Filters








504 Hits in 1.8 sec

Explainable Enterprise Credit Rating via Deep Feature Crossing Network [article]

Weiyu Guo, Zhijiang Yang, Shu Wu, Fu Chen
2021 arXiv   pre-print
Due to the powerful learning ability on high-rank and non-linear features, deep neural networks (DNNs) are being applied to data mining and machine learning in various fields, and exhibit higher discrimination performance than conventional methods. However, the applications based on DNNs are rare in enterprise credit rating tasks because most of DNNs employ the "end-to-end" learning paradigm, which outputs the high-rank representations of objects and predictive results without any explanations.
more » ... Thus, users in the financial industry cannot understand how these high-rank representations are generated, what do they mean and what relations exist with the raw inputs. Then users cannot determine whether the predictions provided by DNNs are reliable, and not trust the predictions providing by such "black box" models. Therefore, in this paper, we propose a novel network to explicitly model the enterprise credit rating problem using DNNs and attention mechanisms. The proposed model realizes explainable enterprise credit ratings. Experimental results obtained on real-world enterprise datasets verify that the proposed approach achieves higher performance than conventional methods, and provides insights into individual rating results and the reliability of model training.
arXiv:2105.13843v1 fatcat:h7jfzrkxvrgx5eyekmgf3qrkam

Nifty Futures Rollover Strategies

Ronald T. Slivka, Jie Yang, Weiyu Wan
2017 Indian Journal of Finance  
Issues pertaining to the investor decision to sell a security and buy another (of the same type and with the same terms) with a longer period until the expiration date (the roll forward decision) are examined. In particular, a framework is developed in which it is possible to test the trade execution quality efficiency of a roll strategy against a mean-variance optimal roll strategy characterized by multiple-day roll. Applying this framework to five leading US grain futures markets (corn,
more » ... soybean, soybean meal and soybean oil) demonstrates that commonly used single-day and multiple-day roll strategies (including the Goldman roll strategy) exhibit considerable inefficiencies. These are consistent over the markets and over the time of the day in which trading occurs, and vary with execution quality risk-aversion in a predictable way. A practical multiple-day roll strategy is proposed that reduces these inefficiencies.
doi:10.17010/ijf/2017/v11i7/116563 fatcat:wc7ebl46q5cl7d6imkqm3avequ

Study of the Ti-20wt. % Mo Composite Coating Prepared by Laser Cladding

Rongjuan Yang, Zongde Liu, Guang Yang, Weiyu Ren
2012 Procedia Engineering  
Laser surface cladding was applied on low carbon steel to improve its surface properties. Mixed Mo and Ti powders with a Ti-to-Mo mass ratio of 4:1 were put onto the low carbon steel and subsequently treated by laser beam. The microstructure and composition modifications in the surface layer were carefully investigated by using SEM, EDX and XRD, respectively. A Ti-Mo intermetallic compound layer with thickness of similar to 300 μm was formed on low carbon steel. The bonding between the coating
more » ... ayer and substrate is good and the diffusion occurs in the interface between them. From the results it is very clear that the phases in the composites coatings are composed of -Ti phase and Fe 2 Ti phases. The micro hardness of the coating reaches to above HV 0.5 900. The improvement of the hardness of the coating on low carbon steel has been realized. Preliminary electrochemical studies have indicated Ti-Mo exhibits excellent corrosion resistance in 0.5M NaCl solution.
doi:10.1016/j.proeng.2012.03.052 fatcat:neljcku6nfgffiyghngwkb5sbu

Proton Irradiation-Decelerated Intergranular Corrosion of Ni-Cr Alloys in Molten Salt [article]

Weiyue Zhou, Yang Yang, Guiqiu Zheng, Kevin B Woller, Peter W Stahle, Andrew M Minor, Michael P Short
2019 arXiv   pre-print
The effects of ionizing radiation on materials often reduce to "bad news." Radiation damage usually leads to detrimental effects such as embrittlement, accelerated creep, phase instability, and radiation-altered corrosion. This last point merits special attention. Elucidating synergies between radiation and corrosion has been one of the most challenging tasks impeding the deployment of advanced reactors, stemming from the combined effects of high temperature, corrosive coolants, and intense
more » ... icle fluxes. Here we report that proton irradiation significantly and repeatably decelerates intergranular corrosion of Ni-Cr alloys in molten fluoride salt at 650C. We demonstrate this effect by showing that the depth of intergranular voids resulting from Cr leaching into the salt is reduced by the proton irradiation alone. Interstitial defects generated from proton irradiation result in radiation-enhanced diffusion, more rapidly replenishing corrosion-injected vacancies with alloy constituents, thus playing the crucial role in decelerating corrosion. Our results show that in industrially-relevant scenarios irradiation can have a positive impact, challenging our view that radiation damage always results in negative effects.
arXiv:1911.11798v1 fatcat:q24ewfpe3rcgfghrkl4dwfkyde

Security-Oriented Network Architecture

Weiyu Jiang, Bingyang Liu, Chuang Wang, Xue Yang, Salvatore D'Antonio
2021 Security and Communication Networks  
Internet benefits societies by constantly connecting devices and transmitting data across the world. However, due to the lack of architectural built-in security, the pervasive network attacks faced by the entire information technology are considered to be unending and inevitable. As Internet evolves, security issues are regularly fixed according to a patch-like strategy. Nevertheless, the patch-like strategy generally results in arms races and passive situations, leaving an endless lag in both
more » ... xisting and emerging attacking surface. In this paper, we present NAIS (Network Architecture with Intrinsic Security)—a network architecture towards trustworthiness and security. By solving stubborn security issues like IP spoofing, MITM (man-in-the-middle) attacks, and DDoS (distributed denial of service) attacks at architectural level, NAIS is envisioned to provide the most secure end-to-end communication in the network layer. This paper first presents a comprehensive analysis of network security at Internet range. Then, the system design of NAIS is elaborated with particular design philosophies and four security techniques. Such philosophies and techniques intertwine internally and contribute to a communication environment with authenticity, privacy, accountability, confidentiality, integrity, and availability. Finally, we evaluate the security functionalities on the packet forwarding performance, demonstrating that NAIS can efficiently provide security and trustworthiness in Internet end-to-end communication.
doi:10.1155/2021/6694650 fatcat:ta3azw6canerlahszyhvuxvhde

A hSCARB2-transgenic mouse model for Coxsackievirus A16 pathogenesis

Yanli Chen, Heng Li, Jinxi Yang, Huiwen Zheng, Lei Guo, Weiyu Li, Zening Yang, Jie Song, Longding Liu
2021 Virology Journal  
Background Coxsackievirus A16 (CA16) is one of the neurotropic pathogen that has been associated with severe neurological forms of hand, foot, and mouth disease (HFMD), but its pathogenesis is not yet clear. The limited host range of CA16 make the establishment of a suitable animal model that can recapitulate the neurological pathology observed in human HFMD more difficult. Because the human scavenger receptor class B, member 2 (hSCARB2) is a cellular receptor for CA16, we used transgenic mice
more » ... earing human SCARB2 and nasally infected them with CA16 to study the pathogenicity of the virus. Methods Coxsackievirus A16 was administered by intranasal instillation to groups of hSCARB2 transgenic mice and clinical signs were observed. Sampled at different time-points to document and characterize the mode of viral dissemination, pathological change and immune response of CA16 infection. Results Weight loss and virus replication in lung and brain were observed in hSCARB2 mice infected with CA16, indicating that these animals could model the neural infection process. Viral antigens were observed in the alveolar epithelia and brainstem cells. The typical histopathology was interstitial pneumonia with infiltration of significant lymphocytes into the alveolar interstitial in lung and diffuse punctate hemorrhages in the capillaries of the brainstem. In addition, we detected the expression levels of inflammatory cytokines and detected high levels of interleukin IL-1β, IL-6, IL-18, and IFN-γ in nasal mucosa, lungs and brain tissues. Conclusions The hSCARB2-transgenic mice can be productively infected with CA16 via respiratory route and exhibited a clear tropism to lung and brain tissues, which can serve as a model to investigate the pathogenesis of CA16 associated respiratory and neurological disease.
doi:10.1186/s12985-021-01557-5 pmid:33882964 pmcid:PMC8061046 fatcat:4vz52rcfyje3hf75pdzgvjphim

Laser induced graphitization of PAN-based carbon fibers

Yang Sha, Weimin Yang, Sanyang Li, Liangbo Yao, Haoyi Li, Lisheng Cheng, Hua Yan, Weiyu Cao, Jing Tan
2018 RSC Advances  
A new method based on laser technique is proposed to graphitize PAN-based CFs.
doi:10.1039/c8ra00497h pmid:35542777 pmcid:PMC9079127 fatcat:fmvvaonksrexxn42vouwq2x5pa

Shape Inpainting using 3D Generative Adversarial Network and Recurrent Convolutional Networks [article]

Weiyue Wang, Qiangui Huang, Suya You, Chao Yang, Ulrich Neumann
2017 arXiv   pre-print
Recent advances in convolutional neural networks have shown promising results in 3D shape completion. But due to GPU memory limitations, these methods can only produce low-resolution outputs. To inpaint 3D models with semantic plausibility and contextual details, we introduce a hybrid framework that combines a 3D Encoder-Decoder Generative Adversarial Network (3D-ED-GAN) and a Long-term Recurrent Convolutional Network (LRCN). The 3D-ED-GAN is a 3D convolutional neural network trained with a
more » ... rative adversarial paradigm to fill missing 3D data in low-resolution. LRCN adopts a recurrent neural network architecture to minimize GPU memory usage and incorporates an Encoder-Decoder pair into a Long Short-term Memory Network. By handling the 3D model as a sequence of 2D slices, LRCN transforms a coarse 3D shape into a more complete and higher resolution volume. While 3D-ED-GAN captures global contextual structure of the 3D shape, LRCN localizes the fine-grained details. Experimental results on both real-world and synthetic data show reconstructions from corrupted models result in complete and high-resolution 3D objects.
arXiv:1711.06375v1 fatcat:gi6o36a2jrcfhamtdjqrhjec4y

Enhanced immunocompatibility of ligand-targeted liposomes by attenuating natural IgM absorption

Juan Guan, Qing Shen, Zui Zhang, Zhuxuan Jiang, Yang Yang, Meiqing Lou, Jun Qian, Weiyue Lu, Changyou Zhan
2018 Nature Communications  
Targeting ligands are anticipated to facilitate the precise delivery of therapeutic agents to diseased tissues; however, they may also severely affect the interaction of nanocarriers with plasma proteins. Here, we study the immunocompatibility of brain-targeted liposomes, which inversely correlates with absorbed natural IgM. Modification of long, stable positively charged peptide ligands on liposomes is inclined to absorb natural IgM, leading to rapid clearance and enhanced immunogenicity.
more » ... peptidomimetic D8 developed by computer-aided peptide design exhibits improved immunocompatibility by attenuating natural IgM absorption. The present study highlights the effects of peptide ligands on the formed protein corona and in vivo fate of liposomes. Stable positively charged peptide ligands play double-edged roles in targeted delivery, preserving in vivo bioactivities for binding receptors and long-term unfavorable interactions with the innate immune system. The development of D8 provides insights into how to rationally design immunocompatible drug delivery systems by modulating the protein corona composition.
doi:10.1038/s41467-018-05384-1 pmid:30061672 pmcid:PMC6065320 fatcat:iqqrtjxf4zcfbedozdqcqnynre

Dextran-based Nanomicelle System with Directly Ester-bound Gadolinium Chelates as a Magnetic Resonance Imaging Contrast Agent for Tumor Detection

Lirong Ma, Weiyu Yang, Dingmei Wu, Mengmeng Jiang, Xiuzhong Yao, Weiyu Weng
2020 Iranian journal of pharmaceutical research  
There is a strong need to develop MRI contrast agents (CAs) with lower in-vivo retention, stronger signal enhancement, and more specific imaging. Here, we report a novel dextran (DEX)-based nanomicelle system as an MRI CA with superior tumor imaging and relatively short intravascular persistence. Gadolinium (Gd)-chelate (DTPA-Gd) was conjugated directly to DEX hydroxyl via a degradable ester bond. DEX-DTPA-Gd was then modified with dodecylsuccinic anhydride to obtain the amphiphilic derivative,
more » ... 2-dodecylsuccinic acid (DSA)-grafted DEX-DTPA-Gd. Nanomicelles were prepared by dissolving DSA-DEX-DTPA-Gd in water using ultrasonication. The physicochemical properties, cytotoxicity, and MRI efficiency of the synthesized CA were evaluated. The synthesized DSA-DEX-DTPA-Gd self-assembled into nanomicelles with an average diameter of 67.80 ± 5.21 nm. Within the given Gd concentration range, DSA-DEX-DTPA-Gd and Magnevist® exhibited similar cytotoxicity. DEX-based CAs resulted in a greater contrast enhancement of T1-weighted signal intensity in the tumor region than Magnevist®, and the tumors were clearly defined for at least 3 h. Simultaneously, the ester bond in DSA-DEX-DTPA-Gd facilitated the elimination of Gd chelates, compared with the relatively more stable amide linker. The DEX-based nanomicelle system with directly ester-bound DTPA-Gd may serve as an MRI CA with superior tumor imaging and relatively rapid elimination.
pmid:33680049 pmcid:PMC7758001 fatcat:fzvzxsytp5hzhlpbcbcswjgqre

Brain-targeted drug delivery by manipulating protein corona functions

Zui Zhang, Juan Guan, Zhuxuan Jiang, Yang Yang, Jican Liu, Wei Hua, Ying Mao, Cheng Li, Weiyue Lu, Jun Qian, Changyou Zhan
2019 Nature Communications  
Protein corona presents a major obstacle to bench-to-bedside translation of targeted drug delivery systems, severely affecting targeting yields and directing unfavorable biodistribution. Corona-mediated targeting provides a new impetus for specific drug delivery by precisely manipulating interaction modes of functional plasma proteins on nano-surface. Here bio-inspired liposomes (SP-sLip) were developed by modifying liposomal surface with a short nontoxic peptide derived from Aβ1-42 that
more » ... cally interacts with the lipid-binding domain of exchangeable apolipoproteins. SP-sLip absorb plasma apolipoproteins A1, E and J, consequently exposing receptor-binding domain of apolipoproteins to achieve brain-targeted delivery. Doxorubicin loaded SP-sLip (SP-sLip/DOX) show significant enhancement of brain distribution and anti-brain cancer effect in comparison to doxorubicin loaded plain liposomes. SP-sLip preserve functions of the absorbed human plasma ApoE, and the corona-mediated targeting strategy works in SP modified PLGA nanoparticles. The present study may pave a new avenue to facilitate clinical translation of targeted drug delivery systems.
doi:10.1038/s41467-019-11593-z pmid:31395892 pmcid:PMC6687821 fatcat:ttqsmuz35bfe3d663vukoq4s7y

Fast Single Image Reflection Suppression via Convex Optimization [article]

Yang Yang, Wenye Ma, Yin Zheng, Jian-Feng Cai, Weiyu Xu
2019 arXiv   pre-print
Removing undesired reflections from images taken through the glass is of great importance in computer vision. It serves as a means to enhance the image quality for aesthetic purposes as well as to preprocess images in machine learning and pattern recognition applications. We propose a convex model to suppress the reflection from a single input image. Our model implies a partial differential equation with gradient thresholding, which is solved efficiently using Discrete Cosine Transform.
more » ... e experiments on synthetic and real-world images demonstrate that our approach achieves desirable reflection suppression results and dramatically reduces the execution time.
arXiv:1903.03889v3 fatcat:xmmjkofnj5fwpnswphkypff6e4

Proton irradiation-decelerated intergranular corrosion of Ni-Cr alloys in molten salt

Weiyue Zhou, Yang Yang, Guiqiu Zheng, Kevin B. Woller, Peter W. Stahle, Andrew M. Minor, Michael P. Short
2020 Nature Communications  
The effects of ionizing radiation on materials often reduce to "bad news". Radiation damage usually leads to detrimental effects such as embrittlement, accelerated creep, phase instability, and radiation-altered corrosion. Here we report that proton irradiation decelerates intergranular corrosion of Ni-Cr alloys in molten fluoride salt at 650 °C. We demonstrate this by showing that the depth of intergranular voids resulting from Cr leaching into the salt is reduced by proton irradiation alone.
more » ... nterstitial defects generated from irradiation enhance diffusion, more rapidly replenishing corrosion-injected vacancies with alloy constituents, thus playing the crucial role in decelerating corrosion. Our results show that irradiation can have a positive impact on materials performance, challenging our view that radiation damage usually results in negative effects.
doi:10.1038/s41467-020-17244-y pmid:32647205 fatcat:y7uz4ajnbvfqlijbmauovhdriq

Force field induced heterogeneous molecular orientation within polyacrylonitrile monofilaments

Yang Sha, Wei Liu, Mengfan Wang, Jianglu Wu, Weiyu Cao
2018 RSC Advances  
The evolution of radial structure of polyacrylonitrile monofilament under force field.
doi:10.1039/c8ra06310a pmid:35547686 pmcid:PMC9086294 fatcat:ybbipwdexnf5pkmmxtpg66klmy

Experimental Analysis and Full Prediction Model of a 5-DOF Motorized Spindle

Weiyu Zhang, Huangqiu Zhu, Hengkun Yang, Tao Chen
2017 Energies  
Hengkun Yang and Tao Chen carried out the experiment. Conflicts of Interest: The authors declare no conflict of interest.  ...  Author Contributions: Weiyu Zhang performed the experiments, analyzed the data and proposed the new model. Huangqiu Zhu contributed the experiment platform.  ... 
doi:10.3390/en10010075 fatcat:kkqc2t5nl5g3vjftefzn3sp35q
« Previous Showing results 1 — 15 out of 504 results