检索结果-南通市图书馆

A Blind Batch Encryption and Public Ledger-Based Protocol for Sharing Sensitive Data

在线全文

维普期刊数据库

学校读者我要写书评

暂无评论

China Communications 2024年第1期21卷 310-322页

作者： Zhiwei Wang Nianhua Yang Qingqing Chen Wei Shen Zhiying Zhang School of Computer sciences Nanjing University of Posts and TelecommunicationsNanjing 210023China Dept.of Intelligent science and Information Law East China University of Political Science and LawShanghai 201620China Key Laboratory of Computational science and Application of Hainan Province Hainan Normal UniversityHaikou 571158China State Key Laboratory of Public Big Data Guizhou UniversityGuizhou 550025China Yunnan Key Laboratory of Blockchain Application technology Yunnan Innovation Research Institute of Beijing University of Aeronautics and AstronauticsKunming 650233

For the goals of security and privacy preservation,we propose a blind batch encryption-and public ledger-based data sharing protocol that allows the integrity of sensitive data to be audited by a public ledger and allows privacy information to be *** owners can tightly manage their data with efficient revocation and only grant one-time adaptive access for the fulfillment of the *** prove that our protocol is semanticallly secure,blind,and secure against oblivious requesters and malicious file *** also provide security analysis in the context of four typical attacks.

关键词： blind batch encryption data sharing onetime adaptive access public ledger security and privacy

Smaller & Smarter: Score-Driven Network Chaining of Smaller Language Models

在线全文

学校读者我要写书评

暂无评论

Journal of Software Engineering and Applications 2024年第1期17卷 23-42页

作者： Gunika Dhingra Siddansh Chawla Vijay K. Madisetti Arshdeep Bahga School of Computer science Engineering & technology Bennett University Greater Noida India School of Cybersecurity and Privacy Georgia Institute of Technology Atlanta USA Cloudemy technology Labs Chandigarh India

With the continuous evolution and expanding applications of Large Language Models (LLMs), there has been a noticeable surge in the size of the emerging models. It is not solely the growth in model size, primarily measured by the number of parameters, but also the subsequent escalation in computational demands, hardware and software prerequisites for training, all culminating in a substantial financial investment as well. In this paper, we present novel techniques like supervision, parallelization, and scoring functions to get better results out of chains of smaller language models, rather than relying solely on scaling up model size. Firstly, we propose an approach to quantify the performance of a Smaller Language Models (SLM) by introducing a corresponding supervisor model that incrementally corrects the encountered errors. Secondly, we propose an approach to utilize two smaller language models (in a network) performing the same task and retrieving the best relevant output from the two, ensuring peak performance for a specific task. Experimental evaluations establish the quantitative accuracy improvements on financial reasoning and arithmetic calculation tasks from utilizing techniques like supervisor models (in a network of model scenario), threshold scoring and parallel processing over a baseline study.

关键词： Large Language Models (LLMs) Smaller Language Models (SLMs) Finance Networking Supervisor Model Scoring Function

Optimization of variables and assessment of in-vitro and in-vivo antihyperlipidemic activity of Eudragit RS nanoparticles containing simvastatin

在线全文

维普期刊数据库

学校读者我要写书评

暂无评论

Particuology 2024年第7期90卷 189-198页

作者： Bina Gidwani Vishal Jain Veenu Joshi Ravindra Kumar Pandey Shiv Shankar Shukla Amber Vyas Columbia institute of Pharmacy Raipur(C.G.)India University institute of Pharmacy Pt Ravishankar Shukla UniversityRaipur(C.G.)India Center for Basic science and Research Pt Ravishankar Shukla UniversityRaipur(C.G.)India

Simvastatin, a BCS class II drug, is associated with poor aqueous solubility, first-pass metabolism and short half-life. In the present work, nanoparticles were prepared and evaluated. Optimization of formulation and process parameters was done through the use of independent and dependent variables. Preliminary studies were done to determine suitable range of the concentration of Eudragit polymer (10%–30%) and the ratio of drug to polymer (1:1 to 1:5) for the formation of nanoparticles by emulsification and a solvent evaporation technique. Results revealed that the mean size of nanoparticles was affected by stirring speed from 5000 RPM, 8000 RPM, and 12000 RPM. The results of increase in association efficiency and percent yield with increase in amount of drug from 100 mg, 150 mg, and 200 mg in selected range were observed. In-vitro release studies showed that two formulations possess highest initial burst and slow sustained drug release. There was 2.93-fold decrease in total cholesterol and 3.27-fold increases in triglyceride level during in-vivo study.

关键词： Simvastatin Eudragit RS 100 nanoparticles Hyperlipidemia

Variational Data Assimilation Method Using Parallel Dual Populations Particle Swarm Optimization Algorithm

在线全文

维普期刊数据库

学校读者我要写书评

暂无评论

Wuhan University Journal of Natural sciences 2024年第1期29卷 59-66页

作者： WU Zhongjian LI Junyan Detroit Green institute of technology Hubei University of TechnologyWuhan 430068HubeiChina School of Information Management Central China Normal UniversityWuhan 430079HubeiChina

In recent years,numerical weather forecasting has been increasingly *** data assimilation furnishes precise initial values for numerical forecasting models,constituting an inherently nonlinear optimization *** enormity of the dataset under consideration gives rise to substantial computational burdens,complex modeling,and high hardware *** paper employs the Dual-Population Particle Swarm Optimization(DPSO)algorithm in variational data assimilation to enhance assimilation *** harnessing parallel computing principles,the paper introduces the Parallel Dual-Population Particle Swarm Optimization(PDPSO)Algorithm to reduce the algorithm processing *** were carried out using partial differential equations,and comparisons in terms of time and accuracy were made against DPSO,the Dynamic Weight Particle Swarm Algorithm(PSOCIWAC),and the TimeVarying Double Compression Factor Particle Swarm Algorithm(PSOTVCF).Experimental results indicate that the proposed PDPSO outperforms PSOCIWAC and PSOTVCF in convergence accuracy and is comparable to *** processing time,PDPSO is 40%faster than PSOCIWAC and PSOTVCF and 70%faster than DPSO.

关键词： parallel algorithm variational data assimilation dual-population particle swarm optimization algorithm diffusion mechanism

Evaluating Privacy Leakage and Memorization Attacks on Large Language Models (LLMs) in Generative AI Applications

在线全文

学校读者我要写书评

暂无评论

Journal of Software Engineering and Applications 2024年第5期17卷 421-447页

作者： Harshvardhan Aditya Siddansh Chawla Gunika Dhingra Parijat Rai Saumil Sood Tanmay Singh Zeba Mohsin Wase Arshdeep Bahga Vijay K. Madisetti School of Computer science Engineering & technology Bennett University Greater Noida India Cloudemy technology Labs Chandigarh India School of Cybersecurity and Privacy Georgia Institute of Technology Atlanta USA

The recent interest in the deployment of Generative AI applications that use large language models (LLMs) has brought to the forefront significant privacy concerns, notably the leakage of Personally Identifiable Information (PII) and other confidential or protected information that may have been memorized during training, specifically during a fine-tuning or customization process. We describe different black-box attacks from potential adversaries and study their impact on the amount and type of information that may be recovered from commonly used and deployed LLMs. Our research investigates the relationship between PII leakage, memorization, and factors such as model size, architecture, and the nature of attacks employed. The study utilizes two broad categories of attacks: PII leakage-focused attacks (auto-completion and extraction attacks) and memorization-focused attacks (various membership inference attacks). The findings from these investigations are quantified using an array of evaluative metrics, providing a detailed understanding of LLM vulnerabilities and the effectiveness of different attacks.

关键词： Large Language Models PII Leakage Privacy Memorization Overfitting Membership Inference Attack (MIA)

Magmatic-tectonic response of the South China Craton to the Paleo-Pacific subduction during the Triassic:a new viewpoint based on Well NK-1

在线全文

维普期刊数据库

学校读者我要写书评

暂无评论

Journal of Oceanology and Limnology 2024年第1期42卷 58-89页

作者： Chan WANG Hao LIU Gang LI Zhen SUN Tingting GONG Li MIAO Xiaowei ZHU Yunying ZHANG Weihai XU Wen YAN Key Laboratory of Ocean and Marginal Sea Geology South China Sea Institute of OceanologyChinese Academy of SciencesGuangzhou 510301China Southern Marine science and Engineering Guangdong Laboratory(Guangzhou) Guangzhou 511458China Key Laboratory of Tectonic Controlled Mineralization and Oil Reservoir of Ministry of Natural Resources Chengdu University of TechnologyChengdu 610059China School of Earth sciences and Resources China University of Geosciences(Beijing)Beijing 100083China College of Earth science Chengdu University of TechnologyChengdu 610059China University of Chinese Academy of sciences Beijing 100049China

The Nansha Block(NB)is one of the blocks separated from the southern margin of the South China Craton(SCC)by the western Pacific subduction,which contains rich information of geodynamic and tectonic *** reveal the essence of western Paleo-Pacific subduction during the Triassic period,Well NK-1 in this block was selected for petrographic study,and published research data from other cooperative teams were compared.A double-cycle pattern of basic to neutral magmatic volcanism was established,and 36 lithological rhythmic layers and representative cryptoexplosive breccia facies and welded tuff bands were *** with a reanalysis of published geochronological data,geochemical elements,and isotope geochemistry,we found that the rock assemblages could be divided into an intermediate-acid dacite(DA)series(SiO_(2)>65%)and basaltic(BA)series(Co<40μg/g),which was formed during the early Late Triassic((218.6±3.2)–(217.9±3.5)Ma).BA exhibits obvious calc-alkaline island-arc magmatic properties:(^(87)Sr/^(86)Sr)_i ratio ranging 0.70377–0.71118(average:0.70645),^(147)Sm/^(144)Nd ratio ranging 0.119–0.193(average:0.168),and chondrite-normalized rare earth element(REE)curves being flat,while DA exhibits remarkable characteristics of subducted island-arc andesitic magma:(^(87)Sr/^(86)Sr)_i ratio(0.70939–0.71129;average:0.71035),εNd(t)value(-6.2–-4.8;average:-5.6)andε_(Hf)(t)value(-2.9–-1.7,average:-2.2)show obvious crust-mantle mixing *** and DA reveal typical characteristics of island-arc magma systems and typeⅡenriched mantle(EM-Ⅱ)*** magma was likely resulted from the process whereby the continental crust frontal accretionary wedge was driven by the Paleo-Pacific slab subduction into the deep and began to melt,resulting in a large amount of melt(fluid)joined the asthenosphere on the side of the continental *** contrast,DA magma was likely resulted from the process whereby the plate front was forced to bend with increasing subducti

关键词： Paleo-Pacific tectonic domain tectonic evolution rift volcanic rock Nansha Block South China Craton

SiO_(2) Induces Iron Overload and Ferroptosis in Cardiomyocytes in a Silicosis Mouse Model

在线全文

学校读者我要写书评

暂无评论

Biomedical and Environmental sciences 2024年第6期37卷 617-627页

作者： Yongheng Wang Ning Li Yi Guan Tong LI Yuxiu Zhang Hong Cao Zhihua Yu Zhiheng Li Shuoyan Li Jiahao Hu Wenxin Zhou Sisi Qin Shuang Li Sanqiao Yao School of Public Health North China University of Science and TechnologyTangshan 063210HebeiChina Hebei Key Laboratory of Occupational Health and Safety for Coal Industry North China University of Science and TechnologyTangshan 063210HebeiChina Occupational Disease Prevention and Control institute of Jinneng Holding Coal Industry Group Co. LtdDatong 037001ShanxiChina China Pingmei Shenma Group Occupational Disease Prevention and Control Hospital Pingdingshan 467000HenanChina School of Public Health Xinxiang Medical UniversityXinxiang 453003HenanChina

Objective The aim of this study was to explore the role and mechanism of ferroptosis in SiO_(2)-induced cardiac injury using a mouse *** Male C57BL/6 mice were intratracheally instilled with SiO_(2) to create a silicosis ***-1(Fer-1)and deferoxamine(DFO)were used to suppress *** biomarkers,oxidative stress markers,histopathology,iron content,and the expression of ferroptosis-related proteins were *** SiO_(2) altered serum cardiac injury biomarkers,oxidative stress,iron accumulation,and ferroptosis markers in myocardial ***-1 and DFO reduced lipid peroxidation and iron overload,and alleviated SiO_(2)-induced mitochondrial damage and myocardial ***_(2) inhibited Nuclear factor erythroid 2-related factor 2(Nrf2)and its downstream antioxidant genes,while Fer-1 more potently reactivated Nrf2 compared to *** Iron overload-induced ferroptosis contributes to SiO_(2)-induced cardiac *** ferroptosis by reducing iron accumulation or inhibiting lipid peroxidation protects against SiO_(2) cardiotoxicity,potentially via modulation of the Nrf2 pathway.

关键词： SiO_(2) exposure Iron overload Ferroptosis Cardiac injury Nrf2

Room-temperature ferromagnetism and half-metallicity in monolayer orthorhombic CrS_(2)

在线全文

学校读者我要写书评

暂无评论

Frontiers of physics 2024年第4期19卷 137-146页

作者：雷博程李奥林周文哲王云鹏熊威陈羽欧阳方平 School of Physics science and technology and Xinjiang Key Laboratory of Solid-State Physics and DevicesXinjiang UniversityUrumqi 830046China School of Physics and Hunan Key Laboratory for Super-Microstructure and Ultrafast Processand Hunan Key Laboratory of Nanophotonics and DevicesCentral South UniversityChangsha 410083China School of Physics science and technology and Xinjiang Laboratory of Phase Transitions and Microstructures in Condensed Matter PhysicsYili Normal UniversityYining 835000China Powder Metallurgy Research institute and State Key Laboratory of Powder Metallurgy Central South UniversityChangsha 410083China

Two-dimensional materials with high-temperature ferromagnetism and half-metallicity have the latest applications in spintronic *** on first-principles calculations,we have investigated a novel two-dimensional CrS_(2) phase with an orthorhombic *** results suggest that it is stable in dynamics,thermodynamics,and *** ground state of monolayer orthorhombic CrS_(2) is both ferromagnetic and half-metallic,with a high Curie temperature of 895 K and a large spin-flipping gap on values of 0.804 *** room-temperature ferromagnetism and halfmetallicity can maintain stability against a strong biaxial strain ranging from–5%to 5%.Meanwhile,increasing strain can significantly maintain the out-of-plane magnetic anisotropy.A density of states analysis,together with the orbital-resolved magnetic anisotropy energy,has revealed that the strain-enhanced MAE is highly related to the 3d-orbital splitting of Cr *** results suggest the monolayer orthorhombic CrS_(2) is an ideal candidate for future spintronics.

关键词： orthorhombic CrS_(2) Curie temperature magnetic anisotropy energy biaxial strain first-principles calculations

GUARDIAN: A Multi-Tiered Defense Architecture for Thwarting Prompt Injection Attacks on LLMs

在线全文

维普期刊数据库

学校读者我要写书评

暂无评论

Journal of Software Engineering and Applications 2024年第1期17卷 43-68页

作者： Parijat Rai Saumil Sood Vijay K. Madisetti Arshdeep Bahga School of Computer science Engineering & technology Bennett University Greater Noida India School of Cybersecurity and Privacy Georgia Institute of Technology Atlanta USA Cloudemy technology Labs Chandigarh India

This paper introduces a novel multi-tiered defense architecture to protect language models from adversarial prompt attacks. We construct adversarial prompts using strategies like role emulation and manipulative assistance to simulate real threats. We introduce a comprehensive, multi-tiered defense framework named GUARDIAN (Guardrails for Upholding Ethics in Language Models) comprising a system prompt filter, pre-processing filter leveraging a toxic classifier and ethical prompt generator, and pre-display filter using the model itself for output screening. Extensive testing on Meta’s Llama-2 model demonstrates the capability to block 100% of attack prompts. The approach also auto-suggests safer prompt alternatives, thereby bolstering language model security. Quantitatively evaluated defense layers and an ethical substitution mechanism represent key innovations to counter sophisticated attacks. The integrated methodology not only fortifies smaller LLMs against emerging cyber threats but also guides the broader application of LLMs in a secure and ethical manner.

关键词： Large Language Models (LLMs) Adversarial Attack Prompt Injection Filter Defense Artificial Intelligence Machine Learning Cybersecurity