Search

Your search keyword '"Feng, Lang"' showing total 284 results
Start Over Author "Feng, Lang"
284 results on '"Feng, Lang"'

1. Diverse Intra- and Inter-Domain Activity Style Fusion for Cross-Person Generalization in Activity Recognition

Author

Zhang, Junru, Feng, Lang, Liu, Zhidan, Wu, Yuhan, He, Yang, Dong, Yabo, and Xu, Duanqing

Subjects
Computer Science - Machine Learning, Computer Science - Artificial Intelligence
Abstract

Existing domain generalization (DG) methods for cross-person generalization tasks often face challenges in capturing intra- and inter-domain style diversity, resulting in domain gaps with the target domain. In this study, we explore a novel perspective to tackle this problem, a process conceptualized as domain padding. This proposal aims to enrich the domain diversity by synthesizing intra- and inter-domain style data while maintaining robustness to class labels. We instantiate this concept using a conditional diffusion model and introduce a style-fused sampling strategy to enhance data generation diversity. In contrast to traditional condition-guided sampling, our style-fused sampling strategy allows for the flexible use of one or more random styles to guide data synthesis. This feature presents a notable advancement: it allows for the maximum utilization of possible permutations and combinations among existing styles to generate a broad spectrum of new style instances. Empirical evaluations on a broad range of datasets demonstrate that our generated data achieves remarkable diversity within the domain space. Both intra- and inter-domain generated data have proven to be significant and valuable, contributing to varying degrees of performance enhancements. Notably, our approach outperforms state-of-the-art DG methods in all human activity recognition tasks., Comment: The 30th ACM SIGKDD Conference on Knowledge Discovery and Data Mining (KDD 2024)

Published
2024

2. Resisting Stochastic Risks in Diffusion Planners with the Trajectory Aggregation Tree

Author

Feng, Lang, Gu, Pengjie, An, Bo, and Pan, Gang

Subjects
Computer Science - Machine Learning, Computer Science - Artificial Intelligence
Abstract

Diffusion planners have shown promise in handling long-horizon and sparse-reward tasks due to the non-autoregressive plan generation. However, their inherent stochastic risk of generating infeasible trajectories presents significant challenges to their reliability and stability. We introduce a novel approach, the Trajectory Aggregation Tree (TAT), to address this issue in diffusion planners. Compared to prior methods that rely solely on raw trajectory predictions, TAT aggregates information from both historical and current trajectories, forming a dynamic tree-like structure. Each trajectory is conceptualized as a branch and individual states as nodes. As the structure evolves with the integration of new trajectories, unreliable states are marginalized, and the most impactful nodes are prioritized for decision-making. TAT can be deployed without modifying the original training and sampling pipelines of diffusion planners, making it a training-free, ready-to-deploy solution. We provide both theoretical analysis and empirical evidence to support TAT's effectiveness. Our results highlight its remarkable ability to resist the risk from unreliable trajectories, guarantee the performance boosting of diffusion planners in $100\%$ of tasks, and exhibit an appreciable tolerance margin for sample quality, thereby enabling planning with a more than $3\times$ acceleration., Comment: ICML 2024 (Spotlight)

Published
2024

3. Temporal Convolutional Explorer Helps Understand 1D-CNN's Learning Behavior in Time Series Classification from Frequency Domain

Author

Zhang, Junru, Feng, Lang, He, Yang, Wu, Yuhan, and Dong, Yabo

Subjects
Computer Science - Machine Learning
Abstract

While one-dimensional convolutional neural networks (1D-CNNs) have been empirically proven effective in time series classification tasks, we find that there remain undesirable outcomes that could arise in their application, motivating us to further investigate and understand their underlying mechanisms. In this work, we propose a Temporal Convolutional Explorer (TCE) to empirically explore the learning behavior of 1D-CNNs from the perspective of the frequency domain. Our TCE analysis highlights that deeper 1D-CNNs tend to distract the focus from the low-frequency components leading to the accuracy degradation phenomenon, and the disturbing convolution is the driving factor. Then, we leverage our findings to the practical application and propose a regulatory framework, which can easily be integrated into existing 1D-CNNs. It aims to rectify the suboptimal learning behavior by enabling the network to selectively bypass the specified disturbing convolutions. Finally, through comprehensive experiments on widely-used UCR, UEA, and UCI benchmarks, we demonstrate that 1) TCE's insight into 1D-CNN's learning behavior; 2) our regulatory framework enables state-of-the-art 1D-CNNs to get improved performances with less consumption of memory and computational overhead., Comment: Accepted as CIKM'23 Long Paper

Published
2023

4. FP3O: Enabling Proximal Policy Optimization in Multi-Agent Cooperation with Parameter-Sharing Versatility

Author

Feng, Lang, Xing, Dong, Zhang, Junru, and Pan, Gang

Subjects
Computer Science - Machine Learning, Computer Science - Artificial Intelligence, Computer Science - Multiagent Systems
Abstract

Existing multi-agent PPO algorithms lack compatibility with different types of parameter sharing when extending the theoretical guarantee of PPO to cooperative multi-agent reinforcement learning (MARL). In this paper, we propose a novel and versatile multi-agent PPO algorithm for cooperative MARL to overcome this limitation. Our approach is achieved upon the proposed full-pipeline paradigm, which establishes multiple parallel optimization pipelines by employing various equivalent decompositions of the advantage function. This procedure successfully formulates the interconnections among agents in a more general manner, i.e., the interconnections among pipelines, making it compatible with diverse types of parameter sharing. We provide a solid theoretical foundation for policy improvement and subsequently develop a practical algorithm called Full-Pipeline PPO (FP3O) by several approximations. Empirical evaluations on Multi-Agent MuJoCo and StarCraftII tasks demonstrate that FP3O outperforms other strong baselines and exhibits remarkable versatility across various parameter-sharing configurations.

Published
2023

5. PREFENDER: A Prefetching Defender against Cache Side Channel Attacks as A Pretender

Author

Li, Luyi, Huang, Jiayi, Feng, Lang, and Wang, Zhongfeng

Subjects
Computer Science - Hardware Architecture
Abstract

Cache side channel attacks are increasingly alarming in modern processors due to the recent emergence of Spectre and Meltdown attacks. A typical attack performs intentional cache access and manipulates cache states to leak secrets by observing the victim's cache access patterns. Different countermeasures have been proposed to defend against both general and transient execution based attacks. Despite their effectiveness, they mostly trade some level of performance for security, or have restricted security scope. In this paper, we seek an approach to enforcing security while maintaining performance. We leverage the insight that attackers need to access cache in order to manipulate and observe cache state changes for information leakage. Specifically, we propose Prefender, a secure prefetcher that learns and predicts attack-related accesses for prefetching the cachelines to simultaneously help security and performance. Our results show that Prefender is effective against several cache side channel attacks while maintaining or even improving performance for SPEC CPU 2006 and 2017 benchmarks., Comment: Submitting to a journal

Published
2023
Full Text
View/download PDF

6. Multi-Level Firing with Spiking DS-ResNet: Enabling Better and Deeper Directly-Trained Spiking Neural Networks

Author

Feng, Lang, Liu, Qianhui, Tang, Huajin, Ma, De, and Pan, Gang

Subjects
Computer Science - Neural and Evolutionary Computing, Computer Science - Artificial Intelligence
Abstract

Spiking neural networks (SNNs) are bio-inspired neural networks with asynchronous discrete and sparse characteristics, which have increasingly manifested their superiority in low energy consumption. Recent research is devoted to utilizing spatio-temporal information to directly train SNNs by backpropagation. However, the binary and non-differentiable properties of spike activities force directly trained SNNs to suffer from serious gradient vanishing and network degradation, which greatly limits the performance of directly trained SNNs and prevents them from going deeper. In this paper, we propose a multi-level firing (MLF) method based on the existing spatio-temporal back propagation (STBP) method, and spiking dormant-suppressed residual network (spiking DS-ResNet). MLF enables more efficient gradient propagation and the incremental expression ability of the neurons. Spiking DS-ResNet can efficiently perform identity mapping of discrete spikes, as well as provide a more suitable connection for gradient propagation in deep SNNs. With the proposed method, our model achieves superior performances on a non-neuromorphic dataset and two neuromorphic datasets with much fewer trainable parameters and demonstrates the great ability to combat the gradient vanishing and degradation problem in deep SNNs., Comment: Accepted by the Thirty-First International Joint Conference on Artificial Intelligence (IJCAI-22)

Published
2022

7. GANDSE: Generative Adversarial Network based Design Space Exploration for Neural Network Accelerator Design

Author

Feng, Lang, Liu, Wenjian, Guo, Chuliang, Tang, Ke, Zhuo, Cheng, and Wang, Zhongfeng

Subjects
Computer Science - Machine Learning, Computer Science - Hardware Architecture
Abstract

With the popularity of deep learning, the hardware implementation platform of deep learning has received increasing interest. Unlike the general purpose devices, e.g., CPU, or GPU, where the deep learning algorithms are executed at the software level, neural network hardware accelerators directly execute the algorithms to achieve higher both energy efficiency and performance improvements. However, as the deep learning algorithms evolve frequently, the engineering effort and cost of designing the hardware accelerators are greatly increased. To improve the design quality while saving the cost, design automation for neural network accelerators was proposed, where design space exploration algorithms are used to automatically search the optimized accelerator design within a design space. Nevertheless, the increasing complexity of the neural network accelerators brings the increasing dimensions to the design space. As a result, the previous design space exploration algorithms are no longer effective enough to find an optimized design. In this work, we propose a neural network accelerator design automation framework named GANDSE, where we rethink the problem of design space exploration, and propose a novel approach based on the generative adversarial network (GAN) to support an optimized exploration for high dimension large design space. The experiments show that GANDSE is able to find the more optimized designs in negligible time compared with approaches including multilayer perceptron and deep reinforcement learning., Comment: Published in ACM Transactions on Design Automation of Electronic Systems

Published
2022
Full Text
View/download PDF

8. Molecular characteristics and pathogenicity of a novel chicken astrovirus variant

Author

Xiaoqing Bi, Zhenrui Song, Fanrun Meng, Shiwei Sun, Xusheng Du, Mengzan Yang, Defang Zhou, Xiangyu Cheng, Longying Ding, Hengyang Shi, Feng Lang, Huaibiao Luan, Bing Deng, Liangyu Yang, and Ziqiang Cheng

Subjects
Chicken astrovirus, molecular characterization, sequence analysis, pathogenic analysis, Veterinary medicine, SF600-1100
Abstract

Abstract It is well-established that the genetic diversity, regional prevalence, and broad host range of astroviruses significantly impact the poultry industry. In July 2022, a small-scale commercial broiler farm in China reported cases of growth retardation and a 3% mortality rate. From chickens displaying proventriculitis and pancreatitis, three chicken astroviruses (CAstV) isolates were obtained and named SDAU2022-1-3. Complete genomic sequencing and analysis revealed the unique characteristics of these isolates from known CAstV strains in ORF1a, ORF1b, and ORF2 genes, characterized by an unusually high variability. Analysis of amino acid mutations in ORF1a, ORF1b, and ORF2 indicated that the accumulation of these mutations played a pivotal role in the emergence of the variant strain. Inoculation experiments demonstrated that affected chickens exhibited liver and kidney enlargement, localized proventricular hemorrhage, and a dark reddish-brown appearance in about two-thirds of the pancreas. Histopathological examination unveiled hepatic lymphocytic infiltration, renal tubular epithelial cell swelling, along with lymphocytic proventriculitis and pancreatitis. Quantitative reverse transcription-polymerase chain reaction (qRT-PCR) analysis indicated viremia and viral shedding at 3 days post-infection (dpi). The proventriculus displayed the highest viral loads, followed by the liver, kidney, duodenum, and pancreas. Liver parameters (AST and ALT) and kidney parameters (UA and UN) demonstrated mild damage consistent with earlier findings. While the possibility of new mutations in the ORF2 gene of CAstV causing proventriculitis and pancreatitis warrants further investigation, these findings deepen our comprehension of CAstV’s pathogenicity in chickens. Additionally, they serve as valuable references for subsequent research endeavors.

Published
2023
Full Text
View/download PDF

9. Can Edu-Metaverse Reshape Virtual Teaching Community (VTC) to Promote Educational Equity? An Exploratory Study

Author

Zhai, Xue-Song, Chu, Xiao-Yan, Chen, Min, Shen, Jun, and Lou, Feng-Lang

Abstract

Education equity has become a critical sustainable development goal worldwide, and it has aroused significant concern among educators. While virtual teaching communities have emerged as an approach to minimizing educational inequity through resource sharing and information exchange, the achievement of deep collaboration among individuals remains an area that requires further study. Metaverse, with its advanced behavior interaction capabilities, has shown the potential in promoting deep collaboration and communication among individuals. However, the pathway toward leveraging Metaverse for enhancing educational equity remains unclear. To address this gap, this study aims to review the concept and educational applications of Metaverse, as well as its evolution to pedagogical interactions within virtual teaching communities. This article, specifically, proposes the possibility of achieving educational equity by establishing a Metaverse-supported virtual community that leverages immersive environments for behavior interaction, employs diversified avatars for affective communication, and implements equitable mechanisms for knowledge creation. Overall, the present study underscores the importance of constructing a warmer virtual teaching community that promotes deep collaboration and equitable knowledge creation. By leveraging the capabilities of Metaverse, educators and learners can work together to create a more inclusive and equitable learning environment.

Published
2023
Full Text
View/download PDF

11. Toward Taming the Overhead Monster for Data-Flow Integrity

Author

Feng, Lang, Huang, Jiayi, Huang, Jeff, and Hu, Jiang

Subjects
Computer Science - Hardware Architecture
Abstract

Data-Flow Integrity (DFI) is a well-known approach to effectively detecting a wide range of software attacks. However, its real-world application has been quite limited so far because of the prohibitive performance overhead it incurs. Moreover, the overhead is enormously difficult to overcome without substantially lowering the DFI criterion. In this work, an analysis is performed to understand the main factors contributing to the overhead. Accordingly, a hardware-assisted parallel approach is proposed to tackle the overhead challenge. Simulations on SPEC CPU 2006 benchmark show that the proposed approach can completely enforce the DFI defined in the original seminal work while reducing performance overhead by 4x, on average., Comment: 24 pages, 16 figures, published in ACM Transactions on Design Automation of Electronic Systems

Published
2021
Full Text
View/download PDF

12. Adacket: ADAptive Convolutional KErnel Transform for Multivariate Time Series Classification

Author

Zhang, Junru, Feng, Lang, Zhang, Haowen, Wu, Yuhan, Dong, Yabo, Goos, Gerhard, Founding Editor, Hartmanis, Juris, Founding Editor, Bertino, Elisa, Editorial Board Member, Gao, Wen, Editorial Board Member, Steffen, Bernhard, Editorial Board Member, Yung, Moti, Editorial Board Member, Koutra, Danai, editor, Plant, Claudia, editor, Gomez Rodriguez, Manuel, editor, Baralis, Elena, editor, and Bonchi, Francesco, editor

Published
2023
Full Text
View/download PDF

14. Research Note: Genetic characterization and pathogenicity of an epidemic variant strain of avian reovirus

Author

Zhenrui Song, Xiaoqing Bi, Fanrun Meng, Qi Yang, Xiangyu Cheng, Defang Zhou, Xinyue Zhang, Longying Ding, Feng Lang, Bing Deng, Kang Wang, and Ziqiang Cheng

Subjects
avian reovirus (ARV), genomic characteristics, phylogenetic analysis, pathogenicity, Animal culture, SF1-1100
Abstract

ABSTRACT: The past few years have witnessed a rapid increase in cases of viral arthritis caused by avian reovirus (ARV) in chicken farms in China, attributed to the emergence of variant strains that render traditional vaccines ineffective, leading to substantial economic losses. In this study, we successfully isolated a novel ARV strain, designated as 2023ARV-GS-SDAU-1, from chickens in a broiler flock vaccinated with an ARV vaccine in Gansu province. We performed whole-genome sequencing and assessed its pathogenicity through 2 infection routes: oral administration and intraperitoneal injection. Our analysis revealed significant variations in the σA gene, associated with the inhibition of interferon secretion, compared to known ARV strains. The highest nucleotide identity observed was below 80%. Additionally, the σC gene exhibited notable variations compared to its homologous strains within the same group. Multiple alignment of the amino acid sequences classified the 2023ARV-GS-SDAU-1 strain under genotype I. Furthermore, our pathogenicity experiments indicated that the isolated strain exhibited more severe pathogenicity when administered via intraperitoneal injection in SPF chickens. In summary, our data suggest that the 2023ARV-GS-SDAU-1 strain represents a novel variant circulating in broiler flocks in China. These findings enrich currently available genetic information on ARV strains and provide a new complete genome sequence.

Published
2024
Full Text
View/download PDF

15. Elevation and deformation extraction from TomoSAR

Author

Feng, Lang

Subjects
616.07
Abstract

3D SAR tomography (TomoSAR) and 4D SAR differential tomography (Diff-TomoSAR) exploit multi-baseline SAR data stacks to provide an essential innovation of SAR Interferometry for many applications, sensing complex scenes with multiple scatterers mapped into the same SAR pixel cell. However, these are still influenced by DEM uncertainty, temporal decorrelation, orbital, tropospheric and ionospheric phase distortion and height blurring. In this thesis, these techniques are explored. As part of this exploration, the systematic procedures for DEM generation, DEM quality assessment, DEM quality improvement and DEM applications are first studied. Besides, this thesis focuses on the whole cycle of systematic methods for 3D & 4D TomoSAR imaging for height and deformation retrieval, from the problem formation phase, through the development of methods to testing on real SAR data. After DEM generation introduction from spaceborne bistatic InSAR (TanDEM-X) and airborne photogrammetry (Bluesky), a new DEM co-registration method with line feature validation (river network line, ridgeline, valley line, crater boundary feature and so on) is developed and demonstrated to assist the study of a wide area DEM data quality. This DEM co-registration method aligns two DEMs irrespective of the linear distortion model, which improves the quality of DEM vertical comparison accuracy significantly and is suitable and helpful for DEM quality assessment. A systematic TomoSAR algorithm and method have been established, tested, analysed and demonstrated for various applications (urban buildings, bridges, dams) to achieve better 3D & 4D tomographic SAR imaging results. These include applying Cosmo-Skymed X band single-polarisation data over the Zipingpu dam, Dujiangyan, Sichuan, China, to map topography; and using ALOS L band data in the San Francisco Bay region to map urban building and bridge. A new ionospheric correction method based on the tile method employing IGS TEC data, a split-spectrum and an ionospheric model via least squares are developed to correct ionospheric distortion to improve the accuracy of 3D & 4D tomographic SAR imaging. Meanwhile, a pixel by pixel orbit baseline estimation method is developed to address the research gaps of baseline estimation for 3D & 4D spaceborne SAR tomography imaging. Moreover, a SAR tomography imaging algorithm and a differential tomography four-dimensional SAR imaging algorithm based on compressive sensing, SAR interferometry phase (InSAR) calibration reference to DEM with DEM error correction, a new phase error calibration and compensation algorithm, based on PS, SVD, PGA, weighted least squares and minimum entropy, are developed to obtain accurate 3D & 4D tomographic SAR imaging results. The new baseline estimation method and consequent TomoSAR processing results showed that an accurate baseline estimation is essential to build up the TomoSAR model. After baseline estimation, phase calibration experiments (via FFT and Capon method) indicate that a phase calibration step is indispensable for TomoSAR imaging, which eventually influences the inversion results. A super-resolution reconstruction CS based study demonstrates X band data with the CS method does not fit for forest reconstruction but works for reconstruction of large civil engineering structures such as dams and urban buildings. Meanwhile, the L band data with FFT, Capon and the CS method are shown to work for the reconstruction of large manmade structures (such as bridges) and urban buildings.

Published
2020

17. In situ tailoring solid electrolyte interphase of three‐dimensional Li metal electrode for enhanced Coulombic efficiency

Author

Jiang‐Peng Wang, Feng Lang, and Quan Li

Subjects
3D host, lithiophilic sites, lithium metal batteries, metal anode, Renewable energy sources, TJ807-830, Environmental sciences, GE1-350
Abstract

Abstract Although three‐dimensional (3D) lithium metal electrode is effective in restricting the Li dendrite growth upon cycling, problems associated with the unstable electrode/electrolyte interphase become more severe due to increased interfacial area that is intrinsic of the 3D structures, being a major cause for the low Columbic efficiency. While building a desirable solid electrolyte interphase (SEI) serves as an effective solution to improve the electrode/electrolyte interfacial stability, the 3D nature of the electrode makes the task challenging. In the present work, we demonstrated the in‐situ formation of SEI on chemically/structurally modified carbon cloth that is used as the 3D host electrode for Li metal. Here we show that ZnS/ZnO nanotube arrays uniformly grown on the carbon cloth served as precursors for the in‐situ formation of Li2S/Li2O/LiZn containing artificial SEI in the first lithiation process. While Li2S and Li2O are preferred components in SEI, the in situ generated Zn functions as a lithiophilic site that guides the uniform lithium deposition upon repeated charging/discharging process. As a result, symmetric cells adopting the O‐, S‐, and Zn‐ modified 3D anode demonstrate significantly improved Coulombic efficiency (99.2% over 400 cycles at 1 mA cm−2/1 mA h cm−2). Furthermore, the Li/ZSONT/CC//LiFePO4 full cell shows a capacity retention of 71% after 4000 cycles at 2C. The present work sheds light on effective design strategies for SEI formation on a 3D electrode host with controllable SEI composition.

Published
2023
Full Text
View/download PDF

20. Determination of Synephrine and Caffeine in Citrus Species Herbal Medicines and Weight-Loss Supplements from Taiwan by High-Performance Liquid Chromatography (HPLC).

Author

Chang, William Chih-Wei, Wu, Yu-Tse, Lin, Chia-En, Lin, Feng-Lang, and Chen, Fu-An

Subjects
HIGH performance liquid chromatography, HERBAL medicine, DIETARY supplements, CHINESE medicine, RUTACEAE
Abstract

Synephrine is widely found in Chinese herbs of the Rutaceae family and is commonly used as a primary substitute for ephedrine in sports supplements. However, its potential cardiotoxicity has necessitated governmental regulations regarding its use. This study determined the contents of synephrine, phenylephrine, and caffeine in 30 Chinese herbal preparations (including five single herbs) and 16 supplements promoted for weight loss. The presence of naturally occurring synephrine was detected in all Chinese herbal medicines, with immature bitter orange having the highest synephrine content (ranging from 233.1 to 505.3 mg/100 g). This was followed by three herbs with comparable amounts of up to 317.7 mg/100 g, namely bitter orange, red tangerine exocarp, and tangerine peel. Euodia fruit had the lowest synephrine content, measuring up to 4.9 mg/100 g. None of these herbal medicines were found to contain detectable phenylephrine. On the other hand, weight-loss supplements, comprising complex combinations of herbal extracts, were found to contain synephrine in concentrations up to 1159.0 mg/100 g and caffeine up to 1659.1 mg/100 g. The synephrine content of these products did not exceed the regulatory limit of six percent, nor did they surpass the respective daily consumption of 20 mg. However, one product was found to contain both synephrine and caffeine, albeit at low concentrations, which could potentially breach the regulation prohibiting the combined use of bitter orange and caffeine as food ingredients. [ABSTRACT FROM AUTHOR]

Published
2024
Full Text
View/download PDF

21. Association between air pollution exposure and coronary heart disease hospitalization in a humid sub-tropical region of China: A time-series study

Author

Ya-Ting Feng, Cui-Feng Lang, Cong Chen, Musonye Harry Asena, Yang Fang, Ruo-Di Zhang, Ling-Qiong Jiang, Xi Fang, Yue Chen, Yi-Sheng He, Peng Wang, and Hai-Feng Pan

Subjects
air pollutants, CHD, NO2, CO, O3, disease risk, Public aspects of medicine, RA1-1270
Abstract

ObjectiveEmerging evidence has highlighted the possible links of environmental pollution with several cardiovascular diseases (CVDs). The current study aimed to explore the impact of short-term air pollution exposure on CHD hospitalization in Hefei.MethodsData about the daily number of CHD admissions (from 2014 to 2021) were retrieved from the First Affiliated Hospital of Anhui Medical University. Air pollutants and meteorological data were obtained from the China Environmental Monitoring Station and the China Meteorological Data Service Center, respectively. The correlation between air pollution and CHD hospitalization was assessed using distributed lag non-linear model (DLNM) and Poisson generalized linear regression.ResultsIn the single-pollutant model, NO2, O3, and CO strongly correlated with CHD hospitalization rate. Specifically, exposure to NO2 (lag0, relative risk [RR]: 1.013, 95%CI: 1.002–1.024, per 10 μg/m3 increase) and CO (lag13, RR: 1.035, 95%CI: 1.001–1.071, per 1 μg/m3 increase) revealed a positive correlation with an increased rate of CHD hospitalization. Interestingly, O3 had a protective association with hospitalization of CHD (lag0, RR: 0.993, 95%CI: 0.988–0.999, per 10 μg/m3 increase). Similar results, to those of the single-pollutant model, were revealed following verification using two-pollutant models. Subgroup analyses indicated that young people, women, and people in hot seasons were more susceptible to NO2 exposure, while the elderly, women, and people in cold seasons were more susceptible to O3. Furthermore, the elderly were more susceptible to CO exposure.ConclusionOverall, exposure to NO2 and CO increases the rate of CHD hospitalization, but exposure to O3 shows a protective association with the rate of CHD hospitalization. Therefore, early preventive measures against air pollutants should be applied to protect vulnerable patients with CHD.

Published
2023
Full Text
View/download PDF

23. FastCFI: Real-Time Control Flow Integrity Using FPGA Without Code Instrumentation

Author

Feng, Lang, Huang, Jeff, Hu, Jiang, Reddy, Abhijith, Goos, Gerhard, Founding Editor, Hartmanis, Juris, Founding Editor, Bertino, Elisa, Editorial Board Member, Gao, Wen, Editorial Board Member, Steffen, Bernhard, Editorial Board Member, Woeginger, Gerhard, Editorial Board Member, Yung, Moti, Editorial Board Member, Finkbeiner, Bernd, editor, and Mariani, Leonardo, editor

Published
2019
Full Text
View/download PDF

28. Specificity, flexibility and valence of DNA bonds guide emulsion architecture

Author

Feng, Lang, Pontani, Lea-Laetitia, Dreyfus, Remi, Chaikin, Paul, and Brujic, Jasna

Subjects
Condensed Matter - Soft Condensed Matter, Condensed Matter - Materials Science
Abstract

The specificity and thermal reversibility of DNA interactions have enabled the self-assembly of crystal structures, self-replicating materials and colloidal molecules. Grafting DNA onto liquid interfaces of emulsions leads to exciting new architectural possibilities due to the mobility of the DNA ligands and the patches they form between bound droplets. Here we show that the size and number of these adhesion patches (valency) can be controlled. Valence 2 leads to flexible polymers of emulsion droplets, while valence above 4 leads to rigid droplet networks. A simple thermodynamic model quantitatively describes the increase in the patch size with droplet radii, DNA concentration and the stiffness of the tether to the sticky-end. The patches are formed between droplets with complementary DNA strands or alternatively with complementary colloidal nanoparticles to mediate DNA binding between droplets. This emulsion system opens the route to directed self-assembly of more complex structures through distinct DNA bonds with varying strengths and controlled valence and flexibility., Comment: This paper has been withdrawn by the author due to publication issue

Published
2013
Full Text
View/download PDF

29. Electrification Mechanism and Influence Factors of Sliding Triboelectric Nanogenerators Based on Micro‐Contact Model.

Author

Li, Tao, Long, Wei, Xu, Xuhui, Feng, Lang, and Chen, Ye

Subjects
NANOGENERATORS, ELECTRIFICATION, INTERFACIAL friction, SLIDING friction, ELECTRON density, RURAL electrification
Abstract

As a new green energy‐harvesting technology, the electrification mechanism and output performance of triboelectric nanogenerators (TENGs) have become the research focuses and difficulties. Based on the basic principles of contact and damage mechanics, a five‐stage micro‐contact model is proposed on the interface of sliding TENGs friction subsurface. Experimental bench is built to complete the experimental validation and factors affecting the output performance of TENGs are experimentally investigated. In the results, it is shown that the contact form of the asperities on the surface of the friction subsurface determines the real contact area and the form of energy conversion, and the dynamical processes between the asperities are influenced by the velocity, load, and material, which further affect the elastic/plastic change of the friction surface and the directional transfer of contact electrons. Normal loading promotes the density and efficiency of electron transfer between the interfaces of the friction sub‐interfaces, but the electrical output performance of the system is no longer enhanced when the load is too high (≥10 N). The faster sliding speed is conducive to improving the output performance of the horizontal sliding TENG in a certain range (0.005–0.100 m s−1). [ABSTRACT FROM AUTHOR]

Published
2024
Full Text
View/download PDF

34. A hybrid approach to resolve ambiguity in new product development.

Author

Liu, Feng-Lang, Chien, Li-Chih, Chang, Ting-Yu, Ku, Cooper Cheng-Yuan, and Chang, Ching-Ter

Subjects
*NEW product development, *AMBIGUITY, *ANALYTIC hierarchy process, *PROSPECT theory, *DELPHI method, *TECHNOLOGICAL innovations
Abstract

Improving technological innovation (TI) capabilities is an integral component of government policies aimed at improving the competitiveness of small and medium enterprises (SMEs). This study aims to address implementation challenges arising from the use of Qualitative Forecasting Method (QFM) in new product development programs and proposes a novel method to aid decision makers (DMs) in their decision-making process. To tackle this issue, a hybrid method is proposed, incorporating Fuzzy Delphi method (FDM), Fuzzy Analytic Hierarchy Process (FAHP), Technique for Order Preference by Similarity to an Ideal Solution (TOPSIS), and Multi-Choice Goal Programming with utility function (MCGP-U), while introducing prospect theory as a novel approach. is proposed. The proposed method offers several advantages, including effective early planning, accurate identification of key success factors (KSFs), selection of the most suitable project leader, and estimation of the most reasonable resource investment, all of which are critical factors for success in TI for enterprises. The research results show that (1) the proposed method reduces project execution time by 20% compared to the original manual planning, (2) it facilitates the acquisition of KSFs using a rational approach to ensure project success, and (3) it increases the financial returns of the company by 17% compared to the company's forecast. In summary, this paper makes a significant contribution to practical applications and additionally contributes to decision-making field by introducing prospect theory into the proposed hybrid method. [ABSTRACT FROM AUTHOR]

Published
2024
Full Text
View/download PDF

38. Static Analysis Complements Machine Learning: A Type Inference Use Case

Author

feng, lang (author) and feng, lang (author)

Abstract

Type inference plays a pivotal role in modern software development as it aids in understanding code, detecting errors, and facilitating code completion. Two main approaches, static analysis, and machine learning, contribute to this process. Each approach has its own benefits and limitations. This thesis investigates the potential of combining static analysis techniques and machine learning (ML) approaches to enhance type inference capabilities., https://zenodo.org/record/8255564 Dataset URL, Computer Science | Software Technology

Published
2023

39. In situ tailoring solid electrolyte interphase of three‐dimensional Li metal electrode for enhanced Coulombic efficiency

Author

Jiangpeng Wang, FENG LANG, and Quan Li

Subjects
Chemistry (miscellaneous), Materials Science (miscellaneous), Physical and Theoretical Chemistry
Abstract

Although three-dimensional (3D) host is effective in restricting Li dendrite growth, problems associated with the unstable electrode/electrolyte interphase becomes more severe due to increased interfacial area that is intrinsic of the 3D structures, being a major cause for the low Coulombic efficiency.While building a desirable solid electrolyte interphase (SEI) serves as an effective solution to improve the electrode/electrolyte interfacial stability, the 3D nature of the electrode makes the task challenging.Herein,we demonstrated the in-situ formation of SEI on chemically/structurally modified carbon cloth that is used as the 3D host.We shown that ZnS/ZnO nanotube arrays uniformly grown on the carbon cloth served as precursors for the in-situ formation of Li2S/Li2O/LiZn containing artificial SEI. While Li2S and Li2O are preferred components in SEI, Zn functions as lithiophilic site that guides the uniform lithium deposition.The present work shed light on effective design strategies for SEI formation on 3D electrode host with controllable SEI composition.

Published
2023

41. Study on the tribological properties of Si3N4/polyamide-imide composites

Author

Weichu Jia, Jingfei Shangguan, Feng Lang, Duxin Li, and Haiyang Yang

Subjects
Polymers and Plastics, Organic Chemistry, Materials Chemistry
Abstract

Si3N4 whiskers were modified with triethoxysilane as a means to improve the interfacial interaction with a poly-amide imide (PAI). The Si3N4/PAI composites were prepared by in-situ polymerization as an approach to improve the friction behavior and mechanical properties in order to meet the performance needs of electrical transformers and motors in harsh environments. 4,4′-diamino diphenyl ether, m-phenylenediamine, triethylamine, trimellitic anhydride chloride, and Si3N4 whiskers were used to manufacture Si3N4/PAI composite, which was then characterized for friction and mechanical behavior. The research showed that the addition of Si3N4 whiskers improved the mechanical properties and friction properties of PAI. When the whisker content was 10 wt%, the tensile strength of the composite was 137.5 MPa, which is 15% higher than pure PAI, and the wear rate was 0.69 × 10-8 g/N·m, which is 57% lower than pure PAI, and the friction coefficient was 0.47, which was an increase of 9%. In aerospace, machinery, and transportation applications, these materials may have better performance due to their improved combination of properties.

Published
2022

42. Preparation and properties of MDA-BAPP-BTDA copolyimide/18-crown ether-6 supramolecular films with inclusion structure and ultralow dielectric constants

Author

Chunguang Xiao, Duxin Li, Feng Lang, Yu Xiang, Yi Lin, and Baoli Ou

Subjects
copolyimide, supramolecular films, crown ether, inclusion structure, dielectric constant, mechanical properties, Materials of engineering and construction. Mechanics of materials, TA401-492, Chemical technology, TP1-1185
Abstract

In order to obtain polyimide composites with ultra-low dielectric constant, a series of MDA-BAPP-BTDA copolyimides/18-crown ether-6 (CPI/18-CE) supramolecular films with inclusion structure were prepared. The effects of 18-CE on CPI’s thermal, mechanical, dielectric and water absorption properties were investigated. The inclusion rate of 18-CE to CPI backbone was ∼50%. The introduction of 18-CE slightly reduced the thermal properties of CPI/18-CE supramolecular films, but greatly improved their mechanical, dielectric and hydrophobic properties. The tensile strength, young’s modulus and elongation at break of CPI/18-CE supramolecular films were maximally increased by 21.8%, 34.1% and 92.9% respectively. Meanwhile their dielectric constant, dielectric loss and water absorption were as low as 2.10, 0.007 and 0.63%, respectively. In summary, CPI/18-CE supramolecular films possess excellent dielectric properties and broad application prospect in the field of microelectronics.

Published
2021
Full Text
View/download PDF

43. Preparation and properties of MoS2 modified polydimethylsiloxane/monomer casting nylon

Author

Jinfei Shangguan, Ling Song, Feng Lang, Haiyang Yang, Duxin Li, Jin Wang, Qibin Jiang, and Yilan You

Subjects
MC nylon, polydimethylsiloxane, MoS2, friction performance, Materials of engineering and construction. Mechanics of materials, TA401-492, Chemical technology, TP1-1185
Abstract

MoS _2 reinforced PDMS/MC nylon composites had been fabricated by in situ polymerization to further enhance the comprehensive properties of nylon material. Polydimethylsiloxane was pre-sythesized as the activator of the system and MoS _2 was compounded as the wear-resisting agent. The thermal stability, mechanical properties, friction and wear properties of composites with different MoS _2 contents were tested. The results showed that the crystallization and thermal stability of nylon matrix were increased with the addition of MoS _2 , while the tensile strength and impact strength were slightly reduced. The composites finally presented excellent friction properties with 3 wt% MoS _2 , whose friction coefficient was reduced to 0.75 (18.48% decrease) corresponding to the wear rate of 0.4145 × 10 ^−8 g Nm ^−1 .

Published
2021
Full Text
View/download PDF

44. RISC-V Custom Instructions of Elementary Functions for IoT Endpoint Devices

Author

Chen, Yuxing, Wang, Xinrui, Song, Suwen, Feng, Lang, and Wang, Zhongfeng

Abstract

The computation of elementary functions is required in many tasks of Internet of Things (IoT) endpoint devices, for example, communications, image processing, and biomedical signal processing. IoT endpoint devices generally adopt software approaches to compute elementary functions, which take many cycles. To improve efficiency, this work proposes custom instructions for elementary functions to the open-source RISC-V instruction set architecture (ISA). In particular, several variants of the custom instructions (fast, intermediate, and tiny variants) are developed to satisfy the needs of various types of IoT devices. Microarchitecture design and VLSI circuit design are then proposed to efficiently support the extended ISA. Both software emulation and on-board evaluation of the new architecture are carried out with testbenches covering typical communication and computation tasks for IoT devices. The custom instructions gain speedups ranging from 3.3 to 18.0 compared to a baseline RV32IM design. ASIC synthesis results under TSMC 28nm technology demonstrate that the power overhead is $ \lt $< 5% with the tiny variant, $ \lt $< 17% with the intermediate variant, and $ \lt $< 26% with the fast variant, which is not significant considering the achieved speedup. The experimental results further confirm that the proposed custom instructions are computation-efficient and versatile to adapt to different IoT devices for various applications.

Published
2024
Full Text
View/download PDF

45. A Fast Transient LDO Regulator Featuring High PSRR Over 100-kHz Frequency Range With Adaptive, Dynamic Biasing, and Current Mode Feed-Forward Amplifier

Author

Fan, Hua, Feng, Lang, Diao, Xiaopeng, Xie, Xiuhua, Wang, Ce, Li, Guo, Wei, Qi, Qiao, Fei, Feng, Quanyuan, and Bonizzoni, Edoardo

Abstract

This brief presents a fast transient LDO with high power supply rejection ratio (PSRR) over 100-kHz based on adaptive biasing, dynamic biasing technique and a current mode feed-forward amplifier (CMFFA). The dynamic biasing improves the load transient response and the adaptive biasing benefits the loop stability. The CMFFA introduces a left-half-plane zero to compensate for the non-dominant pole without large current consumption. Moreover, a impedance adaptive circuit is also used in this design: it ensures the DC gain and pushes a non-dominant pole to high frequency. Through these methods, the proposed LDO is stable over full load range, from 0mA to 150mA, it achieves fast transient response and high PSRR with a low quiescent current. The LDO is fabricated in a 0.6- $\mu $ m CMOS technology. The output voltage can be regulated from 1.8V to 3.3V and the load capacitance is 1 $\mu $ F. For a 150mA load step and a 3.3V output voltage, the maximum undershoot voltage is 38.6mV. At maximum load condition, a 41dB PSRR is achieved at 100-kHz and the loop gain bandwidth product is 1.9MHz. The DC gain is around 70dB over the full load range. The measured load regulation and line regulation are 0.06mV/mA and 1mV/V, respectively. The LDO has a minimum quiescent current of 9.6 $\mu $ A without load. Finally, the proposed LDO achieves a FOM1 of 16.47ps and a FOM2 of 123.52 $\mu $ V.

Published
2024
Full Text
View/download PDF

46. Prefender: A <underline>Pref</underline>etching Def<underline>en</underline>der Against Cache Side Channel Attacks as a Preten<underline>der</underline>

Author

Li, Luyi, Huang, Jiayi, Feng, Lang, and Wang, Zhongfeng

Abstract

Cache side channel attacks are increasingly alarming in modern processors due to the recent emergence of Spectre and Meltdown attacks. A typical attack performs intentional cache access and manipulates cache states to leak secrets by observing the victim's cache access patterns. Different countermeasures have been proposed to defend against both general and transient execution based attacks. Despite their effectiveness, they mostly trade some level of performance for security, or have restricted security scope. In this paper, we seek an approach to enforcing security while maintaining performance. We leverage the insight that attackers need to access cache in order to manipulate and observe cache state changes for information leakage. Specifically, we propose Prefender, a secure prefetcher that learns and predicts attack-related accesses for prefetching the cachelines to simultaneously help security and performance. Our results show that Prefender is effective against several cache side channel attacks while maintaining or even improving performance for SPEC CPU 2006 and 2017 benchmarks.

Published
2024
Full Text
View/download PDF

49. Array focused Lamb wave Hilbert holographic imaging method

Author

Yan-Feng Lang, Shao-Hua Tian, Zhi-Bo Yang, Ruo-Bin Sun, Wei Zhang, De-Tong Kong, and Xue-Feng Chen

Subjects
Mechanics of Materials, Mechanical Engineering, General Mathematics, General Materials Science, Civil and Structural Engineering
Published
2022

Catalog

Books, media, physical & digital resources
See catalog results