|
|
|
题名
|
作者
|
年代
|
出处
|
被引量
|
| 1 | The Deep Ritz Method: A Deep Learning-Based Numerical Algorithm for Solving Variational Problems显示文摘We propose a deep learning-based method,the Deep Ritz Method,for numerically solving variational problems,particularly the ones that arise from par-tial differential equations.The Deep Ritz Method is naturally nonlinear,naturally adaptive and has the potential to work in rather high dimensions.The framework is quite simple and fits well with the stochastic gradient descent method used in deep learning.We illustrate the method on several problems including some eigenvalue problems. | Weinan E Bing Yu | 2018 | Communications in Mathematics and Statistics2018，6，1： | 22 |
| 2 | The heterogeneity of islet autoantibodies and the progression of islet failure in type 1 diabetic patients显示文摘Type 1 diabetes mellitus is heterogeneous in many facets. The patients suffered from type 1 diabetes present several levels of islet function as well as variable number and type of islet-specific autoantibodies. This study was to investigate prevalence and heterogeneity of the islet autoantibodies and clinical phenotypes of type 1 diabetes mellitus; and also discussed the process of islet failure and its risk factors in Chinese type 1 diabetic patients. A total of 1,291 type 1 diabetic patients were enrolled in this study. Demographic information was collected. Laboratory tests including mixed-meal tolerance test, human leukocyte antigen alleles, hemoglobin A1 c, lipids, thyroid function and islet autoantibodies were conducted. The frequency of islet-specific autoantibody in newly diagnosed T1 DM patients(duration shorter than half year) was 73% in East China. According to binary logistic regressions, autoantibody positivity, longer duration and lower Body Mass Index were the risk factors of islet failure. As the disease developed, autoantibodies against glutamic acid decarboxylase declined as well as the other two autoantibodies against zinc transporter 8 and islet antigen 2. The decrease of autoantibodies was positively correlated with aggressive beta cell destruction. Autoantibodies can facilitate the identification of classic T1 DM from other subtypes and predict the progression of islet failure. As there were obvious heterogeneity in autoantibodies and clinical manifestation in different phenotypes of the disease, we should take more factors into consideration when identifying type 1 diabetes mellitus. | Jin Liu Lingling Bian Li Ji Yang Chen Heng Chen Yong Gu Bingqin Ma Wei Gu Xinyu Xu Yun Shi Jian Wang Dalong Zhu Zilin Sun Jianhua Ma Hui Jin Xing Shi Heng Miao Bing Xin Yan Zhu Zhenwen Zhang Ruifang Bu Lan Xu Guangde Shi Wei Tang Wei Li Dongmei Zhou Jun Liang Xingbo Cheng Bimin Shi Jixiang Dong Ji Hu Chen Fang Shao Zhong Weinan Yu Weiping Lu Chenguang Wu Li Qian Jiancheng Yu Jialin Gao Xiaoqiang Fei Qingqing Zhang Xueqin Wang Shiwei Cui Jinluo Cheng Ning Xu Guofeng Wang Guoqing Han Chunrong Xu Yun Xie Minmin An Wei Zhang Zhixiao Wang Yun Cai Qi Fu Yu Fu Shuai Zheng Fan Yang Qingfang Hu Hao Dai Yu Jin Zheng Zhang Kuanfeng Xu Yifan Li Jie Shen Hongwen Zhou Wei He Xuqin Zheng Xiao Han Liping Yu Jinxiong She Mei Zhang Tao Yang | 2016 | Science China(Life Sciences)2016，59，9： | 5 |
| 3 | Variable elliptical vibrating screen: Particles kinematics and industrial application显示文摘Traditional vibrating screen usually adopts the linear centralized excitation mode,which causes the difficulty in particles loosening and low screening efficiency.The variable elliptical vibrating screen(VEVS)trajectory is regulated to adapt the material mass along the direction of the screen length,improving the particles distribution as well as the screening efficiency.In this work,a theoretical model was developed for analyzing the screen surface motion law during VEVS-based screening process.An equation was obtained to show the relationship between the horizontal amplitude and the vertical amplitude.The materials kinetic characteristics were studied by using high-speed camera during screening process.Compared with equal-amplitude screen(EAS),the material moving velocity was increased by 13.03%on the first half but decreased by 3.52% on the second half,and the total screening time was reduced by 9.42% by using VEVS.In addition,-6 mm screening test was carried out.At the length of VEVS equaled to 1.2 m,the screening efficiency and the total misplaced material content were 92.50% and 2.90%,respectively.However,the screening efficiency was 89.91% and the total misplaced material content was 3.76% during EAS-based screening process.Furthermore,when external moisture is 5.96%,the screening efficiency of VEVS could reach 86.95%.The 2 TKB50113 type VEVS with double-layered screen surface used in Huoshizui Coal Mine was 5.0 m in width and 11.3 m in length.The areas of single layer and double layer were 56.5 and 113 m~2,respectively.In industrial production,the processing capacity was 2500-3000 t/h and the screening efficiency was larger than 90%. | Chenlong Duan Jiale Yuan Miao Pan Tao Huang Haishen Jiang Yuemin Zhao Jinpeng Qiao Weinan Wang Shijie Yu Jiawang Lu | 2021 | International Journal of Mining Science and Technology2021，31，6： | 2 |
| 4 | Effects of amendment of biochar-manure compost in conjunction with pyroligneous solution on soil quality and wheat yield of a salt-stressed cropland from Central China Great Plain显示文摘 | Muhammad Siddique Lashari Yuming Liu Lianqing Li Weinan Pan Jiaying Fu Genxing Pan Jufeng Zheng Jinwei Zheng Xuhui Zhang Xinyan Yu | 2012 | Field Crops Research2012，，： | 1 |
| 5 | Generalized variational principles, gloval weak solutions and behavior with random initial data for systems of conservation laws arising in adhesion particle dynamics 显示文摘 | WEINAN E RYKOV YU G SINAI YA G | 1996 | Comm Math Phys1996，177，： | 1 |
| 6 | Gated recurrent unit model for a sequence tagging problem显示文摘Combinatory categorial grammer(CCG) supertagging is an important subtask that takes place before full parsing and can benefit many natural language processing(NLP) tasks like question answering and machine translation. CCG supertagging can be regarded as a sequence labeling problem that remains a challenging problem where each word is assigned to a CCG lexical category and the number of the probably associated CCG supertags to each word is large. To address this, recently recurrent neural networks(RNNs), as extremely powerful sequential models, have been proposed for CCG supertagging and achieved good performances. In this paper, a variant of recurrent networks is proposed whose design makes it much easier to train and memorize information for long range dependencies based on gated recurrent units(GRUs), which have been recently introduced on some but not all tasks. Results of the experiments revealed the effectiveness of the proposed method on the CCGBank datasets and show that the model has comparable accuracy with the previously proposed models for CCG supertagging. | Rekia Kadari Zhang Yu Zhang Weinan Liu Ting | 2019 | High Technology Letters2019，25，1： | 1 |
| 7 | Neural recovery machine for Chinese dropped pronoun显示文摘Dropped pronouns (DPs) are ubiquitous in pro-drop languages like Chinese, Japanese etc. Previous work mainly focused on painstakingly exploring the empirical features for DPs recovery. In this paper, we propose a neural recovery machine (NRM) to model and recover DPs in Chinese to avoid the non-trivial feature engineering process. The experimental results show that the proposed NRM significantly outperforms the state-of-the-art approaches on two heterogeneous datasets. Further experimental results of Chinese zero pronoun (ZP) resolution show that the performance of ZP resolution can also be improved by recovering the ZPs to DPs. | Weinan ZHANG Ting LIU Qingyu YIN Yu ZHANG | 2019 | Frontiers of Computer Science2019，13，5： | 0 |
| 8 | Efficient policy evaluation by matrix sketching显示文摘In the reinforcement learning,policy evaluation aims to predict long-term values of a state under a certain policy.Since high-dimensional representations become more and more common in the reinforcement learning,how to reduce the computational cost becomes a significant problem to the policy evaluation.Many recent works focus on adopting matrix sketching methods to accelerate least-square temporal difference(TD)algorithms and quasi-Newton temporal difference algorithms.Among these sketching methods,the truncated incremental SVD shows better performance because it is stable and efficient.However,the convergence properties of the incremental SVD is still open.In this paper,we first show that the conventional incremental SVD algorithms could have enormous approximation errors in the worst case.Then we propose a variant of incremental SVD with better theoretical guarantees by shrinking the singular values periodically.Moreover,we employ our improved incremental SVD to accelerate least-square TD and quasi-Newton TD algorithms.The experimental results verify the correctness and effectiveness of our methods. | Cheng CHEN Weinan ZHANG Yong YU | 2022 | Frontiers of Computer Science2022，16，5： | 0 |
| 9 | A survey on model-based reinforcement learning显示文摘Reinforcement learning(RL)interacts with the environment to solve sequential decision-making problems via a trial-and-error approach.Errors are always undesirable in real-world applications,even though RL excels at playing complex video games that permit several trial-and-error attempts.To improve sample efficiency and thus reduce errors,model-based reinforcement learning(MBRL)is believed to be a promising direction,as it constructs environment models in which trial-and-errors can occur without incurring actual costs.In this survey,we investigate MBRL with a particular focus on the recent advancements in deep RL.There is a generalization error between the learned model of a non-tabular environment and the actual environment.Consequently,it is crucial to analyze the disparity between policy training in the environment model and that in the actual environment,guiding algorithm design for improved model learning,model utilization,and policy training.In addition,we discuss the recent developments of model-based techniques in other forms of RL,such as offline RL,goal-conditioned RL,multi-agent RL,and meta-RL.Furthermore,we discuss the applicability and benefits of MBRL for real-world tasks.Finally,this survey concludes with a discussion of the promising future development prospects for MBRL.We believe that MBRL has great unrealized potential and benefits in real-world applications,and we hope this survey will encourage additional research on MBRL. | Fan-Ming LUO Tian XU Hang LAI Xiong-Hui CHEN Weinan ZHANG Yang YU | 2024 | Science China(Information Sciences)2024，67，2： | 0 |
| 10 | Effective strategies to promote Z(S)-scheme photocatalytic water splitting显示文摘Artificial Z(S)-scheme photocatalytic water splitting systems have attracted extensive attention due to their advantages such as wide light absorption range,high charge separation efficiency and strong carrier redox ability.However,it is still challenging to design and prepare Z(S)-scheme photocatalysts with low-cost and highly stability for efficiently photocatalytic overall water splitting using solar energy.This review mainly introduces various strategies to improve the photocatalytic water splitting performance of Z(S)-scheme systems.These strategies mainly focus on enhancing or extending the range of light absorption,promoting charge separation,and enhancing surface redox reaction in Z(S)-scheme systems.Finally,the main challenges of Z(S)-scheme photocatalytic water splitting systems and their future development directions are pointed out.This review would be beneficial to understanding the challenges and opportunities faced by the research field of Z(S)-scheme photocatalytic systems,and has important guiding significance for the development and utilization of high-performance Z(S)-scheme photocatalytic reaction system in the future. | Ye Yuan Junan Pan Weinan Yin Haoxuan Yu Fengshun Wang Weifeng Hu Longlu Wang Dafeng Yan | 2024 | Chinese Chemical Letters2024，35，3： | 0 |
| 11 | Revealing the interaction mechanism of pulsed laser processing with the application of acoustic emission显示文摘The mechanisms of interaction between pulsed laser and materials are complex and indistinct,severely infuencing the stability and quality of laser processing.This paper proposes an intelligent method based on the acoustic emission(AE)technique to monitor laser processing and explore the interaction mechanisms.The validation experiment is designed to perform nanosecond laser dotting on foat glass.Processing parameters are set diferently to generate various outcomes:ablated pits and irregular-shaped cracks.In the signal processing stage,we divide the AE signals into two bands,main and tail bands,according to the laser processing duration,to study the laser ablation and crack behavior,respectively.Characteristic parameters extracted by a method that combines framework and frame energy calculation of AE signals can efectively reveal the mechanisms of pulsed laser processing.The main band features evaluate the degree of laser ablation from the time and intensity scales,and the tail band characteristics demonstrate that the cracks occur after laser dotting.In addition,from the analysis of the parameters of the tail band very large cracks can be efciently distinguished.The intelligent AE monitoring method was successfully applied in exploring the interaction mechanism of nanosecond laser dotting foat glass and can be used in other pulsed laser processing felds. | Weinan Liu Youmin Rong Ranwu Yang Congyi Wu Guojun Zhang Yu Huang | 2023 | Frontiers of Optoelectronics2023，16，2： | 0 |
您所在的机构暂未开通此部分的功能，请联系管理员