Search

Your search keyword '"Li, Xinglin"' showing total 31 results
Start Over Author "Li, Xinglin" Publication Year Range This year
31 results on '"Li, Xinglin"'

1. Defense Against Syntactic Textual Backdoor Attacks with Token Substitution

Author

Li, Xinglin, He, Xianwen, Li, Yao, and Cheng, Minhao

Subjects
Computer Science - Computation and Language
Abstract

Textual backdoor attacks present a substantial security risk to Large Language Models (LLM). It embeds carefully chosen triggers into a victim model at the training stage, and makes the model erroneously predict inputs containing the same triggers as a certain class. Prior backdoor defense methods primarily target special token-based triggers, leaving syntax-based triggers insufficiently addressed. To fill this gap, this paper proposes a novel online defense algorithm that effectively counters syntax-based as well as special token-based backdoor attacks. The algorithm replaces semantically meaningful words in sentences with entirely different ones but preserves the syntactic templates or special tokens, and then compares the predicted labels before and after the substitution to determine whether a sentence contains triggers. Experimental results confirm the algorithm's performance against these two types of triggers, offering a comprehensive defense strategy for model integrity.

Published
2024

2. BRIEDGE: EEG-Adaptive Edge AI for Multi-Brain to Multi-Robot Interaction

Author

Ouyang, Jinhui, Wu, Mingzhu, Li, Xinglin, Deng, Hanhui, and Wu, Di

Subjects
Electrical Engineering and Systems Science - Signal Processing, Computer Science - Artificial Intelligence, Computer Science - Human-Computer Interaction, Computer Science - Machine Learning, Computer Science - Robotics
Abstract

Recent advances in EEG-based BCI technologies have revealed the potential of brain-to-robot collaboration through the integration of sensing, computing, communication, and control. In this paper, we present BRIEDGE as an end-to-end system for multi-brain to multi-robot interaction through an EEG-adaptive neural network and an encoding-decoding communication framework, as illustrated in Fig.1. As depicted, the edge mobile server or edge portable server will collect EEG data from the users and utilize the EEG-adaptive neural network to identify the users' intentions. The encoding-decoding communication framework then encodes the EEG-based semantic information and decodes it into commands in the process of data transmission. To better extract the joint features of heterogeneous EEG data as well as enhance classification accuracy, BRIEDGE introduces an informer-based ProbSparse self-attention mechanism. Meanwhile, parallel and secure transmissions for multi-user multi-task scenarios under physical channels are addressed by dynamic autoencoder and autodecoder communications. From mobile computing and edge AI perspectives, model compression schemes composed of pruning, weight sharing, and quantization are also used to deploy lightweight EEG-adaptive models running on both transmitter and receiver sides. Based on the effectiveness of these components, a code map representing various commands enables multiple users to control multiple intelligent agents concurrently. Our experiments in comparison with state-of-the-art works show that BRIEDGE achieves the best classification accuracy of heterogeneous EEG data, and more stable performance under noisy environments.

Published
2024

3. HAIFIT: Human-to-AI Fashion Image Translation

Author

Jiang, Jianan, Li, Xinglin, Yu, Weiren, and Wu, Di

Subjects
Computer Science - Computer Vision and Pattern Recognition
Abstract

In the realm of fashion design, sketches serve as the canvas for expressing an artist's distinctive drawing style and creative vision, capturing intricate details like stroke variations and texture nuances. The advent of sketch-to-image cross-modal translation technology has notably aided designers. However, existing methods often compromise these sketch details during image generation, resulting in images that deviate from the designer's intended concept. This limitation hampers the ability to offer designers a precise preview of the final output. To overcome this challenge, we introduce HAIFIT, a novel approach that transforms sketches into high-fidelity, lifelike clothing images by integrating multi-scale features and capturing extensive feature map dependencies from diverse perspectives. Through extensive qualitative and quantitative evaluations conducted on our self-collected dataset, our method demonstrates superior performance compared to existing methods in generating photorealistic clothing images. Our method excels in preserving the distinctive style and intricate details essential for fashion design applications. In addition, our method also has obvious advantages in model training and inference speed, contributing to reducing designers' time costs and improving design efficiency., Comment: 10 pages,8 figures

Published
2024

4. EXGC: Bridging Efficiency and Explainability in Graph Condensation

Author

Fang, Junfeng, Li, Xinglin, Sui, Yongduo, Gao, Yuan, Zhang, Guibin, Wang, Kun, Wang, Xiang, and He, Xiangnan

Subjects
Computer Science - Machine Learning
Abstract

Graph representation learning on vast datasets, like web data, has made significant strides. However, the associated computational and storage overheads raise concerns. In sight of this, Graph condensation (GCond) has been introduced to distill these large real datasets into a more concise yet information-rich synthetic graph. Despite acceleration efforts, existing GCond methods mainly grapple with efficiency, especially on expansive web data graphs. Hence, in this work, we pinpoint two major inefficiencies of current paradigms: (1) the concurrent updating of a vast parameter set, and (2) pronounced parameter redundancy. To counteract these two limitations correspondingly, we first (1) employ the Mean-Field variational approximation for convergence acceleration, and then (2) propose the objective of Gradient Information Bottleneck (GDIB) to prune redundancy. By incorporating the leading explanation techniques (e.g., GNNExplainer and GSAT) to instantiate the GDIB, our EXGC, the Efficient and eXplainable Graph Condensation method is proposed, which can markedly boost efficiency and inject explainability. Our extensive evaluations across eight datasets underscore EXGC's superiority and relevance. Code is available at https://github.com/MangoKiller/EXGC.

Published
2024

6. Machine Learning Test for Modulation Range of Ammonium Metatungstate Based Liquid Electrochromic Devices

Author

Yan, Haoyang, Li, Muyun, Ning, Honglong, Guo, Chenxiao, Li, Xinglin, Zhang, Zihan, Jiang, Bocheng, Xu, Wei, Yao, Rihui, Peng, Junbiao, Angrisani, Leopoldo, Series Editor, Arteaga, Marco, Series Editor, Chakraborty, Samarjit, Series Editor, Chen, Jiming, Series Editor, Chen, Shanben, Series Editor, Chen, Tan Kay, Series Editor, Dillmann, Rüdiger, Series Editor, Duan, Haibin, Series Editor, Ferrari, Gianluigi, Series Editor, Ferre, Manuel, Series Editor, Jabbari, Faryar, Series Editor, Jia, Limin, Series Editor, Kacprzyk, Janusz, Series Editor, Khamis, Alaa, Series Editor, Kroeger, Torsten, Series Editor, Li, Yong, Series Editor, Liang, Qilian, Series Editor, Martín, Ferran, Series Editor, Ming, Tan Cher, Series Editor, Minker, Wolfgang, Series Editor, Misra, Pradeep, Series Editor, Mukhopadhyay, Subhas, Series Editor, Ning, Cun-Zheng, Series Editor, Nishida, Toyoaki, Series Editor, Oneto, Luca, Series Editor, Panigrahi, Bijaya Ketan, Series Editor, Pascucci, Federica, Series Editor, Qin, Yong, Series Editor, Seng, Gan Woon, Series Editor, Speidel, Joachim, Series Editor, Veiga, Germano, Series Editor, Wu, Haitao, Series Editor, Zamboni, Walter, Series Editor, Zhang, Junjie James, Series Editor, Tan, Kay Chen, Series Editor, Song, Huihui, editor, Xu, Min, editor, Yang, Li, editor, Zhang, Linghao, editor, and Yan, Shu, editor

Published
2024
Full Text
View/download PDF

7. Transcription factor ZmNAC59 regulates plant salt resistance in Zea mays L.

Author

ZHANG Yuhan, LI Xinglin, LIU Lijun, FU Jingye, and WANG Qiang

Subjects
maize, transcription factors, salt stress, ion transport, Biology (General), QH301-705.5, Botany, QK1-989
Abstract

[ Objective ] The NAC transcription factor family is a class of transcription factors that is widely studied in plants. It plays an important role in regulating plant growth and development and responding to abiotic stress. Maize is one of the three major food crops , which faces various adverse stresses in growth and development. Salt stress is one of the main environmental factors limiting crop growth and produc- tion. Therefore , it is of great significance to identify maize salt resistance genes and analyze their salt re- sistance mechanism for breeding stress-resistant maize varieties. [ Methods ] This study cloned the maize transcription factor ZmNAC59 and analyzed its conserved domain and phylogenetic relationship using bioinformatics methods. The expression pattern of the gene in leaves under NaCl and MeJA treatments was analyzed using the quantitative real-time PCR method. The stable transgenic system was used to het- erologously express the gene in Arabidopsis thaliana for phenotype observation. At the same time , virus- induced silencing technology was used to silence the gene in maize , followed by phenotype observation un- der salt stress and enzyme activity detection. [ Results ] The qRT-PCR results showed that ZmNAC59 was up-regulated by NaCl and MeJA treatments. After virus-induced silencing of ZmNAC59 , the silenced strains were more sensitive to salt stress and accumulated more ROS. After over expression of ZmNAC59 in A . thaliana , the over expression lines had higher survival rate , less accumulation of ROS , and lower Na + / K + ratios under salt stress , indicating that ZmNAC59 , as a positive regulator of salt stress , can im- prove plant salt resistance by regulating ion homeostasis. qRT-PCR results showed that Na + and K + transport related genes were significantly up-regulated in A . thaliana overexpression lines. After transient expression of ZmNAC59 in maize protoplasts , the expression of ZmSOS1 , ZmNHX1 , and ZmNHX7 were significantly up-regulated after salt treatment. The Dual-LUC experiment showed that ZmSOS1 was the target gene of ZmNAC59. [ Conclusion ] This study finds that ZmNAC59 activates ZmSOS1 promoter and regulates its expression to promote ion transport under salt stress , thereby improving plant salt resist- ance. This provides a basis for the screening of maize stress resistant genes and variety cultivation.

Published
2024
Full Text
View/download PDF

23. Sensitivity analysis based on the fuzzy safety index and its application in possibility-based design optimization.

Author

Wei, Ning, Lu, Zhenzhou, and Li, Xinglin

Subjects
SENSITIVITY analysis, POSSIBILITY, DEFINITIONS, SAFETY
Abstract

The fuzzy safety index (FSI) is put forward in this article under the assumption that the fuzzy inputs are independently and identically distributed. Then, the failure possibility-based constraints in original possibility-based design optimization (PBDO) are equivalently converted to the FSI-based constraints, and the fuzzy safety index approach (FSIA) is presented for solving PBDO. Based on the definition of FSI and the relationship between FSI and the fuzzy most possible failure point (F-MPP), the analytical expression is derived for the local sensitivity of FSI (LS-FSI) with respect to the design parameters, and it is used to provide the exact gradient for FSIA. An efficient iterative method is proposed for solving the FSI, which can estimate LS-FSI as a by-product of analysing the FSI without any additional computational cost. Finally, five numerical examples are demonstrated to show the feasibility of the derived analytical expression of LS-FSI and the proposed FSIA for PBDO. [ABSTRACT FROM AUTHOR]

Published
2024
Full Text
View/download PDF

28. Efficacy of non-pharmacological interventions for primary dysmenorrhoea: a systematic review and Bayesian network meta-analysis

Author

Li, Xinglin, Hao, Xinyu, Liu, Jian-hua, and Huang, Jian-peng

Abstract

ObjectivesTo assess the relative benefits of various non-pharmacological interventions on treating primary dysmenorrhoea within a network meta-analysis.Study designSystematic review and Bayesian network meta-analysis.Inclusion criteriaRandomised controlled trial involving patient with primary dysmenorrhoea and received non-pharmacological interventions.Data sourcesFour databases (Medline, Embase, Cochrane Library and Web of Science) were searched from inception to October first, 2022.Risk-of-bias (RoB) assessmentRoB 2.0 assessment tools was used to assess the risk of bias in the included studies.Synthesis of resultsConventional meta-analysis was conducted by pairwise comparison between non-pharmacological therapy and control treatment. The Bayesian network meta-analysis was conducted by the Aggregate Data Drug Information System Software based on the consistency or inconsistency model, and rank probability was used to indicate the priority of non-pharmacological therapy.Results33 studies involving eight non-pharmacological interventions were included. With regard to conventional meta-analysis, we selected Visual Analogue Scale (VAS) as primary outcome to evaluate the pain intensity. The result showed that eight interventions (Exercise, Herb, Acupuncture, Aromatherapy, Transcutaneous Electrical Nerve Stimulation, Topical heat, Acupressure, Yoga) displayed positive effect on reduction of menstrual pain compared with placebo or no treatment. A Bayesian network meta-analysis revealed that exercise −3.20 (95% CI −4.01 to −2.34), acupuncture −2.90 (95% CI −3.97 to −2.85) and topical heat −2.97 (95% CI −4.66 to −1.29) probably resulted in a reduction in pain intensity (VAS) .ConclusionsNon-pharmacological interventions may result in a reduction or slight reduction in pain intensity compared with no treatment or placebo. Specifically, exercise and acupuncture are considered as potentially effective non-pharmacological treatments in short-term treatment. Indeed, larger and better methodological quality research is needed.Trial registration numberCRD42022351021.

Published
2024
Full Text
View/download PDF

29. Brave the Wind and the Waves: Discovering Robust and Generalizable Graph Lottery Tickets

Author

Wang, Kun, Liang, Yuxuan, Li, Xinglin, Li, Guohao, Ghanem, Bernard, Zimmermann, Roger, Zhou, Zhengyang, Yi, Huahui, Zhang, Yudong, and Wang, Yang

Abstract

The training and inference of Graph Neural Networks (GNNs) are costly when scaling up to large-scale graphs. Graph Lottery Ticket (GLT) has presented the first attempt to accelerate GNN inference on large-scale graphs by jointly pruning the graph structure and the model weights. Though promising, GLT encounters robustness and generalization issues when deployed in real-world scenarios, which are also long-standing and critical problems in deep learning ideology. In real-world scenarios, the distribution of unseen test data is typically diverse. We attribute the failures on out-of-distribution (OOD) data to the incapability of discerning causal patterns, which remain stable amidst distribution shifts. In traditional spase graph learning, the model performance deteriorates dramatically as the graph/network sparsity exceeds a certain high level. Worse still, the pruned GNNs are hard to generalize to unseen graph data due to limited training set at hand. To tackle these issues, we propose the Resilient Graph Lottery Ticket (RGLT) to find more robust and generalizable GLT in GNNs. Concretely, we reactivate a fraction of weights/edges by instantaneous gradient information at each pruning point. After sufficient pruning, we conduct environmental interventions to extrapolate potential test distribution. Finally, we perform last several rounds of model averages to further improve generalization. We provide multiple examples and theoretical analyses that underpin the universality and reliability of our proposal. Further, RGLT has been experimentally verified across various independent identically distributed (IID) and out-of-distribution (OOD) graph benchmarks.

Published
2024
Full Text
View/download PDF

30. Activator Heavy Solid Solution in Rigid Framework: An Effective Strategy Toward Highly Efficient and Thermally Stable Near‐Infrared Emission for Wireless Communication.

Author

Jia, Shiyu, Shao, Zifan, Zhang, Chuang, Wang, Wenjie, Li, Xinglin, Zhou, Rongfu, Zhou, Yayun, Deng, Tingting, Yu, Ting, and Song, Enhai

Subjects
*HELMHOLTZ free energy, *SOLID solutions, *OPTICAL communications, *ION emission, *ORGANIC products
Abstract

Broadband near‐infrared (NIR) phosphors are crucial components of next‐generation NIR lighting sources. However, the design of high‐efficiency and thermally stable NIR phosphors still poses a significant challenge, whose quantum efficiencies (QEs) are directly limited by their absorption efficiency (AE) toward incident light. Here, an efficient and thermally stable NIR emission with AE up to 64.9% and emission keeping of 91.23% at 423 K is demonstrated via Cr3+ heavy solid solution in rigid framework LiCaGaF6:Cr3+ (LCGFC). Isomorphic LiCaAlF6:Cr3+ also exhibits thermal robustness, while traps in low doping concentration and low QEs. Comparative studies on crystal structure, formation energy, and Helmholtz free energy disclose that Cr3+ substitution on equivalent and equiradius Ga3+ site versus radii differential Al3+ site generates heavier solid solution and sustainable structural rigidity with acquirement of higher AE and better thermal stability. Incorporating LCGFC with a blue InGaN chip, a NIR phosphor‐converted light‐emitting diode is fabricated to realize stable wireless optical communication with good penetrability through biological tissue and some organic products. These findings develop a strategy based on activator heavy solid solutions in a rigid framework to achieve high‐efficiency and thermally stable NIR phosphors but also advance their novel optoelectronic applications. [ABSTRACT FROM AUTHOR]

Published
2024
Full Text
View/download PDF

31. Research and Progress of Inorganic Infrared Electrochromic Materials and Devices.

Author

Li X, Yao R, Li M, Guo C, Luo D, Zhong J, Xu Z, Ning H, and Peng J

Abstract

Background: Electrochromic materials can dynamically change their optical properties (such as transmittance, absorbance, and reflectance under the action of an applied voltage, and their research and application in the visible band have been widely concerned. In recent years, with the continuous development of electrochromic technology, the related research has been gradually extended to the infrared region., Objective: This invited review aims to provide an overview of the current status of several inorganic infrared electrochromic materials, to provide some references for future research, and to promote the research and application of electrochromic technology in the infrared region., Methods: This review summarizes various research results in the field of infrared electrochromic, which includes a detailed literature review and patent search. Starting from the key performance parameters and device structure characteristics of infrared electrochromic devices (ECDs), the research and progress of several types of inorganic infrared electrochromic materials, including metal oxides, plasma nanocrystals, and carbon nanomaterials, are mainly presented, and feasible optimization directions are also discussed., Conclusion: We believe that the potential of these materials for civilian and military applications, for example, infrared electrochromic smart windows, infrared stealth/disguise, and thermal control of spacecraft, can be fully exploited by optimizing the materials and their devices to improve their performance., (Copyright© Bentham Science Publishers; For any queries, please email at epub@benthamscience.net.)

Published
2024
Full Text
View/download PDF

Catalog

Books, media, physical & digital resources
See catalog results