Search

Your search keyword '"Ran, Ran"' showing total 1,753 results
Start Over Author "Ran, Ran" Search Limiters Available in Library Collection
1,753 results on '"Ran, Ran"'

1. An Evidential-enhanced Tri-Branch Consistency Learning Method for Semi-supervised Medical Image Segmentation

Author

Zhang, Zhenxi, Zhou, Heng, Shi, Xiaoran, Ran, Ran, Tian, Chunna, and Zhou, Feng

Subjects
Computer Science - Computer Vision and Pattern Recognition
Abstract

Semi-supervised segmentation presents a promising approach for large-scale medical image analysis, effectively reducing annotation burdens while achieving comparable performance. This methodology holds substantial potential for streamlining the segmentation process and enhancing its feasibility within clinical settings for translational investigations. While cross-supervised training, based on distinct co-training sub-networks, has become a prevalent paradigm for this task, addressing critical issues such as predication disagreement and label-noise suppression requires further attention and progress in cross-supervised training. In this paper, we introduce an Evidential Tri-Branch Consistency learning framework (ETC-Net) for semi-supervised medical image segmentation. ETC-Net employs three branches: an evidential conservative branch, an evidential progressive branch, and an evidential fusion branch. The first two branches exhibit complementary characteristics, allowing them to address prediction diversity and enhance training stability. We also integrate uncertainty estimation from the evidential learning into cross-supervised training, mitigating the negative impact of erroneous supervision signals. Additionally, the evidential fusion branch capitalizes on the complementary attributes of the first two branches and leverages an evidence-based Dempster-Shafer fusion strategy, supervised by more reliable and accurate pseudo-labels of unlabeled data. Extensive experiments conducted on LA, Pancreas-CT, and ACDC datasets demonstrate that ETC-Net surpasses other state-of-the-art methods for semi-supervised segmentation. The code will be made available in the near future at https://github.com/Medsemiseg.

Published
2024

3. LinGCN: Structural Linearized Graph Convolutional Network for Homomorphically Encrypted Inference

Author

Peng, Hongwu, Ran, Ran, Luo, Yukui, Zhao, Jiahui, Huang, Shaoyi, Thorat, Kiran, Geng, Tong, Wang, Chenghong, Xu, Xiaolin, Wen, Wujie, and Ding, Caiwen

Subjects
Computer Science - Machine Learning, Computer Science - Artificial Intelligence, Computer Science - Cryptography and Security, E.3, I.2, B.0
Abstract

The growth of Graph Convolution Network (GCN) model sizes has revolutionized numerous applications, surpassing human performance in areas such as personal healthcare and financial systems. The deployment of GCNs in the cloud raises privacy concerns due to potential adversarial attacks on client data. To address security concerns, Privacy-Preserving Machine Learning (PPML) using Homomorphic Encryption (HE) secures sensitive client data. However, it introduces substantial computational overhead in practical applications. To tackle those challenges, we present LinGCN, a framework designed to reduce multiplication depth and optimize the performance of HE based GCN inference. LinGCN is structured around three key elements: (1) A differentiable structural linearization algorithm, complemented by a parameterized discrete indicator function, co-trained with model weights to meet the optimization goal. This strategy promotes fine-grained node-level non-linear location selection, resulting in a model with minimized multiplication depth. (2) A compact node-wise polynomial replacement policy with a second-order trainable activation function, steered towards superior convergence by a two-level distillation approach from an all-ReLU based teacher model. (3) an enhanced HE solution that enables finer-grained operator fusion for node-wise activation functions, further reducing multiplication level consumption in HE-based inference. Our experiments on the NTU-XVIEW skeleton joint dataset reveal that LinGCN excels in latency, accuracy, and scalability for homomorphically encrypted inference, outperforming solutions such as CryptoGCN. Remarkably, LinGCN achieves a 14.2x latency speedup relative to CryptoGCN, while preserving an inference accuracy of 75% and notably reducing multiplication depth., Comment: NeurIPS 2023 accepted publication

Published
2023

4. Implicit Neural Feature Fusion Function for Multispectral and Hyperspectral Image Fusion

Author

Deng, ShangQi, Wu, RuoCheng, Deng, Liang-Jian, Ran, Ran, and Vivone, Gemine

Subjects
Computer Science - Computer Vision and Pattern Recognition
Abstract

Multispectral and Hyperspectral Image Fusion (MHIF) is a practical task that aims to fuse a high-resolution multispectral image (HR-MSI) and a low-resolution hyperspectral image (LR-HSI) of the same scene to obtain a high-resolution hyperspectral image (HR-HSI). Benefiting from powerful inductive bias capability, CNN-based methods have achieved great success in the MHIF task. However, they lack certain interpretability and require convolution structures be stacked to enhance performance. Recently, Implicit Neural Representation (INR) has achieved good performance and interpretability in 2D tasks due to its ability to locally interpolate samples and utilize multimodal content such as pixels and coordinates. Although INR-based approaches show promise, they require extra construction of high-frequency information (\emph{e.g.,} positional encoding). In this paper, inspired by previous work of MHIF task, we realize that HR-MSI could serve as a high-frequency detail auxiliary input, leading us to propose a novel INR-based hyperspectral fusion function named Implicit Neural Feature Fusion Function (INF). As an elaborate structure, it solves the MHIF task and addresses deficiencies in the INR-based approaches. Specifically, our INF designs a Dual High-Frequency Fusion (DHFF) structure that obtains high-frequency information twice from HR-MSI and LR-HSI, then subtly fuses them with coordinate information. Moreover, the proposed INF incorporates a parameter-free method named INR with cosine similarity (INR-CS) that uses cosine similarity to generate local weights through feature vectors. Based on INF, we construct an Implicit Neural Fusion Network (INFN) that achieves state-of-the-art performance for MHIF tasks of two public datasets, \emph{i.e.,} CAVE and Harvard. The code will soon be made available on GitHub.

Published
2023

5. PASNet: Polynomial Architecture Search Framework for Two-party Computation-based Secure Neural Network Deployment

Author

Peng, Hongwu, Zhou, Shanglin, Luo, Yukui, Xu, Nuo, Duan, Shijin, Ran, Ran, Zhao, Jiahui, Wang, Chenghong, Geng, Tong, Wen, Wujie, Xu, Xiaolin, and Ding, Caiwen

Subjects
Computer Science - Cryptography and Security, E.3, I.2, B.0
Abstract

Two-party computation (2PC) is promising to enable privacy-preserving deep learning (DL). However, the 2PC-based privacy-preserving DL implementation comes with high comparison protocol overhead from the non-linear operators. This work presents PASNet, a novel systematic framework that enables low latency, high energy efficiency & accuracy, and security-guaranteed 2PC-DL by integrating the hardware latency of the cryptographic building block into the neural architecture search loss function. We develop a cryptographic hardware scheduler and the corresponding performance model for Field Programmable Gate Arrays (FPGA) as a case study. The experimental results demonstrate that our light-weighted model PASNet-A and heavily-weighted model PASNet-B achieve 63 ms and 228 ms latency on private inference on ImageNet, which are 147 and 40 times faster than the SOTA CryptGPU system, and achieve 70.54% & 78.79% accuracy and more than 1000 times higher energy efficiency., Comment: DAC 2023 accepeted publication, short version was published on AAAI 2023 workshop on DL-Hardware Co-Design for AI Acceleration: RRNet: Towards ReLU-Reduced Neural Network for Two-party Computation Based Private Inference

Published
2023

6. Cross-supervised Dual Classifiers for Semi-supervised Medical Image Segmentation

Author

Zhang, Zhenxi, Ran, Ran, Tian, Chunna, Zhou, Heng, Yang, Fan, Li, Xin, and Jiao, Zhicheng

Subjects
Computer Science - Computer Vision and Pattern Recognition
Abstract

Semi-supervised medical image segmentation offers a promising solution for large-scale medical image analysis by significantly reducing the annotation burden while achieving comparable performance. Employing this method exhibits a high degree of potential for optimizing the segmentation process and increasing its feasibility in clinical settings during translational investigations. Recently, cross-supervised training based on different co-training sub-networks has become a standard paradigm for this task. Still, the critical issues of sub-network disagreement and label-noise suppression require further attention and progress in cross-supervised training. This paper proposes a cross-supervised learning framework based on dual classifiers (DC-Net), including an evidential classifier and a vanilla classifier. The two classifiers exhibit complementary characteristics, enabling them to handle disagreement effectively and generate more robust and accurate pseudo-labels for unlabeled data. We also incorporate the uncertainty estimation from the evidential classifier into cross-supervised training to alleviate the negative effect of the error supervision signal. The extensive experiments on LA and Pancreas-CT dataset illustrate that DC-Net outperforms other state-of-the-art methods for semi-supervised segmentation. The code will be released soon., Comment: 13 pages, 4 figures, 5 tables. Code will come soon

Published
2023

7. Self-aware and Cross-sample Prototypical Learning for Semi-supervised Medical Image Segmentation

Author

Zhang, Zhenxi, Ran, Ran, Tian, Chunna, Zhou, Heng, Li, Xin, Yang, Fan, and Jiao, Zhicheng

Subjects
Computer Science - Computer Vision and Pattern Recognition
Abstract

Consistency learning plays a crucial role in semi-supervised medical image segmentation as it enables the effective utilization of limited annotated data while leveraging the abundance of unannotated data. The effectiveness and efficiency of consistency learning are challenged by prediction diversity and training stability, which are often overlooked by existing studies. Meanwhile, the limited quantity of labeled data for training often proves inadequate for formulating intra-class compactness and inter-class discrepancy of pseudo labels. To address these issues, we propose a self-aware and cross-sample prototypical learning method (SCP-Net) to enhance the diversity of prediction in consistency learning by utilizing a broader range of semantic information derived from multiple inputs. Furthermore, we introduce a self-aware consistency learning method that exploits unlabeled data to improve the compactness of pseudo labels within each class. Moreover, a dual loss re-weighting method is integrated into the cross-sample prototypical consistency learning method to improve the reliability and stability of our model. Extensive experiments on ACDC dataset and PROMISE12 dataset validate that SCP-Net outperforms other state-of-the-art semi-supervised segmentation methods and achieves significant performance gains compared to the limited supervised training. Our code will come soon., Comment: 14 pages, Early accepted in MICCAI 2023, code will be released soon

Published
2023

8. DDRF: Denoising Diffusion Model for Remote Sensing Image Fusion

Author

Cao, ZiHan, Cao, ShiQi, Wu, Xiao, Hou, JunMing, Ran, Ran, and Deng, Liang-Jian

Subjects
Computer Science - Computer Vision and Pattern Recognition, Computer Science - Artificial Intelligence, Electrical Engineering and Systems Science - Image and Video Processing
Abstract

Denosing diffusion model, as a generative model, has received a lot of attention in the field of image generation recently, thanks to its powerful generation capability. However, diffusion models have not yet received sufficient research in the field of image fusion. In this article, we introduce diffusion model to the image fusion field, treating the image fusion task as image-to-image translation and designing two different conditional injection modulation modules (i.e., style transfer modulation and wavelet modulation) to inject coarse-grained style information and fine-grained high-frequency and low-frequency information into the diffusion UNet, thereby generating fused images. In addition, we also discussed the residual learning and the selection of training objectives of the diffusion model in the image fusion task. Extensive experimental results based on quantitative and qualitative assessments compared with benchmarks demonstrates state-of-the-art results and good generalization performance in image fusion tasks. Finally, it is hoped that our method can inspire other works and gain insight into this field to better apply the diffusion model to image fusion tasks. Code shall be released for better reproducibility.

Published
2023

9. Numerical simulation and optimization of extrusion runner of thermoplastic polyurethane limit spring based on response surface method

Author

ZHU Jin-mei, LI Gao-wei, LI Jia, JIAN Ran-ran, MIAO Qing, ZENG Xian-kui

Subjects
thermoplastic polyurethane, response surface method, rheological property, extrusion runner, numerical simulation, standard deviation of velocity, structure parameter, Organic chemistry, QD241-441, Chemical engineering, TP155-156, Chemicals: Manufacture, use, etc., TP200-248
Abstract

With thermoplastic polyurethane (TPU) limit spring extruder head as the research object, the relevant material parameters of the constitutive equation of raw TPU were obtained by the rheological property test. Numerical simulation of the flow process of TPU in the extrusion runner was carried out based on the software Polyflow, and in the meanwhile, a second-order regression model of standard deviation of velocity (Sd) at the exit of the forming section was established by response surface method, and the influence law of extrusion runner structure on the melt flow uniformity of TPU was explored. The results showed that after optimization the structure parameters of extrusion runner were as follows: the length of the forming section of 165 mm, the expansion angle of 25°, and the number of diverter shuttles of 6. After optimization the predicted value of Sd at the exit of the forming section of extrusion runner was 0.531 m/s and was only 0.188% different from the actual value of Sd(0.532 m/s), which indicated that the model was more accurate, and the model could be used for the prediction and optimization of the structure parameters of extrusion runner.

Published
2024
Full Text
View/download PDF

10. Application of back propagation neural network optimized by different algorithms in prediction of Mooney viscosity of ethylene-propylene-diene monomer compound

Author

LI Gao-wei, LI Jia, ZHU Jin-mei, JIAN Ran-ran, MIAO Qing, ZENG Xian-kui

Subjects
back propagation neural network, genetic algorithm, particle swarm optimization, ethylene-propylene-diene monomer, compound, mooney viscosity, forecast model, Organic chemistry, QD241-441, Chemical engineering, TP155-156, Chemicals: Manufacture, use, etc., TP200-248
Abstract

Genetic algorithm (GA) and particle swarm optimization (PSO) were used to optimize the back propagation（BP） neural network to establish the prediction model of Mooney viscosity of ethylene-propylene-diene monomer (EPDM) compound, and the error of the prediction results was compared and analyzed. The results showed that the predicted va-lues of the BP neural network model optimized by the two algorithms all maintained a high degree of fit and correlation with the measured values. Compared with the single BP neural network, the accuracy of the GA-BP neural network prediction model increased by 58.9% and the accuracy of the PSO-BP model increased by 3.57%, which indicated that the prediction accuracy of the prediction model optimized by the two algorithms, especially the BP neural network prediction model optimized by GA, improved significantly Mooney viscosity of EPDM compound.

Published
2024
Full Text
View/download PDF

15. RRNet: Towards ReLU-Reduced Neural Network for Two-party Computation Based Private Inference

Author

Peng, Hongwu, Zhou, Shanglin, Luo, Yukui, Xu, Nuo, Duan, Shijin, Ran, Ran, Zhao, Jiahui, Huang, Shaoyi, Xie, Xi, Wang, Chenghong, Geng, Tong, Wen, Wujie, Xu, Xiaolin, and Ding, Caiwen

Subjects
Computer Science - Cryptography and Security, Computer Science - Machine Learning, I.2
Abstract

The proliferation of deep learning (DL) has led to the emergence of privacy and security concerns. To address these issues, secure Two-party computation (2PC) has been proposed as a means of enabling privacy-preserving DL computation. However, in practice, 2PC methods often incur high computation and communication overhead, which can impede their use in large-scale systems. To address this challenge, we introduce RRNet, a systematic framework that aims to jointly reduce the overhead of MPC comparison protocols and accelerate computation through hardware acceleration. Our approach integrates the hardware latency of cryptographic building blocks into the DNN loss function, resulting in improved energy efficiency, accuracy, and security guarantees. Furthermore, we propose a cryptographic hardware scheduler and corresponding performance model for Field Programmable Gate Arrays (FPGAs) to further enhance the efficiency of our framework. Experiments show RRNet achieved a much higher ReLU reduction performance than all SOTA works on CIFAR-10 dataset., Comment: This is work is a updated version of arXiv:2209.09424, the original version has been withdrawn

Published
2023

16. Bactericidal Effect of Combined Slightly Acidic Electrolyzed Water and Ultrasonic Treatment against Escherichia coli on Fresh-Cut Lettuce

Author

GAO Qing, LIU Peihong, RAN Ran, FAN Jiangping, HE Jinsong

Subjects
slightly acidic electrolyzed water, ultrasonic, combined treatment, escherichia coli bactericidal kinetics, synergistic effect, Food processing and manufacture, TP368-456
Abstract

This study aimed to investigate the bactericidal effect and dynamics of slightly acidic electrolyzed water (SAEW) combined with ultrasonic (US) treatment on Escherichia coli on fresh-cut lettuce. Fresh-cut lettuce was treated with SAEW-US under different conditions of solid/liquid ratio (1:5, 1:10 and 1:15 (g/mL)) and temperature (25, 35 and 45 ℃). The colony number of E. coli on lettuce was determined by the plate counting method, and the synergistic effect of combined treatment with SAEW and US on reducing the colony number of E. coli was determined and kinetically analyzed. The sublethal damage of SAEW-US treated E. coli was detected by flow cytometry and the double-layer culture counting method. The results showed that the efficacy of SAEW-US treatment in killing E. coli was significantly better than that of SAEW treatment (p < 0.05), and increased with treatment time, solid/liquid ratio and temperature. SAEW-US treatment had a synergistic effect on killing E. coli, and the sterilization process followed a first-order kinetic model. Moreover, SAEW-US treatment reduced the rate of sublethal damage of E. coli. The results of this study will provide a theoretical basis for the sterilization of fresh agricultural products.

Published
2024
Full Text
View/download PDF

17. Response mechanisms of different Saccharomyces cerevisiae strains to succinic acid

Author

Cai-Yun Xie, Ran-Ran Su, Bo Wu, Zhao-Yong Sun, and Yue-Qin Tang

Subjects
Saccharomyces cerevisiae, Succinic acid, Genetic background, Comparative transcriptomic analysis, Response mechanism, Microbiology, QR1-502
Abstract

Abstract Background The production of succinic acid (SA) from biomass has attracted worldwide interest. Saccharomyces cerevisiae is preferred for SA production due to its strong tolerance to low pH conditions, ease of genetic manipulation, and extensive application in industrial processes. However, when compared with bacterial producers, the SA titers and productivities achieved by engineered S. cerevisiae strains were relatively low. To develop efficient SA-producing strains, it’s necessary to clearly understand how S. cerevisiae cells respond to SA. Results In this study, we cultivated five S. cerevisiae strains with different genetic backgrounds under different concentrations of SA. Among them, KF7 and NBRC1958 demonstrated high tolerance to SA, whereas NBRC2018 displayed the least tolerance. Therefore, these three strains were chosen to study how S. cerevisiae responds to SA. Under a concentration of 20 g/L SA, only a few differentially expressed genes were observed in three strains. At the higher concentration of 60 g/L SA, the response mechanisms of the three strains diverged notably. For KF7, genes involved in the glyoxylate cycle were significantly downregulated, whereas genes involved in gluconeogenesis, the pentose phosphate pathway, protein folding, and meiosis were significantly upregulated. For NBRC1958, genes related to the biosynthesis of vitamin B6, thiamin, and purine were significantly downregulated, whereas genes related to protein folding, toxin efflux, and cell wall remodeling were significantly upregulated. For NBRC2018, there was a significant upregulation of genes connected to the pentose phosphate pathway, gluconeogenesis, fatty acid utilization, and protein folding, except for the small heat shock protein gene HSP26. Overexpression of HSP26 and HSP42 notably enhanced the cell growth of NBRC1958 both in the presence and absence of SA. Conclusions The inherent activities of small heat shock proteins, the levels of acetyl-CoA and the strains’ potential capacity to consume SA all seem to affect the responses and tolerances of S. cerevisiae strains to SA. These factors should be taken into consideration when choosing host strains for SA production. This study provides a theoretical basis and identifies potential host strains for the development of robust and efficient SA-producing strains.

Published
2024
Full Text
View/download PDF

18. Epidemiological Survey of Urinary Incontinence and Its Impact on Quality of Life in Women Aged 40-60 Years in Guizhou Province

Author

WU Chunyan, DENG Yinglan, ZHAO Ping, HU Shixiu, WU Xiaoyan, ZHU Yan, LI Xiangyan, XIE Zhengzhou, WANG Rong, GAO Yan, YANG Xiuying, XU Haina, LU Ran, RAN Limei

Subjects
urinary incontinence, urinary incontinence, stress, urinary incontinence, urge, mixed urinary incontinence, quality of life, epidemiology, guizhou province, Medicine
Abstract

Background Urinary incontinence often occurs in women aged 40-60 years and contributes to a significant impact on women's physical health and quality of life. Objective To explore the prevalence and influencing factors of urinary incontinence and its impact on quality of life in women aged 40-60 years in Guizhou Province. Methods From June 2022 to January 2023, women aged 40-60 years who underwent health checkups in 10 hospitals in 7 regions of Guizhou Province were selected as the study subjects. A self-designed general data questionnaire was used to collect the basic data of the study subjects, and International Consultation on Incontinence Questionnaire-Female Lower Urinary Tract Symptoms (ICIQ-FLUTS) and Incontinence Quality of Life Measure (I-QoL) were used to determine the severity of urinary incontinence and the level of quality of life of the study subjects, respectively. Different types of incontinence, including stress incontinence, urge incontinence, and mixed incontinence, were determined according to the diagnostic criteria established by the International Continence Society (ICS). Multivariate Logistic regression analysis was used to explore the influencing factors of urinary incontinence, and explore the effects of different types and severity levels of urinary incontinence on patients' quality of life. Results A total of 3 000 questionnaires were distributed, and 2 966 valid questionnaires were collected, with a valid recovery rate of 98.9%. The prevalence of urinary incontinence among the 2 966 study subjects was 33.4% (991/2 966), including 20.1% (596/2 966) for stress incontinence, 2.0% (58/2 966) for urge incontinence, and 11.4% (337/2 966) for mixed incontinence. Multivariate Logistic regression analysis showed that physical exercise, menstrual status, parity, history of pelvic organ prolapse, perineal laceration, and lateral episiotomy were the influencing factors of stress incontinence (P

Published
2024
Full Text
View/download PDF

19. Pharmacokinetics, pharmacodynamics, and safety of GS‐3583, a FLT3 agonist Fc fusion protein, from single‐ascending‐dose phase I study in healthy participants

Author

Anees M. Dauki, Nishanthan Rajakumaraswamy, Torsten Trowe, Winnie Weng, Kai‐Wen Lin, Emon Elboudjwarej, Ann Ran‐Ran Qin, Christian Schwabe, Michelle R. Kuhne, and Ahmed A. Othman

Subjects
Therapeutics. Pharmacology, RM1-950, Public aspects of medicine, RA1-1270
Abstract

Abstract Conventional dendritic cells subtype 1 (cDC1) play a vital role in the priming and expansion of tumor‐specific CD8+ T cells and their recruitment to tumor microenvironment. However, cDC1s are often underrepresented in the microenvironment. Systemic administration of Fms‐like tyrosine kinase 3 ligand, a hematopoietic growth factor that binds to FLT3 on myeloid and lymphoid progenitor cells, leads to cDC1 expansion in the periphery and recruitment into the microenvironment. FLT3 pathway stimulation using GS‐3583, a novel FLT3 agonistic Fc fusion protein, has the potential to promote T‐cell mediated antitumor activity. This was a first‐in‐human, placebo‐controlled study of GS‐3583 in healthy participants to evaluate the safety, pharmacokinetics (PK), and pharmacodynamic (PD) of escalating single doses (75–2000 μg) of GS‐3583. Each dose cohort enrolled 8–12 healthy participants who received GS‐3583 or placebo as single IV infusion at 3:1 ratio. As part of the PD evaluation, the changes in the number of cDC1 cells were investigated. GS‐3583 was well‐tolerated in healthy participants up to the highest evaluated dose (2000 μg). There have been no serious or grade III or higher adverse events. PK analysis suggested a dose‐dependent increase in GS‐3583 exposure with target‐mediated disposition characteristics at low doses. PD analysis shows that administration of GS‐3583 resulted in transient, dose‐dependent increases in cDC1 cells that returned to baseline within 3 weeks of drug administration. The pharmacokinetics and pharmacodynamics of GS‐3583 following single dosing were characterized in this study which enabled subsequent phase Ib assessments in patients with advanced solid tumors.

Published
2024
Full Text
View/download PDF

20. CryptoGCN: Fast and Scalable Homomorphically Encrypted Graph Convolutional Network Inference

Author

Ran, Ran, Xu, Nuo, Wang, Wei, Quan, Gang, Yin, Jieming, and Wen, Wujie

Subjects
Computer Science - Cryptography and Security, Computer Science - Artificial Intelligence, Computer Science - Machine Learning
Abstract

Recently cloud-based graph convolutional network (GCN) has demonstrated great success and potential in many privacy-sensitive applications such as personal healthcare and financial systems. Despite its high inference accuracy and performance on cloud, maintaining data privacy in GCN inference, which is of paramount importance to these practical applications, remains largely unexplored. In this paper, we take an initial attempt towards this and develop $\textit{CryptoGCN}$--a homomorphic encryption (HE) based GCN inference framework. A key to the success of our approach is to reduce the tremendous computational overhead for HE operations, which can be orders of magnitude higher than its counterparts in the plaintext space. To this end, we develop an approach that can effectively take advantage of the sparsity of matrix operations in GCN inference to significantly reduce the computational overhead. Specifically, we propose a novel AMA data formatting method and associated spatial convolution methods, which can exploit the complex graph structure and perform efficient matrix-matrix multiplication in HE computation and thus greatly reduce the HE operations. We also develop a co-optimization framework that can explore the trade offs among the accuracy, security level, and computational overhead by judicious pruning and polynomial approximation of activation module in GCNs. Based on the NTU-XVIEW skeleton joint dataset, i.e., the largest dataset evaluated homomorphically by far as we are aware of, our experimental results demonstrate that $\textit{CryptoGCN}$ outperforms state-of-the-art solutions in terms of the latency and number of homomorphic operations, i.e., achieving as much as a 3.10$\times$ speedup on latency and reduces the total Homomorphic Operation Count by 77.4\% with a small accuracy loss of 1-1.5$\%$., Comment: Accepted in Conference on Neural Information Processing Systems (NeurIPS 2022)

Published
2022

21. PolyMPCNet: Towards ReLU-free Neural Architecture Search in Two-party Computation Based Private Inference

Author

Peng, Hongwu, Zhou, Shanglin, Luo, Yukui, Duan, Shijin, Xu, Nuo, Ran, Ran, Huang, Shaoyi, Wang, Chenghong, Geng, Tong, Li, Ang, Wen, Wujie, Xu, Xiaolin, and Ding, Caiwen

Subjects
Computer Science - Cryptography and Security, Computer Science - Machine Learning, I.2, E.3, C.3
Abstract

The rapid growth and deployment of deep learning (DL) has witnessed emerging privacy and security concerns. To mitigate these issues, secure multi-party computation (MPC) has been discussed, to enable the privacy-preserving DL computation. In practice, they often come at very high computation and communication overhead, and potentially prohibit their popularity in large scale systems. Two orthogonal research trends have attracted enormous interests in addressing the energy efficiency in secure deep learning, i.e., overhead reduction of MPC comparison protocol, and hardware acceleration. However, they either achieve a low reduction ratio and suffer from high latency due to limited computation and communication saving, or are power-hungry as existing works mainly focus on general computing platforms such as CPUs and GPUs. In this work, as the first attempt, we develop a systematic framework, PolyMPCNet, of joint overhead reduction of MPC comparison protocol and hardware acceleration, by integrating hardware latency of the cryptographic building block into the DNN loss function to achieve high energy efficiency, accuracy, and security guarantee. Instead of heuristically checking the model sensitivity after a DNN is well-trained (through deleting or dropping some non-polynomial operators), our key design principle is to em enforce exactly what is assumed in the DNN design -- training a DNN that is both hardware efficient and secure, while escaping the local minima and saddle points and maintaining high accuracy. More specifically, we propose a straight through polynomial activation initialization method for cryptographic hardware friendly trainable polynomial activation function to replace the expensive 2P-ReLU operator. We develop a cryptographic hardware scheduler and the corresponding performance model for Field Programmable Gate Arrays (FPGA) platform., Comment: Uploaded a new version of the paper in another new submission: RRNet: Towards ReLU-Reduced Neural Network for Two-party Computation Based Private Inference [arXiv:2302.02292]

Published
2022

22. EVE: Environmental Adaptive Neural Network Models for Low-power Energy Harvesting System

Author

Islam, Sahidul, Zhou, Shanglin, Ran, Ran, Jin, Yufang, Wen, Wujie, Ding, Caiwen, and Xie, Mimi

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

IoT devices are increasingly being implemented with neural network models to enable smart applications. Energy harvesting (EH) technology that harvests energy from ambient environment is a promising alternative to batteries for powering those devices due to the low maintenance cost and wide availability of the energy sources. However, the power provided by the energy harvester is low and has an intrinsic drawback of instability since it varies with the ambient environment. This paper proposes EVE, an automated machine learning (autoML) co-exploration framework to search for desired multi-models with shared weights for energy harvesting IoT devices. Those shared models incur significantly reduced memory footprint with different levels of model sparsity, latency, and accuracy to adapt to the environmental changes. An efficient on-device implementation architecture is further developed to efficiently execute each model on device. A run-time model extraction algorithm is proposed that retrieves individual model with negligible overhead when a specific model mode is triggered.Experimental results show that the neural networks models generated by EVE is on average 2.5X times faster than the baseline models without pruning and shared weights.

Published
2022

23. Etiological analysis of 167 cases of drug-resistant epilepsy in children

Author

Ran-Ran Zuo, Mei Jin, and Su-Zhen Sun

Subjects
Drug-resistant epilepsy, Etiology, Children, Phenotype, Pediatrics, RJ1-570
Abstract

Abstract Background To analyze the etiological distribution characteristics of drug-resistant epilepsy (DRE) in children, with the aim of providing valuable perspectives to enhance clinical practice. Methods In this retrospective study, clinical data were collected on 167 children with DRE who were hospitalized between January 2020 and December 2022, including gender, age of onset, seizure types, video electroencephalogram(VEEG) recordings, neuroimaging, and genetic testing results. Based on the etiology of epilepsy, the enrolled children were categorized into different groups. The rank-sum test was conducted to compare the age of onset for different etiologies. Results Of the 167 cases, 89 (53.3%) had a clear etiology. Among them, structural factors account for 23.4%, genetic factors for 19.2%, multiple factors for 7.2%, and immunological factors for 3.6%. The age of onset was significantly earlier in children with genetic causes than those with structural (P

Published
2024
Full Text
View/download PDF

24. Expression and in vitro glycosylation of recombinant edible bird nest (EBN) mucin

Author

Ge Cheng, Yunsheng Lyu, Ran Ran, Li Liu, and Josef Voglmeir

Subjects
edible bird's nest, recombinant mucin, polypeptide galnac-transferase, o-glycosylation, functional glycoproteins, Food processing and manufacture, TP368-456
Abstract

Edible bird's nest (EBN) is considered a delicacy in Chinese cuisine. The nests are produced by male swiftlets (Aerodramus sp.). Due to their high price, the nutritional benefit from EBN can currently only be afforded by affluent consumers. Therefore, the ability to identify and produce recombinant, functional EBN mucins is highly desirable. Herein, we report the identification, bacterial expression, purification and glycosylation of the first recombinant mucin derived from the Edible-nest Swiftlet Aerodramus fuciphagus (AfMuc). Recombinant AfMuc was successfully expressed at high expression levels and purified to near homogeneity by nickel-chelation affinity chromatography. The in vitro glycosylation of the recombinant AfMuc protein backbone resulted in the addition of up to four GalNAc residues and could be achieved by using a human and a previously unstudied Aerodramus fuciphagus polypeptide N-acetylgalactosaminyltransferase (HsGalNT2 and AfGalNT2, respectively). A comparison of the biochemical properties of these enzymes revealed characteristics of these enzymes.

Published
2024
Full Text
View/download PDF

25. Synergistic dual-phase air electrode enables high and durable performance of reversible proton ceramic electrochemical cells

Author

Zuoqing Liu, Yuesheng Bai, Hainan Sun, Daqin Guan, Wenhuai Li, Wei-Hsiang Huang, Chih-Wen Pao, Zhiwei Hu, Guangming Yang, Yinlong Zhu, Ran Ran, Wei Zhou, and Zongping Shao

Subjects
Science
Abstract

Abstract Reversible proton ceramic electrochemical cells are promising solid-state ion devices for efficient power generation and energy storage, but necessitate effective air electrodes to accelerate the commercial application. Here, we construct a triple-conducting hybrid electrode through a stoichiometry tuning strategy, composed of a cubic phase Ba0.5Sr0.5Co0.8Fe0.2O3−δ and a hexagonal phase Ba4Sr4(Co0.8Fe0.2)4O16−δ. Unlike the common method of creating self-assembled hybrids by breaking through material tolerance limits, the strategy of adjusting the stoichiometric ratio of the A-site/B-site not only achieves strong interactions between hybrid phases, but also can efficiently modifies the phase contents. When operate as an air electrode for reversible proton ceramic electrochemical cell, the hybrid electrode with unique dual-phase synergy shows excellent electrochemical performance with a current density of 3.73 A cm−2 @ 1.3 V in electrolysis mode and a peak power density of 1.99 W cm−2 in fuel cell mode at 650 °C.

Published
2024
Full Text
View/download PDF

26. NeuGuard: Lightweight Neuron-Guided Defense against Membership Inference Attacks

Author

Xu, Nuo, Wang, Binghui, Ran, Ran, Wen, Wujie, and Venkitasubramaniam, Parv

Subjects
Computer Science - Cryptography and Security, Computer Science - Artificial Intelligence, Computer Science - Machine Learning
Abstract

Membership inference attacks (MIAs) against machine learning models can lead to serious privacy risks for the training dataset used in the model training. In this paper, we propose a novel and effective Neuron-Guided Defense method named NeuGuard against membership inference attacks (MIAs). We identify a key weakness in existing defense mechanisms against MIAs wherein they cannot simultaneously defend against two commonly used neural network based MIAs, indicating that these two attacks should be separately evaluated to assure the defense effectiveness. We propose NeuGuard, a new defense approach that jointly controls the output and inner neurons' activation with the object to guide the model output of training set and testing set to have close distributions. NeuGuard consists of class-wise variance minimization targeting restricting the final output neurons and layer-wise balanced output control aiming to constrain the inner neurons in each layer. We evaluate NeuGuard and compare it with state-of-the-art defenses against two neural network based MIAs, five strongest metric based MIAs including the newly proposed label-only MIA on three benchmark datasets. Results show that NeuGuard outperforms the state-of-the-art defenses by offering much improved utility-privacy trade-off, generality, and overhead.

Published
2022
Full Text
View/download PDF

27. Resonance of fatty acid metabolism and immune infiltration in anti-PD-1 monotherapy for breast cancer

Author

Andi Zhao, Jin Yang, Ran Ran, Shidi Zhao, Yuxin Cui, Fang Hu, and Yan Zhou

Subjects
Tumor immunology, Immunogenetics, Tumor immune microenvironment, Fatty acid metabolism, Biomarkers of breast cancer, Neoplasms. Tumors. Oncology. Including cancer and carcinogens, RC254-282
Abstract

The interaction between tumor fatty acid metabolism and immune microenvironment is a novel topic in oncology research, and the relationship of lipid-derived factors with immune editing in tumor is unclear. The breast cancer samples from the TCGA database were used as the training set, and samples from GSE42568 were employed as the validation set for constructing a model to identify a signature associated with fatty acid metabolism through Lasso Cox regression. And the changes in immune related signatures and risk score before and after anti-PD-1 monotherapy were caught by the differential analysis in GSE225078. A 14-gene prognostic risk scoring model identifying by fatty acid metabolism relevant signature was conducted, and the high risk group had shorter overall survival and progression free survival than low risk group. Many metabolism-related pathways were enriched in the high risk group, and many immune-related pathways were enriched in low risk group. The crucial differentially expressed genes between the high/low risk groups, CYP4F8 and CD52, were found to be strongly associated with SUCLA2 and ACOT4 of 14-gene model, and strongly related to immune infiltration. Immune related signatures, fatty acid metabolism-risk score and the expression level of ALDH1A1 (in 14-gene-model) changed after anti-PD-1 monotherapy. And the mice model results also showed anti-PD-1 mAb could significantly reduce the expression level of ALDH1A1 (p < 0.01). These results brought up the crosstalk between immune components and fatty acid metabolism in breast cancer microenvironment, which provided a new possibility of targeting fatty acid metabolism for combination therapy in breast cancer immunotherapy.

Published
2024
Full Text
View/download PDF

28. Dynamics of respiratory infectious diseases under rapid urbanization and COVID-19 pandemic in the subcenter of Beijing during 2014–2022

Author

Chang-Yu Guo, Wan-Xue Zhang, Yi-Guo Zhou, Shan-Shan Zhang, Lu Xi, Ran-Ran Zheng, Juan Du, Jianming Zhang, Yan Cui, and Qing-Bin Lu

Subjects
Urbanization, Infectious diseases, Respiratory infectious diseases, COVID-19, Science (General), Q1-390, Social sciences (General), H1-99
Abstract

Objective: The study analyzed the impact of urbanization on epidemiological characteristics of respiratory infectious disease in Tongzhou District, Beijing during 2014–2022 to provide reference for prevention and control priorities of respiratory infectious diseases during the innovative urbanization process in China. Methods: The incidence data of notifiable respiratory infectious diseases (NRIDs) in Tongzhou Beijing during 2014–2022 were summarized. The trend of incidence rate was analyzed by Joinpoint regression model, and entropy method was performed to construct the comprehensive index of urbanization (CIU) and generalized linear model was used to analyze the influence of CIU on the incidence rate of respiratory infectious diseases. Results: Totally 72616 NRIDs cases were reported in Tongzhou District during 2014–2022, and the incidence rate of NRIDs was higher during 2017–2019 (153/100 000) than during 2014–2016 (930/100 000) and during 2020–2022 (371/100 000), respectively (both P

Published
2024
Full Text
View/download PDF

29. Detailed survey of an in vitro intestinal epithelium model by single-cell transcriptomics

Author

Ran Ran, Javier Muñoz Briones, Smrutiti Jena, Nicole L. Anderson, Matthew R. Olson, Leopold N. Green, and Douglas K. Brubaker

Subjects
Cell biology, Expression study, Systems biology, Transcriptomics, Science
Abstract

Summary: The co-culture of two adult human colorectal cancer cell lines, Caco-2 and HT29, on Transwell is commonly used as an in vitro gut mimic, yet the translatability of insights from such a system to adult human physiological contexts is not fully characterized. Here, we used single-cell RNA sequencing on the co-culture to obtain a detailed survey of cell type heterogeneity in the system and conducted a holistic comparison with human physiology. We identified the intestinal stem cell-, transit amplifying-, enterocyte-, goblet cell-, and enteroendocrine-like cells in the system. In general, the co-culture was fetal intestine-like, with less variety of gene expression compared to the adult human gut. Transporters for major types of nutrients were found in the majority of the enterocytes-like cells in the system. TLR 4 was not expressed in the sample, indicating that the co-culture model is incapable of mimicking the innate immune aspect of the human epithelium.

Published
2024
Full Text
View/download PDF

31. Effects of different treatment subjects of Lactiplantibacillus plantarum SCS5 on oxidative stress in type 2 diabetic mice

Author

Xiao Meng, Siheng Ren, Yongkun Hu, Ying Zhang, Ran Ran, Weixi Ge, Lishi Jiang, and Shukun Liu

Subjects
Lactiplantibacillus plantarum SCS5, type 2 diabetes, oxidative stress, pancreatic tissues, Nutrition. Foods and food supply, TX341-641, Food processing and manufacture, TP368-456
Abstract

ABSTRACTThe purpose of this study was to determine the effect of different subjects from Lactiplantibacillus plantarum SCS5 (L. plantarum SCS5) on type 2 diabetes by modulating oxidative stress levels.The normal group (NG) and the model group (MG) of mice were fed normally, while the three test groups (LP1,LP2,LP3) received 10 ml/kg (body weight) of the L. plantarum SCS5 suspension, intracellular material, or heat-killed intracellular material daily for 10 weeks. The weight loss and the increase in blood glucose were alleviated in the LP2 and LP3 groups. Insulin, catalase, superoxide dismutase, and glutathione peroxidase significantly increased in the LP3 group of mice compared with the MG group. Additionally, the pathological sections showed that the lesions in the LP3 group were less severe than those in the MG group. These findings indicate that the heat-killed intracellular L. plantarum SCS5 material was most effective in alleviating oxidative stress in type 2 diabetic mice.

Published
2023
Full Text
View/download PDF

38. LILRB2/PirB mediates macrophage recruitment in fibrogenesis of nonalcoholic steatohepatitis

Author

Li, Dan-Pei, Huang, Li, Kan, Ran-Ran, Meng, Xiao-Yu, Wang, Shu-Yun, Zou, Hua-Jie, Guo, Ya-Ming, Luo, Pei-Qiong, Pan, Li-Meng, Xiang, Yu-Xi, Mao, Bei-Bei, Xie, Yu-Yu, Wang, Zhi-Han, Yang, Min, He, Rui, Yang, Yan, Liu, Zhe-Long, Xie, Jun-Hui, Ma, De-Lin, Zhang, Ben-Ping, Shao, Shi-Ying, Chen, Xi, Xu, Si-Miao, He, Wen-Tao, Li, Wen-Jun, Chen, Yong, and Yu, Xue-Feng

Published
2023
Full Text
View/download PDF

42. Genomic insights into local adaptation and phenotypic diversity of Wenchang chickens

Author

Li-Hong Gu, Ran-Ran Wu, Xin-Li Zheng, An Fu, Zeng-Yang Xing, Yi-Yong Chen, Zhong-Chun He, Li-Zhi Lu, Yan-Tao Qi, An-Hong Chen, Ya-Ping Zhang, Tie-Shan Xu, Min-Sheng Peng, and Cheng Ma

Subjects
Wenchang chicken, population structure, feather color, local adaptation, GWAS, Animal culture, SF1-1100
Abstract

ABSTRACT: Wenchang chicken, a prized local breed in Hainan Province of China renowned for its exceptional adaptability to tropical environments and good meat quality, is deeply favored by the public. However, an insufficient understanding of its population architecture and the unclear genetic basis that governs its typical attributes have posed challenges in the protection and breeding of this precious breed. To address these gaps, we conducted whole-genome resequencing on 200 Wenchang chicken samples derived from 10 distinct strains, and we gathered data on an array of 21 phenotype traits. Population genomics analysis unveiled distinctive population structures in Wenchang chickens, primarily attributed to strong artificial selection for different feather colors. Selection sweep analysis identified a group of candidate genes, including PCDH9, DPF3, CDIN1, and SUGCT, closely linked to adaptations that enhance resilience in tropical island habitats. Genome-wide association studies (GWAS) highlighted potential candidate genes associated with diverse feather color traits, encompassing TYR, RAB38, TRPM1, GABARAPL2, CDH1, ZMIZ1, LYST, MC1R, and SASH1. Through the comprehensive analysis of high-quality genomic and phenotypic data across diverse Wenchang chicken resource groups, this study unveils the intricate genetic backgrounds and population structures of Wenchang chickens. Additionally, it identifies multiple candidate genes linked to environmental adaptation, feather color variations, and production traits. These insights not only provide genetic reference for the purification and breeding of Wenchang chickens but also broaden our understanding of the genetic basis of phenotypic diversity in chickens.

Published
2024
Full Text
View/download PDF

43. Characterization of bacterial diversity and screening of cellulose-degrading bacteria in the gut system of Glenea cantor (Fabricius) larvae

Author

Ran-Ran Su, Bi-Qiong Pan, You-Xi Luo, Xia-Lin Zheng, Wen Lu, and Xiao-Yun Wang

Subjects
gut bacterial community, Glenea cantor, illumina MiSeq 16S rRNA sequencing, lignocellulose degradation, cellulose degrading bacteria, Biotechnology, TP248.13-248.65
Abstract

The intestinal bacteria of longhorn beetles would be ideal targets for pest control and lignocellulosic resources by destroying or exploiting their cellulose-degrading function. This article aims to investigate the diversity and community structure of intestinal bacteria the oligophagous longhorn beetle Glenea cantor. Additionally, it seeks to identify the presence of lignocellulose-degrading bacteria in the gut, and explore their role in consuming host kapok trees Bombax malabaricum. In this study, the bacterial community from G. cantor was examined by Illumina sequencing of 16S ribosomal RNA (rRNA) targeting the V3 and V4 regions. A total of 563,201 valid sequences and 814 OTUs were obtained. The dominant phyla were Proteobacteria, and the dominant genera were Acinetobacter and Lactococcus. The analysis of microbial diversity revealed a high bacterial diversity in the samples, with the gut bacteria playing a crucial role in the physiological activities of the host, particularly, 9 genera of intestinal bacteria with cellulose degradation function were found, highlighting their vital role in cellulose degradation. Five strains of cellulose-degrading bacteria, belonging to the genus Pseudomonas, were obtained from the intestinal tract of G. cantor larvae using traditional isolation and culture techniques as well as 16S rDNA sequencing. Among these strains, A4 exhibited a cellulase activity of 94.42 ± 0.42 U/mL, while A5 displayed the highest filter paper enzyme activity of 127.46 ± 3.54 U/mL. These results offered valuable insights into potential targets for pest control through internal attack digestion and cellulose-degrading bacteria in longhorn beetles.

Published
2024
Full Text
View/download PDF

44. Assessing risk factors associated with breakthrough early post-traumatic seizures in patients receiving phenytoin prophylaxis

Author

Eugene Generoso, Calvin Diep, Carolyn Hua, Elizabeth Rader, Ran Ran, Nathanael J. Lee, and Lucia Rivera-Lara

Subjects
traumatic brain injury, post-traumatic epilepsy, phenytoin, risk factors, seizure prophylaxis, Neurology. Diseases of the nervous system, RC346-429
Abstract

ObjectivePost-traumatic seizure (PTS) is a well-known complication of traumatic brain injury (TBI). The objective of this study was to identify risk factors associated with breakthrough early PTS in TBI patients receiving phenytoin prophylaxis.MethodsThis was a single-centered retrospective study including adult patients admitted to the intensive care unit (ICU), had a TBI, and started on phenytoin for seizure prophylaxis within 24 h of admission. The primary outcome was the incidence and factors associated with early PTS, defined as a confirmed seizure on a continuous electroencephalogram within 7 days of TBI. Secondary outcomes included the association between early post-traumatic seizures and ICU length of stay, hospital length of stay, and in-hospital mortality.ResultsA total of 105 patients were included in the final analysis. Patients with early PTS were older (65 vs. 48 years old, p = 0.01), had a higher Marshall score (5 vs. 2, p = 0.01), were more likely to have a Marshall score > 2 (73 vs. 37%, p = 0.01), and had more neurosurgeries for hematoma evacuation (57 vs. 19%, p = 0.01). In patients with early PTS, 57% had a level at the time of seizure, and of those, 87.5% had a therapeutic level (>10 mcg/mL). Patients with early PTS had a longer ICU length of stay (14.7 vs. 5.9 days, p = 0.04) and a greater proportion of hospital mortality (21 vs. 2%, p = 0.02).ConclusionPatients with higher age, Marshall score, and neurosurgical procedures for hematoma evacuation had higher incidences of breakthrough early PTS despite the use of phenytoin prophylaxis. The majority of patients with early PTS had therapeutic phenytoin levels at the time of seizure when a level was available; however, approximately half (43%) did not have a level.

Published
2024
Full Text
View/download PDF

45. Prevention of exacerbation in patients with moderate-to-very severe COPD with the intent to modulate respiratory microbiome: a pilot prospective, multi-center, randomized controlled trial

Author

Jian-lan Hua, Zi-feng Yang, Qi-jian Cheng, Yao-pin Han, Zheng-tu Li, Ran-ran Dai, Bin-feng He, Yi-xing Wu, and Jing Zhang

Subjects
chronic obstructive pulmonary disease, vaccination, inhaled antibiotics, probiotic, amikacin, Medicine (General), R5-920
Abstract

IntroductionConsidering the role of bacteria in the onset of acute exacerbation of COPD (AECOPD), we hypothesized that the use of influenza-Streptococcus pneumoniae vaccination, oral probiotics or inhaled amikacin could prevent AECOPD.MethodsIn this pilot prospective, muti-central, randomized trial, moderate-to-very severe COPD subjects with a history of moderate-to-severe exacerbations in the previous year were enrolled and assigned in a ratio of 1:1:1:1 into 4 groups. All participants were managed based on the conventional treatment recommended by GOLD 2019 report for 3 months, with three groups receiving additional treatment of inhaled amikacin (0.4 g twice daily, 5–7 days monthly for 3 months), oral probiotic Lactobacillus rhamnosus GG (1 tablet daily for 3 months), or influenza-S. pneumoniae vaccination. The primary endpoint was time to the next onset of moderate-to-severe AECOPD from enrollment. Secondary endpoints included CAT score, mMRC score, adverse events, and survival in 12 months.ResultsAmong all 112 analyzed subjects (101 males, 96 smokers or ex-smokers, mean ± SD age 67.19 ± 7.39 years, FEV1 41.06 ± 16.09% predicted), those who were given dual vaccination (239.7 vs. 198.2 days, p = 0.044, 95%CI [0.85, 82.13]) and oral probiotics (248.8 vs. 198.2 days, p = 0.017, 95%CI [7.49, 93.59]) had significantly delayed onset of next moderate-to-severe AECOPD than those received conventional treatment only. For subjects with high symptom burden, the exacerbations were significantly delayed in inhaled amikacin group as compared to the conventional treatment group (237.3 vs. 179.1 days, p = 0.009, 95%CI [12.40,104.04]). The three interventions seemed to be safe and well tolerated for patient with stable COPD.ConclusionThe influenza-S. pneumoniae vaccine and long-term oral probiotic LGG can significantly delay the next moderate-to-severe AECOPD. Periodically amikacin inhalation seems to work in symptomatic patients. The findings in the current study warrants validation in future studies with microbiome investigation.Clinical trial registration:https://clinicaltrials.gov/, identifier NCT03449459.

Published
2024
Full Text
View/download PDF

46. What can GPT-4 do for Diagnosing Rare Eye Diseases? A Pilot Study

Author

Xiaoyan Hu, An Ran Ran, Truong X. Nguyen, Simon Szeto, Jason C. Yam, Carmen K. M. Chan, and Carol Y. Cheung

Subjects
Rare eye disease, Generative pretrained transformer, Large language models, Artificial intelligence, Ophthalmology, RE1-994
Abstract

Abstract Introduction Generative pretrained transformer-4 (GPT-4) has gained widespread attention from society, and its potential has been extensively evaluated in many areas. However, investigation of GPT-4’s use in medicine, especially in the ophthalmology field, is still limited. This study aims to evaluate GPT-4’s capability to identify rare ophthalmic diseases in three simulated scenarios for different end-users, including patients, family physicians, and junior ophthalmologists. Methods We selected ten treatable rare ophthalmic disease cases from the publicly available EyeRounds service. We gradually increased the amount of information fed into GPT-4 to simulate the scenarios of patient, family physician, and junior ophthalmologist using GPT-4. GPT-4’s responses were evaluated from two aspects: suitability (appropriate or inappropriate) and accuracy (right or wrong) by senior ophthalmologists (> 10 years’ experiences). Results Among the 30 responses, 83.3% were considered "appropriate" by senior ophthalmologists. In the scenarios of simulated patient, family physician, and junior ophthalmologist, seven (70%), ten (100%), and eight (80%) responses were graded as “appropriate” by senior ophthalmologists. However, compared to the ground truth, GPT-4 could only output several possible diseases generally without “right” responses in the simulated patient scenarios. In contrast, in the simulated family physician scenario, 50% of GPT-4's responses were “right,” and in the simulated junior ophthalmologist scenario, the model achieved a higher “right” rate of 90%. Conclusion To our knowledge, this is the first proof-of-concept study that evaluates GPT-4’s capacity to identify rare eye diseases in simulated scenarios involving patients, family physicians, and junior ophthalmologists. The results indicate that GPT-4 has the potential to serve as a consultation assisting tool for patients and family physicians to receive referral suggestions and an assisting tool for junior ophthalmologists to diagnose rare eye diseases. However, it is important to approach GPT-4 with caution and acknowledge the need for verification and careful referrals in clinical settings.

Published
2023
Full Text
View/download PDF

47. LILRB2/PirB mediates macrophage recruitment in fibrogenesis of nonalcoholic steatohepatitis

Author

Dan-Pei Li, Li Huang, Ran-Ran Kan, Xiao-Yu Meng, Shu-Yun Wang, Hua-Jie Zou, Ya-Ming Guo, Pei-Qiong Luo, Li-Meng Pan, Yu-Xi Xiang, Bei-Bei Mao, Yu-Yu Xie, Zhi-Han Wang, Min Yang, Rui He, Yan Yang, Zhe-Long Liu, Jun-Hui Xie, De-Lin Ma, Ben-Ping Zhang, Shi-Ying Shao, Xi Chen, Si-Miao Xu, Wen-Tao He, Wen-Jun Li, Yong Chen, and Xue-Feng Yu

Subjects
Science
Abstract

Abstract Inhibition of immunocyte infiltration and activation has been suggested to effectively ameliorate nonalcoholic steatohepatitis (NASH). Paired immunoglobulin-like receptor B (PirB) and its human ortholog receptor, leukocyte immunoglobulin-like receptor B (LILRB2), are immune-inhibitory receptors. However, their role in NASH pathogenesis is still unclear. Here, we demonstrate that PirB/LILRB2 regulates the migration of macrophages during NASH by binding with its ligand angiopoietin-like protein 8 (ANGPTL8). Hepatocyte-specific ANGPTL8 knockout reduces MDM infiltration and resolves lipid accumulation and fibrosis progression in the livers of NASH mice. In addition, PirB−/− bone marrow (BM) chimeras abrogate ANGPTL8-induced MDM migration to the liver. And yet, PirB ectodomain protein could ameliorate NASH by sequestering ANGPTL8. Furthermore, LILRB2-ANGPTL8 binding-promoted MDM migration and inflammatory activation are also observed in human peripheral blood monocytes. Taken together, our findings reveal the role of PirB/LILRB2 in NASH pathogenesis and identify PirB/LILRB2-ANGPTL8 signaling as a potential target for the management or treatment of NASH.

Published
2023
Full Text
View/download PDF

Catalog

Books, media, physical & digital resources
See catalog results