The recently discovered titanium-based kagome metal ATi_(3)Bi_(5)(A=Cs,Rb)provides a new platform to explore novel quantum phenomena.In this work,the transport properties of ATi_(3)Bi_(5)(A=Cs,Rb)are systematically in...The recently discovered titanium-based kagome metal ATi_(3)Bi_(5)(A=Cs,Rb)provides a new platform to explore novel quantum phenomena.In this work,the transport properties of ATi_(3)Bi_(5)(A=Cs,Rb)are systematically investigated under high pressure.Although ATi_(3)Bi_(5)(A=Cs,Rb)shows no evidence of superconductivity at ambient pressure,the pressure-induced double-dome superconductivity is observed in both compounds,resembling the superconducting phase diagram of AV_(3)Sb_(5)(A=Cs,Rb,and K)under pressure.High-pressure X-ray difraction measurements exclude the pressure-induced structural phase transition.A slope change in the c/a ratio was found between 12.4 and 14.9 GPa,indicating the occurrence of lattice distortion.The distinct changes in the electronic band structure revealed by frst-principles calculations further explain the emergence of superconductivity in the two domes.These fndings suggest that pressure can efectively tune the electronic properties of ATi_(3)Bi_(5),providing new insights into the rich physics of kagome metals.展开更多
THE well-known ancient Chinese philosopher Lao Tzu(老子)or Laozi(6th~4th century BC during the Spring and Autumn period)started his classic Tao Teh Ching《道德经》or Dao De Jing(see Fig.1)with six Chinese characters:&...THE well-known ancient Chinese philosopher Lao Tzu(老子)or Laozi(6th~4th century BC during the Spring and Autumn period)started his classic Tao Teh Ching《道德经》or Dao De Jing(see Fig.1)with six Chinese characters:"道(Dao)可(Ke)道(Dao)非(Fei)常(Chang)道(Dao)",which has been traditionally interpreted as“道可道,非常道”or"The Dao that can be spoken is not the eternal Dao".展开更多
Atmospheric particulate matter pollution has attracted much wider attention globally.In recent years,the development of atmospheric particle collection techniques has put forwards new demands on the real-time source a...Atmospheric particulate matter pollution has attracted much wider attention globally.In recent years,the development of atmospheric particle collection techniques has put forwards new demands on the real-time source apportionments techniques.Such demands are summarized,in this paper,as how to set up new restraints in apportionment and how to develop a non-linear regression model to process complicated circumstances,such as the existence of secondary source and similar source.In this study,we firstly analyze the possible and potential restraints in single particle source apportionment,then propose a novel three-step self-feedback long short-term memory(SF-LSTM)network for approximating the source contribution.The proposed deep learning neural network includes three modules,as generation,scoring and refining,and regeneration modules.Benefited from the scoring modules,SF-LSTM implants four loss functions representing four restraints to be followed in the apportionment,meanwhile,the regeneration module calculates the source contribution in a non-linear way.The results show that the model outperforms the conventional regression methods in the overall performance of the four evaluation indicators(residual sum of squares,stability,sparsity,negativity)for the restraints.Additionally,in short time-resolution analyzing,SF-LSTM provides better results under the restraint of stability.展开更多
BACKGROUND Understanding a virus shedding patterns in body fluids/secretions is importantto determine the samples to be used for diagnosis and to formulate infectioncontrol measures.AIM To investigate the severe acute...BACKGROUND Understanding a virus shedding patterns in body fluids/secretions is importantto determine the samples to be used for diagnosis and to formulate infectioncontrol measures.AIM To investigate the severe acute respiratory syndrome coronavirus 2 (SARS-CoV-2)shedding patterns and its risk factors.METHODS All laboratory-confirmed coronavirus disease 2019 patients with completemedical records admitted to the Shenzhen Third People’s Hospital from January28, 2020 to March 8, 2020 were included. Among 145 patients (54.5% males;median age, 46.1 years), three (2.1%) died. The bronco-alveolar lavage fluid(BALF) had the highest virus load compared with the other samples. The viralload peaked at admission (3.3 × 108 copies) and sharply decreased 10 d afteradmission.RESULTS The viral load was associated with prolonged intensive care unit (ICU) duration.Patients in the ICU had significantly longer shedding time compared to those inthe wards (P < 0.0001). Age > 60 years [hazard ratio (HR) = 0.6;95% confidenceinterval (CI): 0.4-0.9] was an independent risk factor for SARS-CoV-2 shedding,while chloroquine (HR = 22.8;95%CI: 2.3-224.6) was a protective factor.CONCLUSION BALF had the highest SARS-CoV-2 load. Elderly patients had higher virus loads,which was associated with a prolonged ICU stay. Chloroquine was associatedwith shorter shedding duration and increased the chance of viral negativity.展开更多
The effects of various contaminants in the electrolytic refinement of indium were investigated using a glow discharge mass spectrometer(GDMS).The effects of several factors such as the indium ion(In3+)concentration,th...The effects of various contaminants in the electrolytic refinement of indium were investigated using a glow discharge mass spectrometer(GDMS).The effects of several factors such as the indium ion(In3+)concentration,the sodium chloride(NaCl)concentration,the current density,the gelatin concentration,the pH,and the electrode distance,were examined.Significant variations in impurity levels concerning gelatin concentration were observed.Both the gelatin and In3+concentration were moderately positively correlated with the Pb content.The Sb concentration was associated positively with the NaCl concentration,while the Ti concentration had an adverse correlation with the NaCl concentration.The Bi element content was positively linked to the electrode distance.As the current density increased,Cu,Pb,and Bi impurities initially rose and then eventually declined.Notably,a critical current density of 45 A·m^(-2) was identified in this behavior.展开更多
Fine particulate matter(PM_(2.5))and ozone(O_(3))pollutions are prevalent air quality issues in China.Volatile organic compounds(VOCs)have significant impact on the formation of O_(3)and secondary organic aerosols(SOA...Fine particulate matter(PM_(2.5))and ozone(O_(3))pollutions are prevalent air quality issues in China.Volatile organic compounds(VOCs)have significant impact on the formation of O_(3)and secondary organic aerosols(SOA)contributing PM_(2.5).Herein,we investigated 54 VOCs,O_(3)and SOA in Tianjin from June 2017 to May 2019 to explore the non-linear relationship among O_(3),SOA and VOCs.The monthly patterns of VOCs and SOA concentrations were characterized by peak values during October to March and reached a minimum from April to September,but the observed O_(3)was exactly the opposite.Machine learning methods resolved the importance of individual VOCs on O_(3)and SOA that alkenes(mainly ethylene,propylene,and isoprene)have the highest importance to O_(3)formation;alkanes(C_(n),n≥6)and aromatics were the main source of SOA formation.Machine learning methods revealed and emphasized the importance of photochemical consumptions of VOCs to O_(3)and SOA formation.Ozone formation potential(OFP)and secondary organic aerosol formation potential(SOAFP)calculated by consumed VOCs quantitatively indicated that more than 80%of the consumed VOCs were alkenes which dominated the O_(3)formation,and the importance of consumed aromatics and alkenes to SOAFP were 40.84%and 56.65%,respectively.Therein,isoprene contributed the most to OFP at 41.45%regardless of the season,while aromatics(58.27%)contributed the most to SOAFP in winter.Collectively,our findings can provide scientific evidence on policymaking for VOCs controls on seasonal scales to achieve effective reduction in both SOA and O_(3).展开更多
The activation of CO on iron-based materials is a key elementary reaction for many chemical processes.We investigate CO adsorption and dissociation on a series of Fe,Fe_(3)C,Fe_(5)C_(2),and Fe_(2)C catalysts through d...The activation of CO on iron-based materials is a key elementary reaction for many chemical processes.We investigate CO adsorption and dissociation on a series of Fe,Fe_(3)C,Fe_(5)C_(2),and Fe_(2)C catalysts through density functional theory calculations.We detect dramatically different performances for CO adsorption and activation on diverse surfaces and sites.The activation of CO is dependent on the local coordination of the molecule to the surface and on the bulk phase of the underlying catalyst.The bulk properties and the different local bonding environments lead to varying interactions between the adsorbed CO and the surface and thus yielding different activation levels of the C-O bond.We also examine the prediction of CO adsorption on different types of Fe-based catalysts by machine learning through linear regression models.We combine the features originating from surfaces and bulk phases to enhance the prediction of the activation energies and perform eight different linear regressions utilizing the feature engineering of polynomial representations.Among them,a ridge linear regression model with2nd-degree polynomial feature generation predicted the best CO activation energy with a mean absolute error of 0.269 eV.展开更多
The Gaia DR3 parallax approach was used to estimate the absolute parameters of 2375δScuti stars from the ASAS catalog.The selected stars have a variety of observational characteristics,with a higher than 80%probabili...The Gaia DR3 parallax approach was used to estimate the absolute parameters of 2375δScuti stars from the ASAS catalog.The selected stars have a variety of observational characteristics,with a higher than 80%probability of beingδScuti stars.We have displayed all the stars in the Hertzsprung-Russell diagram along with theδScuti instability strip,the Zero Age Main Sequence and the Terminal Age Main Sequence.Then,we determined which fundamental and overtone modes each star belongs to using pulsation constant(Q)calculations.In addition,we evaluated the parameters in the Q calculation equation using three machine learning methods,which showed that surface gravity and temperature have the greatest effect on its calculation.The Period-Luminosity(P-L)relationship of theδScuti stars was also revisited.Eventually,using least squares linear regression,we made four linear fits for fundamental and overtone modes and updated their relationships.展开更多
AIM:To establish a meaningful standard for diagnosing ocular metastasis(OM)in menopausal breast cancer(BC)women,and explore the relationship between CA-153,CA-125,apolipoprotein A,and OM.METHODS:A total of 1362 menopa...AIM:To establish a meaningful standard for diagnosing ocular metastasis(OM)in menopausal breast cancer(BC)women,and explore the relationship between CA-153,CA-125,apolipoprotein A,and OM.METHODS:A total of 1362 menopausal female BC patients with OM volunteered to take part in this study between July 2012 and July 2022.Women with BC who are menopausal were found to have an OM incidence of 1.6%.Furthermore,CA-153,CA-125,and apolipoprotein A(Apo A)all contributed to OM in women with BC who are postmenopausal according to binary logistic regression.Receiver operating curve(ROC)analysis was used to assess the diagnostic value of OM in patients with BC.RESULTS:Both CA-153 and CA-153+CA-125 showed a higher sensitivity of 95.45%,whereas CA-153+Apo A illustrated the highest specificity of 99.02%.Moreover,CA-153 and CA-153+CA-125 had higher areas under the curve(AUC)of 0.973.CONCLUSION:The data indicate that the serum concentrations of CA-153 exhibited the most significant predictors of the diagnosis of OM in menopausal women with BC.The current study researches the utility of risk factors in predicting of OM in menopausal BC women and put forward the latest suggestions on their clinical application.展开更多
Large models have been widely used in the field of neural language processing,information retrieving,etc.With the development of the large models,not only is the parameter scale increased,but the model architecture ha...Large models have been widely used in the field of neural language processing,information retrieving,etc.With the development of the large models,not only is the parameter scale increased,but the model architecture has also become more complex.For example,the multi-modal transformer-based model mainly has concurrent branches,which we denoted as the concurrent branch model(CBM).Many CBMs have enlarged to tens of billions of parameters,and require distributed resources to train this kind of model.Existing distributed training systems cannot fully handle this type of model architecture because there are interactions between branches.Inspired by the unbalanced resource usage of pipeline parallelism,we prefer to organize different branches with a fine-grained bidirectional pipeline schedule of communication and computation.However,improper coordination between branches leads to idle time for computation and low training efficiency.In this paper,we present Flexpipe,a pipeline engine for c3oncurrent-branch models.We first introduce a branch-aware pipeline parallelism(BAPP)to make full use of the concurrent characteristic of the model architecture.Then,based on a multi-branch pipeline simulator,we propose an adaptive interaction coordinator,which facilitates the low-overhead branch interactions during the distributed model training.We evaluate our approach on popular concurrent branch models combined with modern training systems.Compared with the Chimera,the experiential results show that our method improves the end-to-end training throughput by 20%on average.展开更多
Transformer-based large language models(LLMs)have made significant strides in the field of artificial intelligence(AI).However,training these LLMs imposes immense demands on computational power and bandwidth for hardw...Transformer-based large language models(LLMs)have made significant strides in the field of artificial intelligence(AI).However,training these LLMs imposes immense demands on computational power and bandwidth for hardware systems.Wafer-scale chips(WSCs)offer a promising solution,yet they struggle with limited on-chip memory and complex tensor partitioning.To fully harness the high-bandwidth,low-latency on-chip interconnect benefits of WSCs and to alleviate the on-chip memory limitations,a specialized mapping and architecture co-exploration method is essential.Despite existing efforts in memory optimization and mapping,current approaches fall short for WSC scenarios.To bridge this gap,we introduce TMAC,an architecture-mapping co-exploration framework that integrates recomputation into the design space,fully exploiting optimization opportunities overlooked by existing works.Further,TMAC takes advantage of the superior on-chip interconnect performance of WSCs by incorporating a more flexible tensor partition scheme.TMAC then introduces a novel operator-centric encoding scheme(OCES)designed to comprehensively describe the mapping space for training LLMs.Unlike previous studies that focus solely on communication volume analysis based on mapping,TMAC explores the design space by evaluating the combined impact of mapping and architecture on training performance.However,fully accounting for these untapped optimization opportunities increases the complexity of the design space.To address this,we streamline the simulation process,reducing the time needed for exploration.Compared to AccPar,Deepspeed and Megatron,TMAC delivers a 3.1×,2.9×,1.6×performance gain.In terms of memory usage,TMAC requires 3.6×,3.1×less memory than AccPar and Deepspeed,respectively and is comparable to Megatron’s full recomputation method.展开更多
Chain-walking has emerged as a promising strategy in organic synthetic methodology,but achieving site-selectivity in reactions involving competition among multiple potential positions remains a challenge.In this study...Chain-walking has emerged as a promising strategy in organic synthetic methodology,but achieving site-selectivity in reactions involving competition among multiple potential positions remains a challenge.In this study,we presented a novel approach to ligand-modulated,nickel-catalyzed regiodivergent alkenylboration of allylarenes.Our method allows for highly chemoselective preparation of two classes of structurally different alkenyl boronates from the same starting by simply switching the ligand.Mechanistic investigations involving experiments and calculations suggest that the ligand-modulating regiodivergence arises from either regioselectiveβ-H elimination or oxidative addition with organic halides.Our findings offer valuable insight into addressing site-selective issues in chain-walking reactions with multiple thermodynamically stable factors.展开更多
Real-world blind image super-resolution is a challenging problem due to the absence of target high resolution images for training.Inspired by the recent success of the single image generation based method SinGAN,we ta...Real-world blind image super-resolution is a challenging problem due to the absence of target high resolution images for training.Inspired by the recent success of the single image generation based method SinGAN,we tackle this challenging problem with a refined model SR-SinGAN,which can learn to perform single real image super-resolution.Firstly,we empirically find that downsampled LR input with an appropriate size can improve the robustness of the generation model.Secondly,we introduce a global contextual prior to provide semantic information.This helps to remove distorted pixels and improve the output fidelity.Finally,we design an image gradient based local contextual prior to guide detail generation.It can alleviate generated artifacts in smooth areas while preserving rich details in densely textured regions(e.g.,hair,grass).To evaluate the effectiveness of these contextual priors,we conducted extensive experiments on both artificial and real images.Results show that these priors can stabilize training and preserve output fidelity,improving the generated image quality.We furthermore find that these single image generation based methods work better for images with repeated textures compared to general images.展开更多
Shadows in document images are undesirable yet inevitable.They can decrease the clarity and readability of the images.The existing methods for removing shadows from documents still face some challenges,such as the tra...Shadows in document images are undesirable yet inevitable.They can decrease the clarity and readability of the images.The existing methods for removing shadows from documents still face some challenges,such as the traditional heuristics lack universality and the optimization goal of subnetworks is not consistent for multistage deep learning methods.In this paper,we introduce an end-to-end direct document shadow removal network(DDSR-Net),where we employ a 3-layer UNet++as the backbone to extract features from diverse scales.To further improve the performance of DDSR-Net,we integrate the multi-scale attention(MSA)blocks into each node.The MSA block allocates different weights to feature vectors at different positions,achieving automatic feature alignment and significantly enhancing the end-to-end network's ability to handle shadow processing.To verify the effectiveness of the proposed DDSR-Net,qualitative and quantitative experiments are conducted on multiple open-source document shadow removal datasets.The experimental results demonstrate that our method outperforms the existing state-of-the-art methods on these datasets.Our code and models will be released to the public.展开更多
1 Introduction Endoscopy plays a crucial role in the diagnoses and treatment of gastrointestinal(GI)diseases[1],as it helps to identify abnormalities,classify lesion,and determine treatment methods.During GI endoscopi...1 Introduction Endoscopy plays a crucial role in the diagnoses and treatment of gastrointestinal(GI)diseases[1],as it helps to identify abnormalities,classify lesion,and determine treatment methods.During GI endoscopic examinations,physicians may encounter practical hindrances,i.e.,fatigue,stress,or limited experience,which can lead to erroneous results.Artificial intelligence(AI)-assisted GI endoscopy technology has emerged to address these limitations[2].展开更多
In this paper,a cross-sensor generative self-supervised learning network is proposed for fault detection of multi-sensor.By modeling the sensor signals in multiple dimensions to achieve correlation information mining ...In this paper,a cross-sensor generative self-supervised learning network is proposed for fault detection of multi-sensor.By modeling the sensor signals in multiple dimensions to achieve correlation information mining between channels to deal with the pretext task,the shared features between multi-sensor data can be captured,and the gap between channel data features will be reduced.Meanwhile,in order to model fault features in the downstream task,the salience module is developed to optimize cross-sensor data features based on a small amount of labeled data to make warning feature information prominent for improving the separator accuracy.Finally,experimental results on the public datasets FEMTO-ST dataset and the private datasets SMT shock absorber dataset(SMT-SA dataset)show that the proposed method performs favorably against other STATE-of-the-art methods.展开更多
Introduction:Seasonal influenza poses a significant public health burden,causing substantial morbidity and mortality worldwide each year.In this context,timely and accurate vaccine strain selection is critical to miti...Introduction:Seasonal influenza poses a significant public health burden,causing substantial morbidity and mortality worldwide each year.In this context,timely and accurate vaccine strain selection is critical to mitigating the impact of influenza outbreaks.This article aims to develop an adaptive,universal,and convenient method for predicting antigenic variation in influenza A(H1N1),thereby providing a scientific basis to enhance the biannual influenza vaccine selection process.Methods:The study integrates adaptive Fourier decomposition(AFD)theory with multiple techniques—including matching pursuit,the maximum selection principle,and bootstrapping—to investigate the complex nonlinear interactions between amino acid substitutions in hemagglutinin(HA)proteins(the primary antigenic protein of influenza virus)and their impact on antigenic changes.Results:Through comparative analysis with classical methods such as Lasso,Ridge,and random forest,we demonstrate that the AFD-type method offers superior accuracy and computational efficiency in identifying antigenic change-associated amino acid substitutions,thus eliminating the need for timeconsuming and expensive experimental procedures.AAW Conclusion:In summary,AFD-based methods represent effective mathematical models for predicting antigenic variations based on HA sequences and serological data,functioning as ensemble algorithms with guaranteed convergence.Following the sequence of indicators specified in I,we perform a series of operations on A_(1),including feature extension,extraction,and rearrangement,to generate a new input dataset for the prediction step.With this newly prepared input,we can compute the predicted results as.展开更多
The emergence of nonconventional luminescent materials(NLMs)has attracted significant attention due to their sustainable synthesis and tunable optical properties.Yet,establishing a clear structure-emission relationshi...The emergence of nonconventional luminescent materials(NLMs)has attracted significant attention due to their sustainable synthesis and tunable optical properties.Yet,establishing a clear structure-emission relationship remains a challenge.In this work,we report a previously unknown class of NLMs:cross-linked protein crystals that exhibit intense photoluminescence(PL)in the visible range(425-680 nm).We systematically investigated seven natural protein crystals(concanavalin,catalase,lysozyme,hemoglobin,α-chymotrypsin,pepsin,and β-lactoglobulin)cross-linked with glutaraldehyde and demonstrated that cross-linking induces broadband emission that is absent in natural crystals.Focusing on polymorphic lysozyme crystals(tetragonal,orthorhombic,and monoclinic),we found excitation-dependent fluorescence with lifetimes in the nanosecond range and quantum yields up to 20%(in themonoclinic phase under 450 nmexcitation).Single-and two-photon spectroscopy,as well as pressure-and solvent-modulated PL studies,confirm that the emission is due to intermolecular through-space interactions(TSI)within the crystal lattice.Compression enhances TSI and redshifts the emission,whereas the solvent(DMSO)-induced swelling reduces TSI and causes a blue shift,establishing a direct structure-emission correlation.This work establishes protein crystals as programmableNLMswith tunable emission and provides a mechanistic framework for the design of nonconventional luminogens through protein crystal engineering.展开更多
Transformers have recently lead to encouraging progress in computer vision.In this work,we present new baselines by improving the original Pyramid Vision Transformer(PVT v1)by adding three designs:(i)a linear complexi...Transformers have recently lead to encouraging progress in computer vision.In this work,we present new baselines by improving the original Pyramid Vision Transformer(PVT v1)by adding three designs:(i)a linear complexity attention layer,(ii)an overlapping patch embedding,and(iii)a convolutional feed-forward network.With these modifications,PVT v2 reduces the computational complexity of PVT v1 to linearity and provides significant improvements on fundamental vision tasks such as classification,detection,and segmentation.In particular,PVT v2 achieves comparable or better performance than recent work such as the Swin transformer.We hope this work will facilitate state-ofthe-art transformer research in computer vision.Code is available at https://github.com/whai362/PVT.展开更多
基金supported by the Natural Science Foundation of China(Grant No.12174064)the Shanghai Municipal Science and Technology Major Project(Grant No.2019SHZDZX01)+5 种基金the Innovation Program for Quantum Science and Technology(Grant No.2024ZD0300104)the support by the Natural Science Foundation of China(Grant No.12204383)the Young Elite Scientists Sponsorship Program by CAST(Grant No.2023QNRC001)the Young Talent Fund of the Association for Science and Technology in Shaanxi(Grant No.CLGC202201)supported by the open project of Beijing National Laboratory for Condensed Matter Physics(Grant No.ZBJ2106110017)the Double First-Class Initiative Fund of Shanghai Tech University。
文摘The recently discovered titanium-based kagome metal ATi_(3)Bi_(5)(A=Cs,Rb)provides a new platform to explore novel quantum phenomena.In this work,the transport properties of ATi_(3)Bi_(5)(A=Cs,Rb)are systematically investigated under high pressure.Although ATi_(3)Bi_(5)(A=Cs,Rb)shows no evidence of superconductivity at ambient pressure,the pressure-induced double-dome superconductivity is observed in both compounds,resembling the superconducting phase diagram of AV_(3)Sb_(5)(A=Cs,Rb,and K)under pressure.High-pressure X-ray difraction measurements exclude the pressure-induced structural phase transition.A slope change in the c/a ratio was found between 12.4 and 14.9 GPa,indicating the occurrence of lattice distortion.The distinct changes in the electronic band structure revealed by frst-principles calculations further explain the emergence of superconductivity in the two domes.These fndings suggest that pressure can efectively tune the electronic properties of ATi_(3)Bi_(5),providing new insights into the rich physics of kagome metals.
基金partially supported by the National Key R&D Program of China(2020YFB2104001)the National Natural Science Foundation of China(62271485,61903363,62203250,U1811463)。
文摘THE well-known ancient Chinese philosopher Lao Tzu(老子)or Laozi(6th~4th century BC during the Spring and Autumn period)started his classic Tao Teh Ching《道德经》or Dao De Jing(see Fig.1)with six Chinese characters:"道(Dao)可(Ke)道(Dao)非(Fei)常(Chang)道(Dao)",which has been traditionally interpreted as“道可道,非常道”or"The Dao that can be spoken is not the eternal Dao".
基金supported by the National Key Research and Development Program of China[grant number 2020YFA0608000]the National Natural Science Foundation of China[grant number 42030605].
基金supported by Key Laboratory For Environmental Factors Control of Agro-product Quality Safety,Ministry of Agriculture and Rural Affairs(No.2018hjyzkfkt-002)Qian Xuesen Laboratory of Space Technology,CAST(No.GZZKFJJ2020002)National Research Program for Key Issues in Air Pollution Control(No.DQGG-05-30)
文摘Atmospheric particulate matter pollution has attracted much wider attention globally.In recent years,the development of atmospheric particle collection techniques has put forwards new demands on the real-time source apportionments techniques.Such demands are summarized,in this paper,as how to set up new restraints in apportionment and how to develop a non-linear regression model to process complicated circumstances,such as the existence of secondary source and similar source.In this study,we firstly analyze the possible and potential restraints in single particle source apportionment,then propose a novel three-step self-feedback long short-term memory(SF-LSTM)network for approximating the source contribution.The proposed deep learning neural network includes three modules,as generation,scoring and refining,and regeneration modules.Benefited from the scoring modules,SF-LSTM implants four loss functions representing four restraints to be followed in the apportionment,meanwhile,the regeneration module calculates the source contribution in a non-linear way.The results show that the model outperforms the conventional regression methods in the overall performance of the four evaluation indicators(residual sum of squares,stability,sparsity,negativity)for the restraints.Additionally,in short time-resolution analyzing,SF-LSTM provides better results under the restraint of stability.
基金Supported by Startup Fund forYouth Faculty of ShenzhenUniversity, No. 2018009.
文摘BACKGROUND Understanding a virus shedding patterns in body fluids/secretions is importantto determine the samples to be used for diagnosis and to formulate infectioncontrol measures.AIM To investigate the severe acute respiratory syndrome coronavirus 2 (SARS-CoV-2)shedding patterns and its risk factors.METHODS All laboratory-confirmed coronavirus disease 2019 patients with completemedical records admitted to the Shenzhen Third People’s Hospital from January28, 2020 to March 8, 2020 were included. Among 145 patients (54.5% males;median age, 46.1 years), three (2.1%) died. The bronco-alveolar lavage fluid(BALF) had the highest virus load compared with the other samples. The viralload peaked at admission (3.3 × 108 copies) and sharply decreased 10 d afteradmission.RESULTS The viral load was associated with prolonged intensive care unit (ICU) duration.Patients in the ICU had significantly longer shedding time compared to those inthe wards (P < 0.0001). Age > 60 years [hazard ratio (HR) = 0.6;95% confidenceinterval (CI): 0.4-0.9] was an independent risk factor for SARS-CoV-2 shedding,while chloroquine (HR = 22.8;95%CI: 2.3-224.6) was a protective factor.CONCLUSION BALF had the highest SARS-CoV-2 load. Elderly patients had higher virus loads,which was associated with a prolonged ICU stay. Chloroquine was associatedwith shorter shedding duration and increased the chance of viral negativity.
基金supported by the National Natural Science Foundation of China(52074180)the Science and Technology Major Project of Yunnan Province(202302AB080020)+2 种基金the Independent Research Project of State Key Laboratory of Advanced Special Steel,Shanghai Key Laboratory of Advanced Ferrometallurgy,Shanghai University(SKLASS 2023-Z07)the Science and Technology Commission of Shanghai Municipality(19DZ2270200)the Program for Professor of Special Appointment(Eastern Scholar)at SIHL,Shanghai Sailing Program(19YF1416500).
文摘The effects of various contaminants in the electrolytic refinement of indium were investigated using a glow discharge mass spectrometer(GDMS).The effects of several factors such as the indium ion(In3+)concentration,the sodium chloride(NaCl)concentration,the current density,the gelatin concentration,the pH,and the electrode distance,were examined.Significant variations in impurity levels concerning gelatin concentration were observed.Both the gelatin and In3+concentration were moderately positively correlated with the Pb content.The Sb concentration was associated positively with the NaCl concentration,while the Ti concentration had an adverse correlation with the NaCl concentration.The Bi element content was positively linked to the electrode distance.As the current density increased,Cu,Pb,and Bi impurities initially rose and then eventually declined.Notably,a critical current density of 45 A·m^(-2) was identified in this behavior.
基金financially supported by the National Key Research and Development Program of China(No.2018 YFE0106900)supported by National Natural Science Foundation of China(Nos.42077191,41775149)+2 种基金Fundamental Research Funds for the Central Universities(No.63213072)National Research Program for Key Issues in Air Pollution Control(No.DQGG-05-30)the Blue Sky Foundation
文摘Fine particulate matter(PM_(2.5))and ozone(O_(3))pollutions are prevalent air quality issues in China.Volatile organic compounds(VOCs)have significant impact on the formation of O_(3)and secondary organic aerosols(SOA)contributing PM_(2.5).Herein,we investigated 54 VOCs,O_(3)and SOA in Tianjin from June 2017 to May 2019 to explore the non-linear relationship among O_(3),SOA and VOCs.The monthly patterns of VOCs and SOA concentrations were characterized by peak values during October to March and reached a minimum from April to September,but the observed O_(3)was exactly the opposite.Machine learning methods resolved the importance of individual VOCs on O_(3)and SOA that alkenes(mainly ethylene,propylene,and isoprene)have the highest importance to O_(3)formation;alkanes(C_(n),n≥6)and aromatics were the main source of SOA formation.Machine learning methods revealed and emphasized the importance of photochemical consumptions of VOCs to O_(3)and SOA formation.Ozone formation potential(OFP)and secondary organic aerosol formation potential(SOAFP)calculated by consumed VOCs quantitatively indicated that more than 80%of the consumed VOCs were alkenes which dominated the O_(3)formation,and the importance of consumed aromatics and alkenes to SOAFP were 40.84%and 56.65%,respectively.Therein,isoprene contributed the most to OFP at 41.45%regardless of the season,while aromatics(58.27%)contributed the most to SOAFP in winter.Collectively,our findings can provide scientific evidence on policymaking for VOCs controls on seasonal scales to achieve effective reduction in both SOA and O_(3).
基金financially supported from the National Natural Science Foundation of China (No.22002008)Ningxia Key Research and Development Project,China (Nos.2022BEE03002 and 2022BSB03056)funding support from Synfuels China,Co.,Ltd.and Beijing Advanced Innovation Center for Materials Genome Engineering。
文摘The activation of CO on iron-based materials is a key elementary reaction for many chemical processes.We investigate CO adsorption and dissociation on a series of Fe,Fe_(3)C,Fe_(5)C_(2),and Fe_(2)C catalysts through density functional theory calculations.We detect dramatically different performances for CO adsorption and activation on diverse surfaces and sites.The activation of CO is dependent on the local coordination of the molecule to the surface and on the bulk phase of the underlying catalyst.The bulk properties and the different local bonding environments lead to varying interactions between the adsorbed CO and the surface and thus yielding different activation levels of the C-O bond.We also examine the prediction of CO adsorption on different types of Fe-based catalysts by machine learning through linear regression models.We combine the features originating from surfaces and bulk phases to enhance the prediction of the activation energies and perform eight different linear regressions utilizing the feature engineering of polynomial representations.Among them,a ridge linear regression model with2nd-degree polynomial feature generation predicted the best CO activation energy with a mean absolute error of 0.269 eV.
文摘The Gaia DR3 parallax approach was used to estimate the absolute parameters of 2375δScuti stars from the ASAS catalog.The selected stars have a variety of observational characteristics,with a higher than 80%probability of beingδScuti stars.We have displayed all the stars in the Hertzsprung-Russell diagram along with theδScuti instability strip,the Zero Age Main Sequence and the Terminal Age Main Sequence.Then,we determined which fundamental and overtone modes each star belongs to using pulsation constant(Q)calculations.In addition,we evaluated the parameters in the Q calculation equation using three machine learning methods,which showed that surface gravity and temperature have the greatest effect on its calculation.The Period-Luminosity(P-L)relationship of theδScuti stars was also revisited.Eventually,using least squares linear regression,we made four linear fits for fundamental and overtone modes and updated their relationships.
基金Supported by National Natural Science Foundation of China(No.82160195,No.82460203)Jiangxi Key R&D Program of Jiangxi Province(No.20223BBH80014)+1 种基金Science and Technology Project of Jiangxi Province Health Commission of Traditional Chinese Medicine(No.2022B258)Science and Technology Project of Jiangxi Health Commission(No.202210017).
文摘AIM:To establish a meaningful standard for diagnosing ocular metastasis(OM)in menopausal breast cancer(BC)women,and explore the relationship between CA-153,CA-125,apolipoprotein A,and OM.METHODS:A total of 1362 menopausal female BC patients with OM volunteered to take part in this study between July 2012 and July 2022.Women with BC who are menopausal were found to have an OM incidence of 1.6%.Furthermore,CA-153,CA-125,and apolipoprotein A(Apo A)all contributed to OM in women with BC who are postmenopausal according to binary logistic regression.Receiver operating curve(ROC)analysis was used to assess the diagnostic value of OM in patients with BC.RESULTS:Both CA-153 and CA-153+CA-125 showed a higher sensitivity of 95.45%,whereas CA-153+Apo A illustrated the highest specificity of 99.02%.Moreover,CA-153 and CA-153+CA-125 had higher areas under the curve(AUC)of 0.973.CONCLUSION:The data indicate that the serum concentrations of CA-153 exhibited the most significant predictors of the diagnosis of OM in menopausal women with BC.The current study researches the utility of risk factors in predicting of OM in menopausal BC women and put forward the latest suggestions on their clinical application.
基金supported by the National Key R&D Program of China(No.2023YFB3001704)NSFC for Young Scientists Fund(No.62402266)NSFC for Distinguished Young Scholar(No.62225206).
文摘Large models have been widely used in the field of neural language processing,information retrieving,etc.With the development of the large models,not only is the parameter scale increased,but the model architecture has also become more complex.For example,the multi-modal transformer-based model mainly has concurrent branches,which we denoted as the concurrent branch model(CBM).Many CBMs have enlarged to tens of billions of parameters,and require distributed resources to train this kind of model.Existing distributed training systems cannot fully handle this type of model architecture because there are interactions between branches.Inspired by the unbalanced resource usage of pipeline parallelism,we prefer to organize different branches with a fine-grained bidirectional pipeline schedule of communication and computation.However,improper coordination between branches leads to idle time for computation and low training efficiency.In this paper,we present Flexpipe,a pipeline engine for c3oncurrent-branch models.We first introduce a branch-aware pipeline parallelism(BAPP)to make full use of the concurrent characteristic of the model architecture.Then,based on a multi-branch pipeline simulator,we propose an adaptive interaction coordinator,which facilitates the low-overhead branch interactions during the distributed model training.We evaluate our approach on popular concurrent branch models combined with modern training systems.Compared with the Chimera,the experiential results show that our method improves the end-to-end training throughput by 20%on average.
基金This work was supported in part by the National Science and Technology Major Project under Grant 2022ZD0115200in part by Frontier Technique Collaboration Project under Grant QYJS-2023-2801-B+3 种基金in part by NSFC under Grant 62125403 and Grant 92164301in part by the Beijing S&T Project under Grant Z221100007722023in part by the Shanghai Municipal Science and Technology Major Project,in part by the 2022 Special Project on Industrial Foundation Reconstruction and High Quality Development of Manufacturing Industry under Grant CEIEC-2022-ZM02-0245in part by the Beijing National Research Center for Information Science and Technology,and in part by the Beijing Advanced Innovation Center for Integrated Circuits.
文摘Transformer-based large language models(LLMs)have made significant strides in the field of artificial intelligence(AI).However,training these LLMs imposes immense demands on computational power and bandwidth for hardware systems.Wafer-scale chips(WSCs)offer a promising solution,yet they struggle with limited on-chip memory and complex tensor partitioning.To fully harness the high-bandwidth,low-latency on-chip interconnect benefits of WSCs and to alleviate the on-chip memory limitations,a specialized mapping and architecture co-exploration method is essential.Despite existing efforts in memory optimization and mapping,current approaches fall short for WSC scenarios.To bridge this gap,we introduce TMAC,an architecture-mapping co-exploration framework that integrates recomputation into the design space,fully exploiting optimization opportunities overlooked by existing works.Further,TMAC takes advantage of the superior on-chip interconnect performance of WSCs by incorporating a more flexible tensor partition scheme.TMAC then introduces a novel operator-centric encoding scheme(OCES)designed to comprehensively describe the mapping space for training LLMs.Unlike previous studies that focus solely on communication volume analysis based on mapping,TMAC explores the design space by evaluating the combined impact of mapping and architecture on training performance.However,fully accounting for these untapped optimization opportunities increases the complexity of the design space.To address this,we streamline the simulation process,reducing the time needed for exploration.Compared to AccPar,Deepspeed and Megatron,TMAC delivers a 3.1×,2.9×,1.6×performance gain.In terms of memory usage,TMAC requires 3.6×,3.1×less memory than AccPar and Deepspeed,respectively and is comparable to Megatron’s full recomputation method.
基金support are from the National Natural Science Foundation of China(22122107)the Fundamental Research Funds for Central Universities(2042021kf0190).
文摘Chain-walking has emerged as a promising strategy in organic synthetic methodology,but achieving site-selectivity in reactions involving competition among multiple potential positions remains a challenge.In this study,we presented a novel approach to ligand-modulated,nickel-catalyzed regiodivergent alkenylboration of allylarenes.Our method allows for highly chemoselective preparation of two classes of structurally different alkenyl boronates from the same starting by simply switching the ligand.Mechanistic investigations involving experiments and calculations suggest that the ligand-modulating regiodivergence arises from either regioselectiveβ-H elimination or oxidative addition with organic halides.Our findings offer valuable insight into addressing site-selective issues in chain-walking reactions with multiple thermodynamically stable factors.
文摘Real-world blind image super-resolution is a challenging problem due to the absence of target high resolution images for training.Inspired by the recent success of the single image generation based method SinGAN,we tackle this challenging problem with a refined model SR-SinGAN,which can learn to perform single real image super-resolution.Firstly,we empirically find that downsampled LR input with an appropriate size can improve the robustness of the generation model.Secondly,we introduce a global contextual prior to provide semantic information.This helps to remove distorted pixels and improve the output fidelity.Finally,we design an image gradient based local contextual prior to guide detail generation.It can alleviate generated artifacts in smooth areas while preserving rich details in densely textured regions(e.g.,hair,grass).To evaluate the effectiveness of these contextual priors,we conducted extensive experiments on both artificial and real images.Results show that these priors can stabilize training and preserve output fidelity,improving the generated image quality.We furthermore find that these single image generation based methods work better for images with repeated textures compared to general images.
基金supported by the National Natural Science Foundation of China,Youth Fund,China(No.62102318)the Basic Research Programs of Taicang,China,2023(No.TC2023JC23)+2 种基金funded in part by the National Natural Science Foundation of China(No.92267203)in part by the Science and Technology Major Project of Guangzhou,China(No.202007030006)in part by the Program for Guangdong Introducing Innovative and Entrepreneurial Teams,China(No.2019ZT08X214).
文摘Shadows in document images are undesirable yet inevitable.They can decrease the clarity and readability of the images.The existing methods for removing shadows from documents still face some challenges,such as the traditional heuristics lack universality and the optimization goal of subnetworks is not consistent for multistage deep learning methods.In this paper,we introduce an end-to-end direct document shadow removal network(DDSR-Net),where we employ a 3-layer UNet++as the backbone to extract features from diverse scales.To further improve the performance of DDSR-Net,we integrate the multi-scale attention(MSA)blocks into each node.The MSA block allocates different weights to feature vectors at different positions,achieving automatic feature alignment and significantly enhancing the end-to-end network's ability to handle shadow processing.To verify the effectiveness of the proposed DDSR-Net,qualitative and quantitative experiments are conducted on multiple open-source document shadow removal datasets.The experimental results demonstrate that our method outperforms the existing state-of-the-art methods on these datasets.Our code and models will be released to the public.
基金supported in part by the National Natural Science Foundation of China(Grant Nos.62272468,62003256,62027813,U1801265,62293543,62322605,62036005,62202015,and U21B2048)the Key-Area Research and Development Program of Shaanxi Province(2023-ZDLSF-41)+2 种基金the Anhui Medical University(2022xkj105,2023cy021)the Anhui Provincial Key R&D Program(2023s07020001)the University Synergy Innovation Program of Anhui Province(GXXT-2022-052).
文摘1 Introduction Endoscopy plays a crucial role in the diagnoses and treatment of gastrointestinal(GI)diseases[1],as it helps to identify abnormalities,classify lesion,and determine treatment methods.During GI endoscopic examinations,physicians may encounter practical hindrances,i.e.,fatigue,stress,or limited experience,which can lead to erroneous results.Artificial intelligence(AI)-assisted GI endoscopy technology has emerged to address these limitations[2].
基金supported by the National Natural Science Foundation of China under Grant No.62173317the Key Research and Development Program of Anhui under Grant No.202104a05020064。
文摘In this paper,a cross-sensor generative self-supervised learning network is proposed for fault detection of multi-sensor.By modeling the sensor signals in multiple dimensions to achieve correlation information mining between channels to deal with the pretext task,the shared features between multi-sensor data can be captured,and the gap between channel data features will be reduced.Meanwhile,in order to model fault features in the downstream task,the salience module is developed to optimize cross-sensor data features based on a small amount of labeled data to make warning feature information prominent for improving the separator accuracy.Finally,experimental results on the public datasets FEMTO-ST dataset and the private datasets SMT shock absorber dataset(SMT-SA dataset)show that the proposed method performs favorably against other STATE-of-the-art methods.
基金Supported by Major Project of Guangzhou National Laboratory,(Grant No.GZNL2024A01004)the National Natural Science Foundation of China(Grant No.82361168672)+4 种基金the Science and Technology Development Fund of Macao SAR(Grant No.FDCT 0111/2023/AFJ,0155/2024/RIA2,005/2022/ALC,0128/2022/A,0020/2023/RIB1)National Key Research and Development Program of China(Grant No.2024YFE0214800)Self-supporting Program of Guangzhou Laboratory(Grant No.SRPG22-007)National Key Research and Development Program of China(Grant No.SQ2024YFE0202244)Engineering Technology Research(Development)Center of Ordinary Colleges and Universities in Guangdong Province(Grant No.2024GCZX010).
文摘Introduction:Seasonal influenza poses a significant public health burden,causing substantial morbidity and mortality worldwide each year.In this context,timely and accurate vaccine strain selection is critical to mitigating the impact of influenza outbreaks.This article aims to develop an adaptive,universal,and convenient method for predicting antigenic variation in influenza A(H1N1),thereby providing a scientific basis to enhance the biannual influenza vaccine selection process.Methods:The study integrates adaptive Fourier decomposition(AFD)theory with multiple techniques—including matching pursuit,the maximum selection principle,and bootstrapping—to investigate the complex nonlinear interactions between amino acid substitutions in hemagglutinin(HA)proteins(the primary antigenic protein of influenza virus)and their impact on antigenic changes.Results:Through comparative analysis with classical methods such as Lasso,Ridge,and random forest,we demonstrate that the AFD-type method offers superior accuracy and computational efficiency in identifying antigenic change-associated amino acid substitutions,thus eliminating the need for timeconsuming and expensive experimental procedures.AAW Conclusion:In summary,AFD-based methods represent effective mathematical models for predicting antigenic variations based on HA sequences and serological data,functioning as ensemble algorithms with guaranteed convergence.Following the sequence of indicators specified in I,we perform a series of operations on A_(1),including feature extension,extraction,and rearrangement,to generate a new input dataset for the prediction step.With this newly prepared input,we can compute the predicted results as.
基金support from the National Key R&D Program of China(2021YFF0502900)the National Natural Science Foundation of China(62127819,T2421003,62435011,32471230,W2431056)+2 种基金the Shenzhen Key Laboratory of Photonics and Biophotonics(ZDSYS20210623092006020)the Shenzhen Science and Technology Program(JCYJ20220818100202005)and the Fundamental Research Funds for the Central Universities(23GH02021).
文摘The emergence of nonconventional luminescent materials(NLMs)has attracted significant attention due to their sustainable synthesis and tunable optical properties.Yet,establishing a clear structure-emission relationship remains a challenge.In this work,we report a previously unknown class of NLMs:cross-linked protein crystals that exhibit intense photoluminescence(PL)in the visible range(425-680 nm).We systematically investigated seven natural protein crystals(concanavalin,catalase,lysozyme,hemoglobin,α-chymotrypsin,pepsin,and β-lactoglobulin)cross-linked with glutaraldehyde and demonstrated that cross-linking induces broadband emission that is absent in natural crystals.Focusing on polymorphic lysozyme crystals(tetragonal,orthorhombic,and monoclinic),we found excitation-dependent fluorescence with lifetimes in the nanosecond range and quantum yields up to 20%(in themonoclinic phase under 450 nmexcitation).Single-and two-photon spectroscopy,as well as pressure-and solvent-modulated PL studies,confirm that the emission is due to intermolecular through-space interactions(TSI)within the crystal lattice.Compression enhances TSI and redshifts the emission,whereas the solvent(DMSO)-induced swelling reduces TSI and causes a blue shift,establishing a direct structure-emission correlation.This work establishes protein crystals as programmableNLMswith tunable emission and provides a mechanistic framework for the design of nonconventional luminogens through protein crystal engineering.
基金National Natural Science Foundation of China under Grant Nos.61672273 and 61832008Science Foundation for Distinguished Young Scholars of Jiangsu under Grant No.BK20160021+1 种基金Postdoctoral Innovative Talent Support Program of China under Grant Nos.BX20200168,2020M681608General Research Fund of Hong Kong under Grant No.27208720。
文摘Transformers have recently lead to encouraging progress in computer vision.In this work,we present new baselines by improving the original Pyramid Vision Transformer(PVT v1)by adding three designs:(i)a linear complexity attention layer,(ii)an overlapping patch embedding,and(iii)a convolutional feed-forward network.With these modifications,PVT v2 reduces the computational complexity of PVT v1 to linearity and provides significant improvements on fundamental vision tasks such as classification,detection,and segmentation.In particular,PVT v2 achieves comparable or better performance than recent work such as the Swin transformer.We hope this work will facilitate state-ofthe-art transformer research in computer vision.Code is available at https://github.com/whai362/PVT.
