检索结果-南通市图书馆

Mitigate noisy data for smart IoT via GAN based machine unlearning

在线全文

维普期刊数据库

学校读者我要写书评

暂无评论

Science China(Information Sciences) 2024年第3期67卷 62-78页

作者： Zhuo MA Yilong YANG Yang liu Xinjing liu Jianfeng MA School of Cyber Engineering Xidian University State Key Laboratory of Integrated Services Networks (ISN)

With the development of IoT applications, machine learning dramatically improves the utility of variable IoT systems such as autonomous driving. Although the pretrain-finetune framework can cope well with data heterogeneity in complex IoT scenarios, the data collected by sensors often contain unexpected noisy data, e.g., out-of-distribution(OOD) data, which leads to the reduced performance of fine-tuned models. To resolve the problem, this paper proposes Mu GAN, a method that can mitigate the side-effect of OOD data via the generative adversarial network(GAN)-based machine unlearning. Mu GAN follows a straightforward but effective idea to mitigate the performance loss caused by OOD data, i.e., “flashbacking” the model to the condition where OOD data are excluded from model training. To achieve the goal, we design an adversarial game, where a discriminator is trained to identify whether a sample belongs to the training set by observing the confidence score. Meanwhile, a generator(i.e., the target model) is updated to fool the discriminator into believing that the OOD data are not included in the training set but others do. The experimental results show that benefiting from the high unlearning rate(more than 90%) and retention rate(99%), Mu GAN succeeds in lowering the model performance degradation caused by OOD data from 5.88% to 0.8%.

关键词： machine unlearning generative adversarial network out of distribution data Internet of Thing neural network

同方期刊数据库

Research on Anthropomorphic Obstacle Avoidance Trajectory Planning for Adaptive Driving Scenarios Based on Inverse Reinforcement Learning Theory

在线全文

学校读者我要写书评

暂无评论

Engineering 2024年第2期33卷 133-145页

作者： Jian Wu Yang Yan Yulong liu Yahui liu School of Mechanical and Automotive Engineering Liaocheng UniversityLiaocheng 252000China 2 School of Vehicle and Mobility Tsinghua UniversityBeijing 100084China

The forward design of trajectory planning strategies requires preset trajectory optimization functions,resulting in poor adaptability of the strategy and an inability to accurately generate obstacle avoidance trajectories that conform to real driver behavior *** addition,owing to the strong time-varying dynamic characteristics of obstacle avoidance scenarios,it is necessary to design numerous trajectory optimization functions and adjust the corresponding ***,an anthropomorphic obstacle-avoidance trajectory planning strategy for adaptive driving scenarios is ***,numerous expert-demonstrated trajectories are extracted from the HighD natural driving ***,a trajectory expectation feature-matching algorithm is proposed that uses maximum entropy inverse reinforcement learning theory to learn the extracted expert-demonstrated trajectories and achieve automatic acquisition of the optimization function of the expert-demonstrated ***,a mapping model is constructed by combining the key driving scenario information that affects vehicle obstacle avoidance with the weight of the optimization function,and an anthropomorphic obstacle avoidance trajectory planning strategy for adaptive driving scenarios is ***,the proposed strategy is verified based on real driving *** results show that the strategy can adjust the weight distribution of the trajectory optimization function in real time according to the“emergency degree”of obstacle avoidance and the state of the ***,this strategy can generate anthropomorphic trajectories that are similar to expert-demonstrated trajectories,effectively improving the adaptability and acceptability of trajectories in driving scenarios.

关键词： Obstacle avoidance trajectory planning Inverse reinforcement theory Anthropomorphic Adaptive driving scenarios

Effect of Pre-strain on Microstructure and Stamping Performance of High-strength Low-alloy Steel

在线全文

学校读者我要写书评

暂无评论

Journal of Wuhan University of Technology(Materials Science) 2024年第3期39卷 774-780页

作者：刘坡 WEN Zhicheng liu Zheng 许峰 School of Mechanical Engineering Guilin University of Aerospace TechnologyGuilin 541004China School of Automotive Engineer-ing Changshu Institute of TechnologyChangshu 215500China

In this study,pre-strain ranging from 0 to 0.12 was applied through uniaxial tension on high-strength low-alloy(HSLA)specimens with four kinds of grain *** of pre-strain and grain size on me-chanical property was investigated through tensile *** of the pre-strained and tensile tested samples were analyzed,*** 30.8°v-bending and following flattening,as well as Erichson cupping tests,were performed on the pre-strained *** show the elongation ratio of grain and dislocation density increases with *** platform is removed when pre-strain is larger than 0.06 while yielding plateau period decreases with pre-strain less than 0.06 due to reduction of pinning *** 30.8°v-bending and the following flattening tests are successfully accomplished on all the pre-strained samples with different grain *** in grain size,along with increase in pre-strain,causes increase in strength and decrease in elongation rate as well as cupping ***-strain causes very slight effect on bending ability,much less than that on mechanical property and cupping test *** impact of the pre-strain and grain size on HSLA steel deformability is inconspicuous.

关键词： strain hardening dislocation texture bending erichson cupping test

Identification of the Worst-case Static Voltage Stability Margin Interval of AC/DC Power System Considering Uncertainty of Renewables

在线全文

学校读者我要写书评

暂无评论

CSEE Journal of Power and Energy Systems 2024年第3期10卷 974-987页

作者： Wanbin liu Shunjiang Lin Yuerong Yang Mingbo liu Qifeng Li School of Electric Power Engineering South China University of TechnologyGuangzhou 510640China Guangdong Key Laboratory of Clean Energy Technology South China University of TechnologyGuangzhou 511458China Department of Electrical and Computer Engineering University of Central FloridaOrlandoFL 32816USA CSEE IEEE

Calculation of static voltage stability margin(SVSM)of AC/DC power systems with lots of renewable energy sources(RESs)integration requires consideration of uncertain load growth and renewable energy generation *** paper presents a bi-level optimal power flow(BLOPF)model to identify the worst-case SVSM of an AC/DC power system with line commutation converter-based HVDC and multi-terminal voltage sourced converter-based HVDC transmission *** of uncertain load growth’s hypercone model and control mode switching of DC converter stations are considered in the BLOPF ***,uncertain RES output fluctuations are described as intervals,and two three-level optimal power flow(TLOPF)models are established to identify interval bounds of the system worst-case *** two TLOPF models are both transformed into max–min bi-level optimization models according to independent characteristics of different uncertain ***,transforming the inner level model into its dual form,max–min BLOPF models are simplified to single-level optimization models for direct *** results on the modified IEEE-39 bus AC/DC case and an actual large-scale AC/DC case in China indicate correctness and efficiency of the proposed identification method.

关键词： AC/DC power system control mode switching convex relaxation dual programming multi-level optimization optimal power flow static voltage stability margin

Quasi-2D weak lensing cosmological constraints using the PDF-SYM method

在线全文

维普期刊数据库

学校读者我要写书评

暂无评论

Science China(Physics,Mechanics & Astronomy) 2024年第7期67卷 44-58页

作者： Zhenjie liu Jun Zhang Hekun Li Zhi Shen Cong liu Department of Astronomy Shanghai Jiao Tong UniversityShanghai 200240China Shanghai Key Laboratory for Particle Physics and Cosmology Shanghai 200240China Shanghai Astronomical Observatory Chinese Academy of SciencesShanghai 200030China

Cosmic shear statistics,such as the two-point correlation function(2PCF),can be evaluated with the PDF-SYM method instead of the traditional weighted-sum *** makes use of the full PDF information of the shear estimators,and does not require weightings on the shear estimators,which can in principle introduce additional systematic *** work presents our constraints on S8and?mfrom the shear-shear correlations using the PDF-SYM *** data we use is from the z-band images of the Dark Energy Camera Legacy Survey(DECa LS),which covers about 10000 deg2with more than 100 million *** shear catalog is produced by the Fourier Quad method,and well tested on the real data itself with the field-distortion *** main approach is called quasi-2D as we do use the photo-z information of each individual galaxy,but without dividing the galaxies into redshift *** mainly use galaxy pairs within the redshift interval between 0.2 and 1.3,and the angular range from 4.7 to 180 *** analysis yields S8=0.762±0.026 and?m=0.234±0.075,with the baryon effects and the intrinsic alignments *** results are robust against redshift *** check the consistency of our results by deriving the cosmological constraints from auto-correlations ofγ1andγ2separately,and find that they are consistent with each other,but the constraints from theγ1component are much weaker than that fromγ*** implies a much worse data quality ofγ1,which is likely due to additional shear uncertainties caused by CCD electronics(according to the survey strategy of DECa LS).We also perform a pure 2D analysis,which gives S8=0.81_(-0.04)^(+0.03)and?m=0.25_(-0.05)^(+0.06).Our findings demonstrate the potential of the PDF-SYM method for precision cosmology.

关键词： weak lensing shear-shear correlations cosmological parameters observations analysis

Modeling Method of Multi-Energy Systems Based on LSTM Algorithm

在线全文

学校读者我要写书评

暂无评论

CSEE Journal of Power and Energy Systems 2024年第4期10卷 1701-1709页

作者： Di Qiu Fei Chen Dong liu Min Cao Siyang liu Shanghai Jiao Tong University Shanghai 200240China Yunnan Power Grid Co. Ltd

The development of the Energy Internet has improved the efficiency of energy utilization and promoted sustainable development of power and energy *** multi-energy system modeling considering the dynamic process of transmission line is one of the key research points of Energy Internet operation *** the energy circuit theory,the lumped parameter model of natural gas pipelines is built and the dynamic characteristic parameters under the control instruction are *** with dynamic characteristic parameters,the long short-term memory(LSTM)neural network is designed to fit the natural gas pipeline dynamic process into discrete linear time-varying(LTV)*** with the equations,an energy hub method is used to build a control model of industrial parks with multi-energy distribution *** the rolling optimal control strategy given in this paper,the model is solved by the Matlab-Yalmip solver and rolling control instructions of each energy conversion unit are ***,the case study demonstrates that the LSTM neural network-based modeling method presented in this paper can accurately fit the dynamic process of a natural gas pipeline *** rolling control model of the multi-energy system can improve the efficiency of energy utilization,exhibit the transmission line status constraints during the optimization control process and improve reliability of the multi-energy system operation.

关键词： Energy hub LSTM multi-energy supply system pipeline system modeling

Recent advances in halide perovskite memristors:From materials to applications

在线全文

维普期刊数据库

学校读者我要写书评

暂无评论

Frontiers of physics 2024年第2期19卷 73-104页

作者： Sixian liu Jianmin Zeng Qilai Chen Gang liu Department of Micro/Nano Electronics School of Electronic Information and Electrical EngineeringShanghai Jiao Tong UniversityShanghai 200240China Aerospace Science&Industry Shenzhen(Group)Co.Ltd. Shenzhen 518000China

With the emergence of the Internet of Things(IoT)and the rapid growth of big data generated by edge devices,there has been a growing need for electronic devices that are capable of processing and transmitting data at low power and high *** Complementary Metal-Oxide-Semiconductor(CMOS)devices are nonvolatile and often limited by their ability for certain IoT applications due to their unnecessary power consumption for data movement in von Neuman architecture-based *** has led to a surge in research and development efforts aimed at creating innovative electronic components and systems that can overcome these shortcomings and meet the evolving needs of the information era,which share features such as improved energy efficiency,higher processing speeds,and increased *** are a novel type of electronic device that has the potential to break down the barrier between storage and *** storing data and processing information within the same device,memristors can minimize the need for data movement,which allows for faster processing speeds and reduced energy *** further improve the energy efficiency and reliability of memristors,there has been a growing trend toward diversifying the selection of dielectric materials used in *** perovskites(HPs)have unique electrical and optical properties,including ion migration,charge trapping effect caused by intrinsic defects,excellent optical absorption efficiency,and high charge mobility,which makes them highly promising in applications of *** this paper,we provide a comprehensive overview of the recent development in resistive switching behaviors of HPs and the underlying ***,we summarize the diverse range of HPs,their respective performance metrics,as well as their applications in various ***,we critically evaluate the current bottlenecks and possible opportunities in the future research of HP memristors.

关键词： halide perovskite memristor mechanism neuromorphic computing non-volatile memory

Peeling of Magneto-responsive Beams with Large Deformation Mediated by the Parallel Magnetic Field

在线全文

维普期刊数据库

学校读者我要写书评

暂无评论

Acta Mechanica Solida Sinica 2024年第3期37卷 457-467页

作者： Gongqi Cao Zhangna Xue Shiyang liu Yuchen Jin Jianlin liu College of Pipeline and Civil Engineering China University of Petroleum(East China)Qingdao266580China

Elasto-capillarity phenomena are prevalent in various industrial fields such as mechanical engineering,material science,aerospace,soft robotics,and *** this study,two typical peeling processes of slender beams driven by the parallel magnetic field are investigated based on experimental and theoretical *** first is the adhesion of two parallel beams,and the second is the self-folding of a long *** these two cases,the energy variation method on the elastica is used,and then,the governing equations and transversality boundary conditions are *** is shown that the analytical solutions are in excellent agreement with the experimental *** effects of magnetic induction intensity,distance,and surface tension on the deflection curve and peeling length of the elastica are fully *** results are instrumental in accurately regulating elasto-capillarity in structures and provide insights for the engineering design of programmable microstructures on surfaces,microsensors,and bionic robots.

关键词： Magneto-responsive beam Adhesion and peeling Self-folding Elastica Transversality condition

Off-Grid Compressed Channel Estimation with Parallel Interference Cancellation for Millimeter Wave Massive MIMO

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3期21卷 51-65页

作者： liu Jinru Tian Yongqing liu Danpu Zhang Zhilong Beijing Laboratory of Advanced Information Network Beijing Key Laboratory of Network System Architecture and ConvergenceBeijing University of Posts and TelecommunicationsBeijing100876China

Millimeter wave(mmWave)massive multiple-input multiple-output(MIMO)plays an important role in the fifth-generation(5G)mobile communications and beyond wireless communication systems owing to its potential of high ***,channel estimation has become very challenging due to the use of massive MIMO antenna ***,the mmWave channel has strong sparsity in the spatial angle domain,and the compressed sensing technology can be used to convert the original channel matrix into the sparse matrix of discrete angle *** the high-dimensional channel matrix estimation is transformed into a sparse recovery problem with greatly reduced computational ***,the path angle in the actual scene appears randomly and is unlikely to be completely located on the quantization angle grid,thus leading to the problem of power ***,multiple paths with the random distribution of angles will bring about serious interpath interference and further deteriorate the performance of channel *** address these off-grid issues,we propose a parallel interference cancellation assisted multi-grid matching pursuit(PIC-MGMP)algorithm in this *** proposed algorithm consists of three stages,including coarse estimation,refined estimation,and inter-path cyclic iterative inter-ference *** specifically,the angular resolution can be improved by locally refining the grid to reduce power leakage,while the inter-path interference is eliminated by parallel interference cancellation(PIC),and the two together improve the estimation *** results show that compared with the traditional orthogonal matching pursuit(OMP)algorithm,the normalized mean square error(NMSE)of the proposed algorithm decreases by over 14dB in the case of 2 paths.

关键词： channel estimation compressed sensing inter-path interference millimeter wave massive MIMO off-grid parallel interference cancellation