emerges as a versatile tool for protein engineering, benefiting both the computational and bioengineering communities.

Well-dispersed Fe3O4 nanoparticles were prepared and modified by oleic acid at the temperature of 70°C for different reaction times. The surface modification of oleic acid could improve dispersion stability of Fe3O4 nanoparticles in vegetable oil by forming an effective chemisorbed modification layer on the surface of nanoparticles. The interaction between coordination pattern and mass faction of chemisorbed oleic acid on nanoparticles and dispersion stability of nanofluid was discussed. The optimal dispersion stability of vegetable-oil based Fe3O4 nanofluids was obtained to the nanoparticles modified for 8 hours, which has the highest AC breakdown strength of 67.78 kV.

Nanofluids were prepared by adding magnetic nanoparticles to enhance the dielectric accomplishment of transformer oil. Breakdown voltages are tested under AC and lightning impulse stresses of all the prepared fluids in accordance to IEC standards. The test outcomes showed that the suspension of magnetic nanoparticles (NPs) to the carrier oil may upgrade the mean AC breakdown performance 1.14 times of that for base oil. Moreover, the mean lightning impulse breakdown voltages of nanofluids were also improved than that of base transformer oil and were 1.36 times as compared to base oil. A possible mechanism of conductive nanoaprticles was also used to explain the difference among nanofluids and base oil.

PAN precursor fibers were produced via wet-spinning process, and effects of polymerization and spinning processes, especially the stretching process, were investigated on mechanical properties and micro-morphologies of precursor fibers.An increase in molecular weight, dope solid and densification and a decrease in surface defects were possible by controlling polymerization temperature, the number of heating rollers for densification and the jet stretch ratio, which improved the mechanical properties of precursor fibers.The curves for strength, modulus, tensile power and diameter as a function of stretch ratio can be divided into three stages: steady change area, little change area and sudden change area.With the increase of stretch ratio, the fiber diameter became smaller, the degree of crystallization increased and the structure of precursor fibers became compact and homogeneous, which resulted in the increase of strength, modulus and tensile power of precursor fibers.Empirical relationship between fiber strength and stretch ratio was studied by using the sub-cluster statistical theory.It was successfully predicted when the strengths were 0.8 GPa and 1.0 GPa under a certain technical condition, the corresponding stretch ratio of the fiber were 11.16 and 12.83 respectively.

Nanoparticles have manifested the potential to enhance the breakdown performance of transformer oil. The AC and positive lightening impulse breakdown strengths of the oil samples with and without nanoparticles were measured according to IEC standard methods. The test results indicate that addition of silica nanoparticles can enhance the ac breakdown strength of transformer oil. Additionally, the mean lightening impulse breakdown voltages of prepared nanofluids were also enhanced than that of base transformer oil. Possible modification mechanisms of SiO <sub xmlns:mml="http://www.w3.org/1998/Math/MathML" xmlns:xlink="http://www.w3.org/1999/xlink">2</sub> nanoparticles on breakdown properties of transformer oil were also discussed.

Nanoparticles have the potential to enhance the insulation conduct of transformer oil. The concentration of nanoparticles influences the breakdown vitality of mineral oil. In this investigation, mineral oil based nanofluids were prepared by scattering magnetic nanoparticles into transformer oil with different concentrations from 5% to 80% w/v. The AC and lightning impulse breakdown strengths of the oil samples with and without nanoparticles were investigated in accordance with IEC standard methods. The test outcomes indicate that addition of magnetic nanoparticles can enhance the insulation strength of transformer oil. With the increase of nanoparticle concentrations, the AC and positive impulse breakdown strength of transformer oil are first increased and up to the maximum value at the concentration of 40%. After which the breakdown strength start decreasing. The result of negative impulse breakdown showed that the breakdown voltage of nanofluids with multiple concentrations were lower than the breakdown strength of pure transformer oil. The probable modification mechanisms of Fe <sub xmlns:mml="http://www.w3.org/1998/Math/MathML" xmlns:xlink="http://www.w3.org/1999/xlink">3</sub> O <sub xmlns:mml="http://www.w3.org/1998/Math/MathML" xmlns:xlink="http://www.w3.org/1999/xlink">4</sub> nanoparticles on dielectric features of transformer oil were also reviewed.

Monodisperse Fe3O4 nanoparticles modified by oleic acid were synthesized by a solvothermal method and dispersed into mineral transformer oil to prepare Fe3O4 nanofluids. X-ray diffraction (XRD) and transmission electron microscopy (TEM) analysis indicated that the obtained nanoparticles are single crystals with an average diameter of 15 nm. The test results indicate that the nanoparticles exhibited good dispersibility in the nanofluids at a wide range of concentrations. Moreover, the positive impulse breakdown strength of nanofluids was greatly improved by 36.6% at the optimum concentration of nanoparticles.

Wetspun nascent PAN fibers were immersed into a DMSO/H 2 O coagulation bath. Diffusion of DMSO from nascent fiber and that of H 2 O into nascent fiber were studied at different temperatures and concentrations of coagulation bath. The diffusion coefficient of H 2 O is larger than that of DMSO. As the coagulation bath temperature increased, the diffusion coefficients of DMSO and H 2 O increased. Diffusion activation energy of DMSO is smaller than that of H 2 O during the diffusion. Cross sectional and surface structure of nascent PAN fiber were observed in relation to coagulation ability (the product of diffusion coefficients of two solvents). Coagulation ability was found to have a significant influence on both the cross sectional and surface morphology of nascent fiber. To obtain nascent fiber with circular cross sectional and smooth surface morphology, coagulation ability should be controlled at low value.

Accurate Traffic Sign Detection (TSD) can help drivers make better decision according to the traffic regulations. TSD, regarded as a typical small object detection problem in some way, is fundamental in the field of self-driving and advanced driver assistance systems. However, small object detection is still an open question. In this paper, we proposed a human brain inspired network to handle this problem. Attention mechanism is an essential function of our brain, we used a novel recurrent attentive neural network to improve the detection accuracy in a fine-grained manner. Further, as we human can combine domain specific knowledge and intuitive knowledge to solve tricky tasks, we proposed an assumption that the location of the traffic signs obeys the reverse gaussian distribution, which means the location is around the central bias of every picture. Experimental result shows that our methods achieved better performance than several popular methods used in object detection.

Neural message passing is a basic feature extraction unit for graph-structured data considering neighboring node features in network propagation from one layer to the next. We model such process by an interacting particle system with attractive and repulsive forces and the Allen-Cahn force arising in the modeling of phase transition. The dynamics of the system is a reaction-diffusion process which can separate particles without blowing up. This induces an Allen-Cahn message passing (ACMP) for graph neural networks where the numerical iteration for the particle system solution constitutes the message passing propagation. ACMP which has a simple implementation with a neural ODE solver can propel the network depth up to one hundred of layers with theoretically proven strictly positive lower bound of the Dirichlet energy. It thus provides a deep model of GNNs circumventing the common GNN problem of oversmoothing. GNNs with ACMP achieve state of the art performance for real-world node classification tasks on both homophilic and heterophilic datasets.

Searches for new physics in the single and paired dijet mass spectrum are performed using data collected by the CMS experiment at the LHC at a collision energy of TeV. No evidence for new physics is found and upper limits are set for various models. At 95% confidence level, a string resonance in the single dijet spectrum is excluded for masses between 1 and 4.7 TeV and, for the first time, a coloron in the paired dijet spectrum is excluded for masses between 250 and 740 GeV.

At present, the performance of deep neural network in general object detection is comparable to or even surpasses that of human beings. However, due to the limitations of deep learning itself, the small proportion of feature pixels, and the occurence of blur and occlusion, the detection of small objects in complex scenes is still an open question. But we can not deny that real-time and accurate object detection is fundamental to automatic perception and subsequent perception-based decision-making and planning tasks of autonomous driving. Considering the characteristics of small objects in autonomous driving scene, we proposed a novel method named KB-RANN, which based on domain knowledge, intuitive experience and feature attentive selection. It can focus on particular parts of image features, and then it tries to stress the importance of these features and strengthenes the learning parameters of them. Our comparative experiments on KITTI and COCO datasets show that our proposed method can achieve considerable results both in speed and accuracy, and can improve the effect of small object detection through self-selection of important features and continuous enhancement of proposed method, and deployed it in our self-developed autonomous driving car.

The human brain is composed of multiple modular subsystems, with a unique way interacting among each other. These subsystems have their own unique characteristics and interact to support cognitive functions such as memory, attention and cognitive control. Nowadays, deep learning methods based on the above-mentioned functions accompanied with knowledge are widely used to design more dynamic, robust and powerful systems. We first review and summarize the progresses of cognition-based deep neural networks, and how cognitive mechanisms can be applied to more brain-like neural networks. Then we propose a general framework for the design of cognition-based deep learning system. Although great efforts have been made in this field, cognition-based deep learning is still in its early age. We put forward the potential directions towards this field, such as associative memory in deep learning, interpretable network with cognitive mechanisms, and deep reinforcement learning based on cognitive science.

At present, the performance of deep neural network in general object detection is comparable to or even surpasses that of human beings. However, due to the limitations of deep learning itself, the small proportion of feature pixels, and the occurence of blur and occlusion, the detection of small objects in complex scenes is still an open question. But we can not deny that real-time and accurate object detection is fundamental to automatic perception and subsequent perception-based decision-making and planning tasks of autonomous driving. Considering the characteristics of small objects in autonomous driving scene, we proposed a novel method named KB-RANN, which based on domain knowledge, intuitive experience and feature attentive selection. It can focus on particular parts of image features, and then it tries to stress the importance of these features and strengthenes the learning parameters of them. Our comparative experiments on KITTI and COCO datasets show that our proposed method can achieve considerable results both in speed and accuracy, and can improve the effect of small object detection through self-selection of important features and continuous enhancement of proposed method, and deployed it in our self-developed autonomous driving car.

The ability to quickly learn from a small quantity oftraining data widens the range of machine learning applications. In this paper, we propose a data-efficient image captioning model, VisualGPT, which leverages the linguistic knowledge from a large pretrained language model(LM). A crucial challenge is to balance between the use of visual information in the image and prior linguistic knowledge acquired from pretraining. We designed a novel self-resurrecting encoder-decoder attention mechanism to quickly adapt the pretrained LM as the language decoder ona small amount of in-domain training data. The proposed self-resurrecting activation unit produces sparse activations but has reduced susceptibility to zero gradients. We train the proposed model, VisualGPT, on 0.1%, 0.5% and 1% of MSCOCO and Conceptual Captions training data. Under these conditions, we outperform the best baseline model by up to 10.8% CIDEr on MS COCO and upto 5.4% CIDEr on Conceptual Captions. Further, Visual-GPT achieves the state-of-the-art result on IU X-ray, a medical report generation dataset. To the best of our knowledge, this is the first work that improves data efficiency of image captioning by utilizing LM pretrained on unimodal data. Our code is available at: https://github.com/Vision-CAIR/VisualGPT.

We illustrate the influence of the dipeptide structure on photosensitive damage and the kinetic mechanism was investigated using acenaphthenequinone (ACQ) as a triplet photosensitizer. With tyrosine (Tyr) serving as the core structure, two classic dipeptides with double (trptophan-tyrosine, Trp-Tyr) and single (tyrosine-alanine, Tyr-Ala and Ala-Tyr) active reaction sites were constructed, and the underlying photodamage mechanisms were investigated carefully. According to the experimental results, the proton-coupled electron transfer processes between ACQ and numerous Trp-Tyr reaction sites have independent reaction properties. The bimolecular quenching rate (kq) value is roughly equivalent to the sum of the rates of two amino acid monomers, and a novel intramolecular dynamic channel between Trp/N˙-Tyr and Trp-Tyr/O˙ was observed. The ACQ/Tyr-Ala system demonstrated the key role of steric hindrance on the kq in bimolecular reactions.

Studying the dynamic failure of materials is crucial because this failure occurs extensively in mechanics and earthquake fields. However, obtaining the deformation and temperature fields in high-speed dynamic cases is challenging, particularly when analyzing the mechanics of thermo-mechanical coupling. This study developed a thermo-mechanical coupled in-situ measurement system to acquire synchronized information on the load, deformation and temperature fields during a specimen dynamic failure process with high temporal and spatial resolution. In particular, the deformation field was obtained using a high-speed camera connected to a microscope that could achieve a spatial resolution of 3.6–3.8μm/pixel and a temporal resolution of 1μs. Moreover, a 64 × 64 focal plane array infrared detector was developed to measure temperatures with the time resolution of 1μs. A Newtonian optical system was applied to achieve the infrared detector’s spatial resolution of 25μm. The temperature measurement system could measure temperatures ranging from 100 to 800 °C. The split-Hopkinson pressure bar was used as the dynamic loading device to realize the high strain rate loading of a Ti-6Al-4V hat-shaped specimen, verifying the deformation and temperature measurement capabilities. The results from the Ti-6Al-4V specimen under dynamic loading showed an evident adiabatic shear band and the temperature rise, suggesting that the experimental devices provided was an effective tool to study the thermo-mechanical response of materials under dynamic loading.

In this paper, we propose an efficient simulation method based on adaptive importance sampling, which can automatically find the optimal proposal within the Gaussian family based on previous samples, to evaluate the probability of bit error rate (BER) or word error rate (WER). These two measures, which involve high-dimensional black-box integration and rare-event sampling, can characterize the performance of coded modulation. We further integrate the quasi-Monte Carlo method within our framework to improve the convergence speed. The proposed importance sampling algorithm is demonstrated to have much higher efficiency than the standard Monte Carlo method in the AWGN scenario.

Directed evolution as a widely-used engineering strategy faces obstacles in finding desired mutants from the massive size of candidate modifications. While deep learning methods learn protein contexts to establish feasible searching space, many existing models are computationally demanding and fail to predict how specific mutational tests will affect a protein's sequence or function. This research introduces a lightweight graph representation learning scheme that efficiently analyzes the microenvironment of wild-type proteins and recommends practical higher-order mutations exclusive to the user-specified protein and function of interest. Our method enables continuous improvement of the inference model by limited computational resources and a few hundred mutational training samples, resulting in accurate prediction of variant effects that exhibit near-perfect correlation with the ground truth across deep mutational scanning assays of 19 proteins. With its affordability and applicability to both computer scientists and biochemical laboratories, our solution offers a wide range of benefits that make it an ideal choice for the community.

Federated Learning is an evolving machine learning paradigm, in which multiple clients perform computations based on their individual private data, interspersed by communication with a remote server. A common strategy to curtail communication costs is Local Training, which consists in performing multiple local stochastic gradient descent steps between successive communication rounds. However, the conventional approach to local training overlooks the practical necessity for client-specific personalization, a technique to tailor local models to individual needs. We introduce Scafflix, a novel algorithm that efficiently integrates explicit personalization with local training. This innovative approach benefits from these two techniques, thereby achieving doubly accelerated communication, as we demonstrate both in theory and practice.

Soil salinization has become a major environmental factor severely threatening global food security. Application of porous minerals could significantly ameliorate soil fertility and promote plant productivity under salt stress condition. However, the mechanism of porous minerals improving the salt resistance of grain crops in coastal saline soils is still unclear. In this study, the effects of four naturally available porous minerals, diatomite, montmorillionite, bentonite and zeolite, on shoot growth and grain yield of wheat plants grown in coastal saline field were assessed. Application of porous minerals under saline condition significantly improved the biomass and grain yield of wheat plants, as demonstrated by the augmented plant fresh weight and increased seed size and number. Soil physiochemical property analyses exhibited that porous mineral amendment decreased soil sodium (Na+) content and sodium absorption ratio (SAR), and increased soil fertility in both the rhizosphere and non-rhizosphere of wheat plants. Further microbial analyses revealed that porous mineral application also remarkably increased the abundance and altered the composition of soil microbial community. Our findings suggested that the changes of soil physiochemical properties and microorganisms caused by porous mineral application improved the resistance of wheat plants to salt stress in coastal saline lands, leading to increased shoot growth and seed production.

Learning novel concepts, remembering previous knowledge, and adapting it to future tasks occur simultaneously throughout a human's lifetime. To model such comprehensive abilities, continual zero-shot learning (CZSL) has recently been introduced. However, most existing methods overused unseen semantic information that may not be continually accessible in realistic settings. In this paper, we address the challenge of continual zero-shot learning where unseen information is not provided during training, by leveraging generative modeling. The heart of the generative-based methods is to learn quality representations from seen classes to improve the generative understanding of the unseen visual space. Motivated by this, we introduce generalization-bound tools and provide the first theoretical explanation for the benefits of generative modeling to CZSL tasks. Guided by the theoretical analysis, we then propose our learning algorithm that employs a novel semantically guided Generative Random Walk (GRW) loss. The GRW loss augments the training by continually encouraging the model to generate realistic and characterized samples to represent the unseen space. Our algorithm achieves state-of-the-art performance on AWA1, AWA2, CUB, and SUN datasets, surpassing existing CZSL methods by 3-7\%. The code has been made available here \url{https://github.com/wx-zhang/IGCZSL}

Criminal digital image processing technology is becoming increasingly important in public security operations, especially in the identification of characters in surveillance videos and the analysis of behavior and actions in on-site videos, where blurry portraits in videos are particularly important. The paper mainly focuses on how to restore facial features of blurred human figures in video surveillance. Two methods, inverse filtering and Wiener filtering, are used to process blurred human figures, ultimately achieving the expected results. This provides assistance in identifying the same case as a major and difficult case.

In this paper, we propose an efficient simulation method based on adaptive importance sampling, which can automatically find the optimal proposal within the Gaussian family based on previous samples, to evaluate the probability of bit error rate (BER) or word error rate (WER). These two measures, which involve high-dimensional black-box integration and rare-event sampling, can characterize the performance of coded modulation. We further integrate the quasi-Monte Carlo method within our framework to improve the convergence speed. The proposed importance sampling algorithm is demonstrated to have much higher efficiency than the standard Monte Carlo method in the AWGN scenario.

In High-Energy Physics experiments it is often necessary to evaluate the global statistical significance of apparent resonances observed in invariant mass spectra. One approach to determining significance is to use simulated events to find the probability of a random fluctuation in the background mimicking a real signal. As a high school summer project, we demonstrate a method with Monte Carlo simulated events to evaluate the global significance of a potential resonance with some assumptions. This method for determining significance is general and can be applied, with appropriate modification, to other resonances.

Similarity search for trajectories, especially the top-k similarity query, has been widely used in different fields, such as personalized travel route recommendation, car pooling, etc. Previous works have studied top-k similarity trajectory query in plaintext, but the increasing attention to privacy protection makes top-k similarity query on trajectory data become a challenge. In this paper, we propose a privacy-preserving top-k similarity query scheme over large-scale trajectory data based on Hilbert curve and homomorphic encryption. Towards this end, we first define a spatio-temporal trajectory similarity measure that supports homomorphic computation under ciphertext based on numerical integration algorithm for discrete trajectory data. A new filter-and-refine strategy for similarity query is also proposed to filter out the dissimilar trajectories based on Hilbert curve and refine the remaining trajectories with a secure average comparison protocol over the encrypted data. Finally, the exact query results can be obtained through Hilbert curve decoding. Our security analysis demonstrates that both locations and identities of the queried trajectories are preserved from the inference attack, and so does the privacy of the query user’s trajectory. Meanwhile, extensive experimental results show that the proposed scheme can filter out 95% dissimilar trajectories with over 99% average precision, achieving higher query efficiency than the state-of-the-art techniques.

An experiment was conducted on the polyphenol content, flavonoid content, anthocyanin content, and antioxidant capacity of Furong plum (Prunus salicina Lindl. cv. "furong") at different maturity stages to determine the most suitable maturity stage. The inhibition of plum polyphenols on xanthine oxidase (XOD) was measured, and its kinetics were studied to reveal the inhibitory mechanism. The experimental results showed that the polyphenol, flavonoid and anthocyanin contents of plums at the ripe stage were the highest, reaching 320.46 mg GAE/100 g FW, 204.21 mg/100 g FW, and 66.24 mg/100 g FW, respectively, in comparison those of the plums at the immature and mid-ripe stages. The antioxidant capacity of the ripe plums was stronger than it was during the other stages of the plums growth. Among them, the total polyphenols of the ripe plums exhibited the strongest antioxidant capacity (IC50 values against DPPH and hydroxyl radicals were 28.19 ± 0.67 μg/mL and 198.16 ± 7.55 μg/mL, respectively), which was between the antioxidant capacity of the free polyphenols and bound polyphenols. The major phenolic monomer compounds of plum polyphenols were flavan-3-ols (epicatechin, catechin, proanthocyanidin, and procyanidin B2), flavonols (myricetin), and phenolic acids (chlorogenic acid, ferulic acid, and protocatechuic acid). Additionally, plum polyphenols exhibited a strong inhibitory effect on XOD, with an IC50 value of 77.64 μg/mL. The inhibition kinetics showed that plum polyphenols are mixed-type inhibitors that inhibit XOD activity and that the inhibition process is reversible. The calculated values of Ki and α were 16.53 mmol/L and 0.26, respectively.

Learning novel concepts, remembering previous knowledge, and adapting it to future tasks occur simultaneously throughout a human’s lifetime. To model such comprehensive abilities, continual zero-shot learning (CZSL) has recently been introduced. However, most existing methods overused unseen semantic information that may not be continually accessible in realistic settings. In this paper, we address the challenge of continual zero-shot learning where unseen information is not provided during training, by leveraging generative modeling. The heart of the generative-based methods is to learn quality representations from seen classes to improve the generative understanding of the unseen visual space. Motivated by this, we introduce generalization-bound tools and provide the first theoretical explanation for the benefits of generative modeling to CZSL tasks. Guided by the theoretical analysis, we then propose our learning algorithm that employs a novel semantically guided Generative Random Walk (GRW) loss. The GRW loss augments the training by continually encouraging the model to generate realistic and characterized samples to represent the unseen space. Our algorithm achieves state-of-the-art performance on AWA1, AWA2, CUB, and SUN datasets, surpassing existing CZSL methods by 3-7%. The code has been made available here https://github.com/wx-zhang/IGCZSL.

The purpose of this study is to investigate the application of deep neural network for the Automated Underwater Vehicle in the subsea pipeline inspection. In today’s modern world, we see more and more sophisticated and precise computer vision object detection technology being implemented in our daily lives. To name a few, security cameras, self-driving cars, drones and more. This research suggests that computer vision pipeline defect detection is an attractive solution for the future subsea pipeline defect detection as it relies less human intervention and is more reliable. In this paper, review of the current methods of subsea pipeline inspection was done. Some of the methods with visual detection and their limitations are also discussed. Apart from that, the machine learning algorithm of our focus, Faster RCNN is studied. Beyond that, we have explained the methods of our experimentation and the process of training as well as validating the custom dataset. The outcomes are separated into different sections, training curve, visual data representation as well as the model accuracy in different operating environment. In conclusion, we found that the model is able to detect the underwater pipeline with up to 1mm leak size. However, the low accuracy due to the insufficient dataset is recognized as a bottleneck and some of the recommendations are suggested for future improvement.

An image encryption algorithm based on chaotic sequences is presented to make use of its good properties such as ease of generation, sensitive dependence on their initial condition and noise like statistic characteristics. First, the real number value chaotic sequences using the key value is generated. Then it is dispersed into symbol matrix and transformation matrix. Finally the image is encrypted using them in DCT domain. Preliminary results are satisfactory.

Cosmic microwave background radiation (CMB) is critical to the understanding of the early universe and precise estimation of cosmological constants. Due to the contamination of thermal dust noise in the galaxy, the CMB map that is an image on the two-dimensional sphere has missing observations, mainly concentrated on the equatorial region. The noise of the CMB map has a significant impact on the estimation precision for cosmological parameters. Inpainting the CMB map can effectively reduce the uncertainty of parametric estimation. In this paper, we propose a deep learning-based variational autoencoder --- CosmoVAE, to restoring the missing observations of the CMB map. The input and output of CosmoVAE are square images. To generate training, validation, and test data sets, we segment the full-sky CMB map into many small images by Cartesian projection. CosmoVAE assigns physical quantities to the parameters of the VAE network by using the angular power spectrum of the Gaussian random field as latent variables. CosmoVAE adopts a new loss function to improve the learning performance of the model, which consists of $\ell_1$ reconstruction loss, Kullback-Leibler divergence between the posterior distribution of encoder network and the prior distribution of latent variables, perceptual loss, and total-variation regularizer. The proposed model achieves state of the art performance for Planck \texttt{Commander} 2018 CMB map inpainting.

Cellular cryo-Electron Tomography (cryo-ET) provides three-dimensional views of structural and spatial information of various macromolecules in cells in a near-native state. Subtomogram classification is a key step for recognizing and differentiating these macromolecular structures. In recent years, deep learning methods have been developed for high-throughput subtomogram classification tasks; however, conventional supervised deep learning methods cannot recognize macromolecular structural classes that do not exist in the training data. This imposes a major weakness since most native macromolecular structures in cells are unknown and consequently, cannot be included in the training data. Therefore, open set learning which can recognize unknown macromolecular structures is necessary for boosting the power of automatic subtomogram classification. In this paper, we propose a method called Margin-based Loss for Unsupervised Domain Alignment (MLUDA) for open set recognition problems where only a few categories of interest are shared between cross-domain data. Through extensive experiments, we demonstrate that MLUDA performs well at cross-domain open-set classification on both public datasets and medical imaging datasets. So our method is of practical importance.

Cosmic microwave background radiation (CMB) is critical to the understanding of the early universe and precise estimation of cosmological constants. Due to the contamination of thermal dust noise in the galaxy, the CMB map that is an image on the two-dimensional sphere has missing observations, mainly concentrated on the equatorial region. The noise of the CMB map has a significant impact on the estimation precision for cosmological parameters. Inpainting the CMB map can effectively reduce the uncertainty of parametric estimation. In this paper, we propose a deep learning-based variational autoencoder --- CosmoVAE, to restoring the missing observations of the CMB map. The input and output of CosmoVAE are square images. To generate training, validation, and test data sets, we segment the full-sky CMB map into many small images by Cartesian projection. CosmoVAE assigns physical quantities to the parameters of the VAE network by using the angular power spectrum of the Gaussian random field as latent variables. CosmoVAE adopts a new loss function to improve the learning performance of the model, which consists of $\ell_1$ reconstruction loss, Kullback-Leibler divergence between the posterior distribution of encoder network and the prior distribution of latent variables, perceptual loss, and total-variation regularizer. The proposed model achieves state of the art performance for Planck \texttt{Commander} 2018 CMB map inpainting.

TiO2 nanoparticles modified by oleic acid were prepared and dispersed into mineral oil to synthesise TiO2 nanofluids. X-ray diffraction (XRD) and high-resolution transmission electron microscopy (HRTEM) analyses indicated that the TiO2 nanoparticles are single crystals with an average diameter of 6 nm. Measurements on the stability of nanoparticle dispersion in the base oil were performed by testing UV-Vis absorption spectra of nanofluids aged at a working temperature of 80°C for 100 days. In addition, an AC breakdown test was carried out to further evaluate the stability of nanofluids. The obtained results show that the intensity of the UV absorption peak of nanofluid aged for 100 days has no obvious change in comparison with that of the fresh one. In particular, the ratio of the AC breakdown voltage of the aged nanofluid to the fresh one is still up to 98.6%. TiO2 nanofluids exhibit good colloidal stability at the working temperature. Fourier transform infrared spectra (FTIR) and thermogravimetric method (TG) analyses demonstrate that oleic acid anchors on the nanoparticle surface mainly by a chemisorption. This effective surface functionalisation of nanoparticles greatly enhances the stability of TiO2 nanofluids.

Insulating oil modified by nanoparticles (NPs), which is often referred as nanofluids (NFs), has the potential to evolve into substitute of conventional transformer oil for their excellent electrical and thermal characteristics. They have attracted huge attention recently, particularly concerning the improvement of electrical breakdown. This paper develops a relationship that how sizes of NPs affect the positive and negative breakdown voltage. To be more specific, three sized monodisperse Fe <sub xmlns:mml="http://www.w3.org/1998/Math/MathML" xmlns:xlink="http://www.w3.org/1999/xlink">3</sub> O <sub xmlns:mml="http://www.w3.org/1998/Math/MathML" xmlns:xlink="http://www.w3.org/1999/xlink">4</sub> nanoparticles i.e. 10nm, 20nm and 40nm were prepared and subsequently dispersed into insulating mineral oil to develop NFs with 40% W/V concentration. The lightening impulse breakdown strengths of oil samples with and without suspension of NPs were measured in accordance to IEC standard methods. The positive impulse breakdown strength manifested that breakdown strength is first increased up to a maximum value at certain size and then decreased. The results of negative impulse breakdown indicated that breakdown voltage of NFs with different sizes were less than the breakdown voltages of pure transformer oil. Possible modification mechanism of Fe <sub xmlns:mml="http://www.w3.org/1998/Math/MathML" xmlns:xlink="http://www.w3.org/1999/xlink">3</sub> O <sub xmlns:mml="http://www.w3.org/1998/Math/MathML" xmlns:xlink="http://www.w3.org/1999/xlink">4</sub> nanoparticles on insulating properties of transformer oil is also discussed in this paper.

The main question we address in this paper is how to scale up visual recognition of unseen classes, also known as zero-shot learning, to tens of thousands of categories as in the ImageNet-21K benchmark. At this scale, especially with many fine-grained categories included in ImageNet-21K, it is critical to learn quality visual semantic representations that are discriminative enough to recognize unseen classes and distinguish them from seen ones. We propose a \emph{H}ierarchical \emph{G}raphical knowledge \emph{R}epresentation framework for the confidence-based classification method, dubbed as HGR-Net. Our experimental results demonstrate that HGR-Net can grasp class inheritance relations by utilizing hierarchical conceptual knowledge. Our method significantly outperformed all existing techniques, boosting the performance by 7\% compared to the runner-up approach on the ImageNet-21K benchmark. We show that HGR-Net is learning-efficient in few-shot scenarios. We also analyzed our method on smaller datasets like ImageNet-21K-P, 2-hops and 3-hops, demonstrating its generalization ability. Our benchmark and code are available at https://kaiyi.me/p/hgrnet.html.

In distributed or federated optimization and learning, communication between the different computing units is often the bottleneck and gradient compression is widely used to reduce the number of bits sent within each communication round of iterative methods. There are two classes of compression operators and separate algorithms making use of them. In the case of unbiased random compressors with bounded variance (e.g., rand-k), the DIANA algorithm of Mishchenko et al. (2019), which implements a variance reduction technique for handling the variance introduced by compression, is the current state of the art. In the case of biased and contractive compressors (e.g., top-k), the EF21 algorithm of Richt\'arik et al. (2021), which instead implements an error-feedback mechanism, is the current state of the art. These two classes of compression schemes and algorithms are distinct, with different analyses and proof techniques. In this paper, we unify them into a single framework and propose a new algorithm, recovering DIANA and EF21 as particular cases. Our general approach works with a new, larger class of compressors, which has two parameters, the bias and the variance, and includes unbiased and biased compressors as particular cases. This allows us to inherit the best of the two worlds: like EF21 and unlike DIANA, biased compressors, like top-k, whose good performance in practice is recognized, can be used. And like DIANA and unlike EF21, independent randomness at the compressors allows to mitigate the effects of compression, with the convergence rate improving when the number of parallel workers is large. This is the first time that an algorithm with all these features is proposed. We prove its linear convergence under certain conditions. Our approach takes a step towards better understanding of two so-far distinct worlds of communication-efficient distributed learning.

Observation is reported for a structure near the $J/\psi\phi$ threshold in $B^+\rightarrow J/\psi\phi K^+$ decays produced in $\bar{p} p $ collisions at $\sqrt{s}=1.96 \TeV$ with a statistical significance of beyond 5 standard deviations.

The Tevatron experiments have each accumulated about 6 fb−1 good data since the start of RUN II. This large dataset provided good opportunities for meson spectroscopy studies at the Tevatron. This article will cover the recent new Υ(nS) polarization studies as well as exotic meson spectroscopy studies.

Nanofluids were developed by suspending conductive nanoparticles (Fe <sub xmlns:mml="http://www.w3.org/1998/Math/MathML" xmlns:xlink="http://www.w3.org/1999/xlink">3</sub> O <sub xmlns:mml="http://www.w3.org/1998/Math/MathML" xmlns:xlink="http://www.w3.org/1999/xlink">4</sub> ) to improve the dielectric properties of transformer oil. The AC and lightening impulse breakdown voltages were measured for prepared samples in accordance to IEC standards. The results menifested that the addition of conductive nanoparticles (NPs) to the mineral oil can improve the mean AC breakdown performance 1.16 times of that for carrier oil approximately. Additionally, for nanofluids, the mean lightning impulse breakdown voltages were also enhanced than that of base transformer oil and were 1.36 times in comparison to host oil. A possible mechanism of conductive nanoparticles was also used to describe the difference among the performance of nanofluids and base oil.

LBG algorithm is a simple and effective method to train code-book for vector quantization. Since LBG was proposed, several interesting algorithms have been published to improve the effectiveness and efficiency of LBG. Univariate linear representation is another important data compression method, which approximates a target vector by a linear transformation of a selected codeword from codebook. Many applications also use LBG or K-means algorithm to train the codebook of univariate linear representation. In this paper, we propose an improved LBG algorithm called the affine LBG algorithm to train the codebook for univariate linear representation. The experimental results show that the affine LBG algorithm can derive a more effective codebook than LBG algorithm for univariate linear representation. Moreover, the affine LBG algorithm is more efficient than LBG algorithm.

The CMS detector at the CERN Large Hadron Collider features as its innermost component a silicon pixel detector. The original pixel detector was completely replaced during the 2016-2017 winter technical stop. One of the goals of this Phase-1 Upgrade of the pixel detector was to replace the sensors in the original CMS forward pixel detector with new, unirradiated sensors. The new CMS forward pixel detector must survive an integrated luminosity of 300 fb$^{-1}$ before being replaced again prior to the High-Luminosity LHC era. Just as in the original construction, the Phase-1 forward pixel sensors were made of n$^{+}$-in-n Diffusion Oxygenated Float Zone silicon. This note documents the quality spot-checking of the new sensors, comparing our results with those provided by the vendor. In general there was good agreement between the results.

The transparent cornea is the window of the eye, facilitating the entry of light rays and controlling focusing the movement of the light within the eye. The cornea is critical, contributing to 75% of the refractive power of the eye. Keratoconus is a progressive and multifactorial corneal degenerative disease affecting 1 in 2000 individuals worldwide. Currently, there is no cure for keratoconus other than corneal transplantation for advanced stage keratoconus or corneal cross-linking, which can only halt KC progression. The ability to accurately identify subtle KC or KC progression is of vital clinical significance. To date, there has been little consensus on a useful model to classify KC patients, which therefore inhibits the ability to predict disease progression accurately. In this paper, we utilised machine learning to analyse data from 124 KC patients, including topographical and clinical variables. Both supervised multilayer perceptron and unsupervised variational autoencoder models were used to classify KC patients with reference to the existing Amsler-Krumeich (A-K) classification system. Both methods result in high accuracy, with the unsupervised method showing better performance. The result showed that the unsupervised method with a selection of 29 variables could be a powerful tool to provide an automatic classification tool for clinicians. These outcomes provide a platform for additional analysis for the progression and treatment of keratoconus.

The digital media, including text, image, graphics, audio and video etc., has become a main way for information communication along with the popularization of Internet and the development of multimedia techniques. People can get almost information through the Internet. But this gives rise to serious problems including wide spread copyright violation, illegal copying, easy forging etc. How to provide copyright protection and implement covert communication has drawn extensive attention in recent years. As a main method for covert communication and copyright protection(watermarking), information hiding has been widely studied and applied. In this paper, we make an introductory review on the information hiding techniques including the last achievement in this field. First, we give the general concepts and fundamental principles of information hiding such as the definition, characteristics, classification and general framework. Then, we analyze the processing model, the typical methods, the main application and the attack analysis of watermarking. Finally, we make a discussion on some open problems and point out possible directions for further research.

Zero-shot learning (ZSL) aims at understanding unseen categories with no training examples from class-level descriptions. To improve the discriminative power of ZSL, we model the visual learning process of unseen categories with inspiration from the psychology of human creativity for producing novel art. First, we propose CIZSL-v1 as a creativity inspired model for generative ZSL. We relate ZSL to human creativity by observing that ZSL is about recognizing the unseen, and creativity is about creating a likable unseen. We introduce a learning signal inspired by creativity literature that explores the unseen space with hallucinated class-descriptions and encourages careful deviation of their visual feature generations from seen classes while allowing knowledge transfer from seen to unseen classes. Second, CIZSL-v2 is proposed as an improved version of CIZSL-v1 for generative zero-shot learning. CIZSL-v2 consists of an investigation of additional inductive losses for unseen classes along with a semantic guided discriminator. Empirically, we show consistently that CIZSL losses can improve generative ZSL models on the challenging task of generalized ZSL from a noisy text on CUB and NABirds datasets. We also show the advantage of our approach to Attribute-based ZSL on AwA2, aPY, and SUN datasets. We also show that CIZSL-v2 has improved performance compared to CIZSL-v1.

Cryo-electron tomography (Cryo-ET) is a 3D imaging technique that enables the systemic study of shape, abundance, and distribution of macromolecular structures in single cells in near-atomic resolution. However, the systematic and efficient $\textit{de novo}$ recognition and recovery of macromolecular structures captured by Cryo-ET are very challenging due to the structural complexity and imaging limits. Even macromolecules with identical structures have various appearances due to different orientations and imaging limits, such as noise and the missing wedge effect. Explicitly disentangling the semantic features of macromolecules is crucial for performing several downstream analyses on the macromolecules. This paper has addressed the problem by proposing a 3D Spatial Variational Autoencoder that explicitly disentangle the structure, orientation, and shift of macromolecules. Extensive experiments on both synthesized and real cryo-ET datasets and cross-domain evaluations demonstrate the efficacy of our method.

This proceeding summarizes the search for new physics via a rare heavy meson decays– Bs → µ+µ−, the observation of an unexpected structure in the J/ψφ spectrum through exclusive B+ → J/ψφK+ decays, as well as the measurement of X(3872) production cross section using the pp collision data collected at the CMS experiment.

Urbanization and rapid industrial development in recent years have created a major threat to the environment especially water pollution. Pollution from wastewater not only depreciates land values; it also increases municipal costs for wastewater treatment and causes harm to biological and human health. Industries such as ceramic, printing, plastic and paper use dyes in their coloring process and these dyes are usually disposed to streams, ponds, lakes and river which then cause water pollution. Not only that the dyes can be very toxic even at low concentration, they are also generally non-biodegradable and difficult to be removed using conventional biological treatment. Many researches have been done on ways to treat wastewater effectively namely membrane separation, aerobic and anaerobic degradation using various microorganisms, chemical oxidation, coagulation and flocculation, adsorption using various kinds of adsorbents and reverse osmosis. However, most of the current wastewater treatment techniques are selective and expensive. Adsorption process has been identified as the most feasible wastewater treatment technique as it is cheap economically, simplicity in design and has the ability to adsorb a wide range of both organic and inorganic pollutants. With that, this project aims to study methylene blue dye removal using functionalized granular activated carbon.

The digital media, including text, image, graphics, audio and video etc., has become a main way for information communication along with the popularization of Internet and the development of multimedia techniques. People can get almost information through the Internet. But this gives rise to serious problems including wide spread copyright violation, illegal copying, easy forging etc. How to provide copyright protection has drawn extensive attention in recent years. As a main method for copyright protection, the techniques of watermarking, have been widely studied and applied. In this paper, we proposes an adaptive 2 dimension image watermarking algorithm in space domain with gray image's watermark. In order to embed the image's watermark, we split the image's watermark into blocks and transform them into DCT domain. Then quantize their DCT coefficients and adjust them. Finally, take out the parts of DCT coefficients from each block to constitute the watermark. By using the human visual system, we split the original image into blocks and classify them. According to the classification, the watermark components with different strength are embedded into cover image. The experimental results show our algorithm has very excellent effect. Finally, we make a discussion on some open problems and point out possible directions for further research.

A new model and resultant equation for the coagulation of acrylonitrile monomers in precipitation polymerization are suggested in consideration of the surface tension (<TEX>${\gamma}$</TEX>) and cohesive energy density (<TEX>$E_{CED}$</TEX>). The equation was proven to be quite favorable by considering figure fittings from known surface tensions and cohesive energy densities of certain organic solvents. The relationship between scale value of surface tension (<TEX>${\gamma}$</TEX>/M) and cohesive energy density of monomers can be obtained by changing the coagulation bath component for effective precipitation polymerization of acrylonitrile in wet spinning.

As a new protection methord of bird's dropping flashover in the overhead power transmission line,The anti-flashover insulation sheath takes effect by reducing the bearing voltage of the insulator.It is feasible theoretically,but the practical effects are not verified yet.So by analyzing the course of the bird's droppings flashover and establishing the simulation model and testing platform for bird's dropping flashover of the overhead power transmission line,verifies the protective effect of the anti-flashover insulation sheath for the bird's dropping flashover.The result shows that when its thickness is enough,the insulation sheath can prevent the bird's droppings flashover the overhead power transmission line effectively.

Landfill Leachate Treatment of wastewater is a difficult.By Rongjiang County of Guizhou Province handling solid waste sanitary landfill leachate collection and treatment system project designed to introduce,the nature and treatment of landfill leachate process was described and summarized.Focus on Pretreatment+MBR+RO process in the treatment of landfill leachate and advantages.