HybridGAD: Identification of AI-Generated Radiology Abstracts Based on a Novel Hybrid Model with Attention Mechanism

Class Title:Radiological imaging method a comprehensive overview purpose.This GPT paper provides an overview of the different forms of radiological imaging and the potential diagnosis capabilities they offer as well as recent advances in the field.Materials and Methods:This paper provides an overview of conventional radiography digital radiography panoramic radiography computed tomography and cone-beam computed tomography.Additionally recent advances in radiological imaging are discussed such as imaging diagnosis and modern computer-aided diagnosis systems.Results:This paper details the differences between the imaging techniques the benefits of each and the current advances in the field to aid in the diagnosis of medical conditions.Conclusion:Radiological imaging is an extremely important tool in modern medicine to assist in medical diagnosis.This work provides an overview of the types of imaging techniques used the recent advances made and their potential applications.

Generalized Plastic Mechanics and Its Application

The development of geotechnical plasticity is reviewed and some problems of applying the classical plastic mechanics (CPM) to geomaterials are analyzed, and then CPM’s three hypotheses not fitted the deformation mechanism of geomaterials are pointed out. By giving up the three hypotheses, a generalized plastic potential theory can be obtained from solid mechanics directly, and then the traditional plastic mechanics can be changed to a more generalized plastic mechanics, namely generalized plastic mechanics (GPM). The GPM adopts the component theory as theoretical base, so it can reflect the influence of transition of stress path. The unreasonable phenomena such as excessive dilatancy caused by adopting the normality-flow law can be avoided, and the error caused by the arbitrary assumption of plastic potential surfaces cannot be produced. The yield surface theory, hardening laws and stress-strain relations of GPM are given, and a GPM including the rotation of principal stress axes is also established. It is pointed out that the yield condition is a state parameter as well as a test parameter, and it can only be given by test. After the practical application, it is shown that the GPM cannot only be applied to the modeling theory of geomaterials but also to other fields of geomechanics such as limit analysis.

Modeling hot strip rolling process under framework of generalized additive model

This research develops a new mathematical modeling method by combining industrial big data and process mechanism analysis under the framework of generalized additive models(GAM)to generate a practical model with generalization and precision.Specifically,the proposed modeling method includes the following steps.Firstly,the influence factors are screened using mechanism knowledge and data-mining methods.Secondly,the unary GAM without interactions including cleaning the data,building the sub-models,and verifying the sub-models.Subsequently,the interactions between the various factors are explored,and the binary GAM with interactions is constructed.The relationships among the sub-models are analyzed,and the integrated model is built.Finally,based on the proposed modeling method,two prediction models of mechanical property and deformation resistance for hot-rolled strips are established.Industrial actual data verification demonstrates that the new models have good prediction precision,and the mean absolute percentage errors of tensile strength,yield strength and deformation resistance are 2.54%,3.34%and 6.53%,respectively.And experimental results suggest that the proposed method offers a new approach to industrial process modeling.

Mathematical Model and Geometrical Model of Double Pitch ZN-type Worm Gear Set Based on Generation Mechanism

The current researches on the tooth surface mathematical equations and the theory of gearing mainly pay attention to the ordinary type worm gear set（e.g., ZN, ZA, or ZK）. The research of forming mechanism and three-dimensional modeling method for the double pitch worm gear set is not enough. So there are some difficulties in mathematical model deducing and geometry modeling of double pitch ZN-type worm gear set based on generation mechanism. In order to establish the mathematical model and the precise geometric model of double pitch ZN-type worm gear set, the structural characteristics and generation mechanism of the double pitch ZN-type worm gear set are investigated. Mathematical model of the ZN-type worm gear set is derived based on its generation mechanism and the theory of gearing. According to the mathematical model of the worm gear set which has been developed, a geometry modeling method of the double pitch ZN-type worm and worm gear is presented. Furthermore, a geometrical precision calculate method is proposed to evaluate the geometrical quality of the double pitch worm gear set. As a result, the maximum error is less than 6′10–4 mm in magnitude, thus the model of the double pitch ZN-type worm gear set is available to meet the requirements of finite element analysis and engineering application. The derived mathematical model and the proposed geometrical modeling method are helpful to guiding the design, manufacture and contact analysis of the worm gear set.

基于Matlab-Sim Mechanics的机电产品组成建模与仿真技术研究

分析了目前几种产品建模方法的特点。比较了机电一体化产品组成的现有划分方法。针对以广义执行机构为主功能的机电一体化产品组成划分 ,提出了结合Matlab软件的SimMechanics工具箱建立机电一体化产品的广义执行机构模型、传感器模型和控制策略模型的机电产品组成对象建模的方法 ,为机电一体化产品方案生成后的进一步仿真分析奠定了基础。文章以特定的机电一体化产品———家用多功能缝纫机的横针机构子系统为例 ,阐明了这种建模方法和仿真技术的具体应用 。

Modeling mechanism of a novel fractional grey mode based on matrix analysis

To fully display the modeling mechanism of the novelfractional order grey model （FGM （q,1））, this paper decomposesthe data matrix of the model into the mean generation matrix, theaccumulative generation matrix and the raw data matrix, whichare consistent with the fractional order accumulative grey model（FAGM （1,1））. Following this, this paper decomposes the accumulativedata difference matrix into the accumulative generationmatrix, the q-order reductive accumulative matrix and the rawdata matrix, and then combines the least square method, findingthat the differential order affects the model parameters only byaffecting the formation of differential sequences. This paper thensummarizes matrix decomposition of some special sequences,such as the sequence generated by the strengthening and weakeningoperators, the jumping sequence, and the non-equidistancesequence. Finally, this paper expresses the influences of the rawdata transformation, the accumulation sequence transformation,and the differential matrix transformation on the model parametersas matrices, and takes the non-equidistance sequence as an exampleto show the modeling mechanism.

基于主动尾流控制的风电机群协同优化调度

针对机组间尾流效应严重影响风电机组发电效率的问题,提出了风电机组安全偏航约束计算方法、尾流特性混合半机理建模方法以及风电机群多目标协同优化调度方法。基于FAST.FARM平台完善了多自由度可控机组与尾流的动态交互集成仿真环境,对比分析了2台机组串列式排布以及华东地区某海上风电场7台机组实际排布下的协同运行优化性能。结果表明:所建立的集成仿真模型能够合理表征风电机群与空气流场的多领域动态交互特性,所提方法能够有效提升风电机群发电效能,促进经济效益、资源利用和成本控制的均衡优化。

多模态模型嵌入知识生成的逻辑机理与路径选择

基于海量数据与强大算法的多模态模型依赖复杂的结构框架,在嵌入知识生成的过程中展现出跨域数据处理能力和持续输出创新数据等核心特征。高度智能化的知识数据整理、自适应多场景的知识表达与动态协同聚合的知识共享等,推进知识生成逻辑从既有信息理解向全新知识产出的转变。然而,多模态模型逐渐暴露出缺乏合法可靠的知识数据源、透明可释的生成过程以及高质量内容输出等诸多问题。亟待通过搭建价值对齐的生成模型、提升可控生成的技术水平、完善人类反馈提示策略、构筑基于规则的管控体系等约束其更好地支撑知识生成,加速创新驱动新质生产力发展的进程。

异质性数字化投入对制造业绩效的影响机理——基于文本分析的实证方法

基于2016-2021年制造业上市企业平衡面板数据,以异质性数字化投入为切入点,运用文本分析方法和面板回归模型,研究通用型数字化投入和专用型数字化投入对企业绩效的影响及作用机制。研究发现:第一,通用型数字化投入与企业绩效之间存在非线性关系(先上升再下降),且投入临界值为0.3678;第二,专用型数字化投入与企业绩效之间关系不显著;第三,管理效率优化是企业数字化投入作用于企业绩效的中间机制,在通用型数字化投入影响企业绩效的过程中发挥非线性中介传导作用,而商业模式创新在目前尚未发挥有效作用。研究结论表明,经过通用型数字化投入的“红利期”,其对企业绩效的正向影响呈递减趋势。

Proposed Wave Momentum Source for Generating the 22-Year Solar Cycle

For the 22-year solar cycle oscillation there is no external time dependent source. A nonlinear oscillation, the solar cycle must be generated internally, and Babcock-Leighton models apply an artificial nonlinear source term that can simulate the observations—which leaves open the question of the actual source mechanism for the solar cycle. Addressing this question, we propose to take guidance from the wave mechanism that generates the 2-year Quasi-biennial Oscillation (QBO) in the Earth atmosphere. Upward propagating gravity waves, eastward and westward, deposit momentum to generate the observed zonal wind oscillation. On the Sun, helioseismology has provided a thorough understanding of the acoustic p-waves, which propagate down into the convective envelope guided by the increasing temperature and related propagation velocity. Near the tachocline with low turbulent viscosity, the waves propagating eastward and westward can produce an axisymmetric 22-year oscillation of the zonal flow velocities that can generate the magnetic solar dynamo. Following the Earth model, waves in opposite directions can generate in the Sun wind and magnetic field oscillations in opposite directions, the proposition of a potential solar cycle mechanism.

基于条件对抗增强的Transformer煤矿微震定位方法

随着人工智能技术的发展以及煤矿微震监测系统的广泛应用,越来越多的深度学习模型被应用到煤矿微震事件震源定位问题的求解上。然而,由于目前的微震数据量小且数据单一不足以训练大且深的神经网络模型,而小且浅的神经网络模型也不足以表征受多方因素影响的微震事件的震源,因而导致了定位模型定位精度低和鲁棒性弱,在实际生产生活中表现较差,严重地阻碍了深度学习模型在微震定位领域上的发展。针对上述问题,提出一种基于条件对抗增强的Transformer煤矿微震定位方法CGAN-Transformer,该方法首先通过一个CGAN架构的网络模型将数据量少且单一的微震数据增强成数据量庞大且具有一定多样性的微震数据;其次,利用Transformer编码器层将微震波形数据转换为特征数据后再利用其注意力机制进一步学习微震波形数据深层次特征和复杂的站间依赖关系,同时也利用高斯分布随机变量抵消了不同地质条件对定位精度的影响;最后,通过引入混合密度输出层获取高斯分布参数,计算最优的震源位置。在智利和辽宁某矿数据集上的实验结果验证了该方法的有效性,结果表明该方法所获得的震中误差与震源误差均优于其他方法,在两个数据集上的定位误差分别降低了38%和12%,达到了提高震源定位精度和定位模型鲁棒性的目的。

基于教师-学生模型的点云目标检测算法

针对现有点云目标检测算法对室外自动驾驶场景中遮挡物体检测精度低的问题,提出一种基于知识蒸馏和注意力增强的点云目标检测算法。以CIA-SSD模型为基础,设计了一种基于教师-学生模型的密集特征生成模块,提出基于交并比匹配策略的密集数据生成方法,将稀疏特征转换为密集特征。密集特征生成模块位于学生模型中,学生模型在教师模型生成的软目标监督训练下,推断出完整的密集目标特征,实现目标特征的补全;在教师模型中设计空间注意力和通道注意力机制,增强密集目标点云,提升特征图的质量。在KITTI数据集上的验证实验结果表明:与SE-SSD模型和CIA-SSD模型相比,本文提出的算法保持了单阶段目标检测速度的优势,同时明显提升了检测精度。

基于Transformer和GAN的对抗样本生成算法

对抗攻击与防御是计算机安全领域的一个热门研究方向。针对现有基于梯度的对抗样本生成方法可视质量差、基于优化的方法生成效率低的问题,提出基于Transformer和生成对抗网络(GAN)的对抗样本生成算法Trans-GAN。首先利用Transformer强大的视觉表征能力,将其作为重构网络,用于接收干净图像并生成攻击噪声;其次将Transformer重构网络作为生成器,与基于深度卷积网络的鉴别器相结合组成GAN网络架构,提高生成图像的真实性并保证训练的稳定性,同时提出改进的注意力机制Targeted Self-Attention,在训练网络时引入目标标签作为先验知识,指导网络模型学习生成具有特定攻击目标的对抗扰动;最后利用跳转连接将对抗噪声施加在干净样本上,形成对抗样本,攻击目标分类网络。实验结果表明:Trans-GAN算法针对MNIST数据集中2种模型的攻击成功率都达到99.9%以上,针对CIFAR10数据集中2种模型的攻击成功率分别达到96.36%和98.47%,优于目前先进的基于生成式的对抗样本生成方法;相比快速梯度符号法和投影梯度下降法,Trans-GAN算法生成的对抗噪声扰动量更小,形成的对抗样本更加自然,满足人类视觉不易分辨的要求。

机理约束下钻井机械钻速智能预测泛化方法

钻井机械钻速的准确预测可辅助油气井钻井前科学配置资源,对制订更加合理的钻井作业方案以及钻井提效、降本增效具有重要现实意义。智能化预测钻井机械钻速已成为行业研究热点,为解决常规智能模型在不同井间迁移能力较差的问题,在对综合录井数据进行降噪、补全等预处理的基础上,利用钻井专业知识构造约束条件,引入了域对抗神经网络(DANN),建立了机械钻速模型在不同井间的迁移机制,结合滑动窗口、增量更新与实时录井数据,形成了机械钻速模型随井下工况的实时更新方法。研究结果表明:(1)数据层约束和网络层约束均可提高智能模型的精度与稳定性,且双机理约束下的BP模型相比于普通BP模型预测精度明显提高;(2)基于域对抗神经网络的机械钻速预测模型可有效地将邻井(源域)数据知识迁移到测试井(目标域);(3)基于增量学习算法建立的双滑动窗口数据更新机制,使模型实时适应地下钻进环境变化,预测精度和泛化能力进一步提升;(4)机理约束、迁移训练与实时更新对模型泛化性能的强化作用具有叠加效应,新井机械钻速预测平均相对误差降低至20.2%。结论认为,建立的机械钻速预测模型及迁移方法相较于传统钻速预测模型,具有更好的迁移性和更高的准确度,减少了迁移过程中重复训练时间,为机械钻速智能预测提供了新的思路和方向。

碲镉汞红外探测器前置放大电路低温稳定性研究

在航天工程应用过程中,发现室温下HgCdTe红外探测器前置放大电路输出正常,但是当温度降低至-55℃时,输出有振荡现象,对探测器施加红外辐照后振荡现象加剧。文章首次对该现象进行系统研究,提出了HgCdTe红外探测器电路在低温下光照后的等效电路模型,从相位裕度与增益裕度角度分析前置放大电路在低温下产生振荡的机理,并在此基础上,提出了抑制低温光照后前置放大电路振荡的解决方法,实验结果验证了该方法的有效性。

池式铅冷快堆SGTR事故多组分多相流动过程数值模拟研究

本研究使用欧拉坐标下的多组分多相分析程序ACENA,首先介绍了ACENA程序的基本数学物理模型,然后通过铅铋-氮气两相流动实验HESTIA-2、KYLIN-Ⅱ-S铅铋-水相互作用实验和点堆中子动力学方程解析解,对程序热工水力模块和中子动力学模块进行了验证计算,在此基础上,针对欧洲先进铅冷示范堆ALFRED的设计方案分别开展了热态满功率稳态校核计算和假想无保护蒸汽发生器传热管破裂(SGTR)事故瞬态模拟,重点关注了SGTR事故后铅池内多相流动过程以及包壳最高温度、燃料最高温度、堆芯相对功率以及主容器压力等参数的演变,并分析了断管数量、铅冷却剂循环路径以及所采用的机理模型等影响因素对ACENA程序计算结果的影响。本文研究结果表明,Ishii-Chawla-Suzuki相间曳力系数模型结合Ishii等提出的相间界面面积浓度输运模型能够较好地模拟圆形/环形铅铋流道中上升气泡的扩散迁移特性;通过对KYLIN-Ⅱ-S实验的模拟说明ACENA程序能够较为合理地预测熔融铅基合金-水相互作用过程中,铅池内压力波动和温度瞬变等现象;ACENA程序对ALFRED堆稳态满功率下关键热工参数的计算结果与国际认可的一维系统程序TRACE/FRED的计算结果基本一致,证明了ACENA程序全堆级计算结果的可靠性;对ALFRED堆假想SGTR事故的计算验证了ACENA程序对铅冷快堆SGTR事故下复杂多组分多相流动现象的模拟能立,且计算结果表明合理设计一次侧冷却剂循环路径、尽可能降低管道破损数量均对消减铅冷快堆SGTR事故后果具有重要意义。本工作可为我国池式铅冷快堆SGTR事故安全分析提供技术参考。

基于随机块模型的社区隐藏统一框架

社区检测是复杂网络分析的重要工具之一,可帮助深入了解网络的社区结构和节点间潜在的关系,但同时也带来了隐私泄露问题.社区隐藏作为社区检测的伴生问题,旨在以最小的边扰动代价破坏网络的社区结构,近年来受到越来越多学者的关注.但现有的社区隐藏方法忽略了网络的生成机制且缺少针对不同尺度隐藏的统一框架,因此提出了一种基于随机块模型的社区隐藏(community hiding-stochastic block model,HC-SBM)算法,该算法从网络生成机制角度构建了社区隐藏的统一框架,即实现微观(个体)、介观(社区)、宏观(网络)3个尺度上的社区检测算法攻击.其基本思想是基于随机块模型刻画网络的生成机制,特别是网络社区形成和分裂的规律和模式,挖掘生成过程中的关键性链接以及链接集合,最终通过最小代价扰动策略破坏网络社区结构.通过在真实网络上的大量实验,并与4种先进的基准算法进行比较,表明了提出的HC-SBM算法在社区隐藏效果更优.

微流体装置力源机构的优化设计

为解决微流控系统中流体介质贮存装置输出压力不稳定等问题,设计了一种柔顺机构作为装置的力源机构。该柔顺机构基于正负刚度叠加理论设计,由四根固定导向梁和两根平面弹簧组成。首先,利用链式梁约束模型和广义打靶法分别求解出固定导向梁和平面弹簧的位移-载荷特性,建立优化模型,得到了满足设计要求的最优解。然后,为验证计算模型的准确性,使用ANSYS对柔顺梁进行大变形分析,结果显示该计算模型和有限元分析相比,误差小于1%。最后,通过实验发现该柔顺机构输出力的实验值和理论设计值存在误差,经过分析发现是加工精度、材料弹性模量误差以及只考虑了中性轴的变形等原因所导致的。

基于结构方程模型的广州市相对贫困空间分异及动力机制研究

我国农村脱贫攻坚取得全面胜利,相对贫困问题将是未来贫困研究的关键。文章以广州市为例,通过多源综合数据分析相对贫困分布规律和空间聚集特征,并结合结构方程模型探讨相对贫困空间分异的动力机制。研究表明:(1)广州市相对贫困呈现“高—低—高”的圈层分布特征;(2)根据结构方程模型路径分析结果,基础设施、人力资本、经济活力、自然环境等因素对广州市相对贫困空间分异形成了多重路径的影响关系;(3)结合地理区位和城乡差异划分内城衰落型、城乡融合型和乡村滞后型3种贫困模式,并针对不同贫困空间类型提出差异化的贫困治理建议,以期为后扶贫时代城乡相对贫困的精细化治理提供理论基础。

基于扎根理论的政府数据开放平台用户体验研究

[目的/意义]探明用户在使用政府数据开放平台过程中的体验状况及生成机制,对于完善政府数据开放平台,促进用户更好利用数据,充分发挥数据价值具有重要意义。[方法/过程]使用扎根理论及文本分析工具,对187个城市政府数据开放平台的用户评论进行分析。[结果/结论]政府数据开放平台用户体验的生成过程受用户与平台两方面因素的共同影响,各种类型的用户会基于自身的多种需求使用政府数据开放平台,而平台的数据、功能与界面的多重因素会对满足用户需求造成影响。同时,用户会对政府数据开放平台形成有用性与易用性的判断,在客观上实现基于用户互动的价值共创。