基于CLIP和交叉注意力的多模态情感分析模型

针对多模态情感分析中存在的标注数据量少、模态间融合不充分以及信息冗余等问题,提出了一种基于对比语言-图片训练(CLIP)和交叉注意力(CA)的多模态情感分析(MSA)模型CLIP-CA-MSA。首先,该模型使用CLIP预训练的BERT模型、PIFT模型来提取视频特征向量与文本特征;其次,使用交叉注意力机制将图像特征向量和文本特征向量进行交互,以加强不同模态之间的信息传递;最后,利用不确定性损失特征融合后计算输出最终的情感分类结果。实验结果表明:该模型比其他多模态模型准确率提高5百分点至14百分点,F1值提高3百分点至12百分点,验证了该模型的优越性,并使用消融实验验证该模型各模块的有效性。该模型能够有效地利用多模态数据的互补性和相关性,同时利用不确定性损失来提高模型的鲁棒性和泛化能力。

基于CLIP与注意力机制的跨模态哈希检索算法

针对传统无监督跨模态检索算法提取样本内部与样本之间的关联语义不充分,导致检索准确率低的问题,提出一种基于CLIP与注意力融合机制的无监督跨模态哈希检索算法CAFM_Net。将多模态预训练模型CLIP运用到样本特征提取阶段,从不同维度挖掘数据的相似信息;使用注意力融合机制对提取的特征进行处理,加强显著区域的权重;引入对抗学习的思想设计模态分类器,生成更趋于语义一致性的跨模态数据哈希编码。与现有的代表性哈希方法相比,CAFM_Net在多模态检索任务上准确率提升至少11%与9%。

基于CLIP模型和文本重建的人脸图像生成方法研究

针对文本生成人脸方法中生成图像与文本描述不一致、图像分辨率较低等问题,提出一种跨模态文本生成人脸图像网络框架。首先,采用CLIP预训练模型对文本进行特征提取,通过条件增强模块增强文本语义特征并生成隐藏向量;然后,将隐藏向量通过映射网络投影到预训练模型StyleGAN的隐式空间中获得解纠缠隐藏向量,将该向量输入到StyleGAN生成器中生成高分辨率人脸图像;最后,采用文本重建模块将人脸图像重新生成文本,计算重建文本和输入文本之间的语义对齐损失,并将其作为语义监督指导网络训练。在Multi-Modal CelebA-HQ和CelebAText-HQ两个数据集上进行训练与测试,实验结果表明,相比其他方法,该方法能生成更加符合文本描述的高分辨率人脸图像。

.NET平台集成CLIPS的实现方法研究

针对CLIPS语言框架及其不足,研究了在.NET中实现CLIPS托管程序集的方法,详细介绍了实现.NET平台CLIPS托管程序集的基本思想和关键技术。概述了CLIPS框架特点及其不足,分析了CLIPS与.NET混合编程的必要性。在保持CLIPS语言框架最大相似性的情况下,采用.NET的平台调用服务完成了对CLIPS框架类、接口的封装和调用。对生成的程序集进行引用,测试了其可行性。测试结果表明,该方法能够确保CLIPS框架的完成结构,实现对框架的完整二次开发和包装,继承了其推理结构的可靠性和系统的完整性,对专家系统的开发具有重要意义。

CLIP170通过TGF-β通路抑制甲状腺乳头状癌的转移

目的:研究细胞质连接蛋白170(cytoplasmic linker protein 170,CLIP170)是否影响甲状腺乳头状癌(papillary thyroid cancer,PTC)细胞的转移和侵袭并阐明其机制。方法:通过GEO和TCGA数据分析CLIP170在PTC中的表达水平;通过慢病毒转染技术构建CLIP170敲减细胞,Transwell转移和侵袭实验评估其功能;通过免疫荧光观察CLIP170对细胞肌动蛋白结构的影响;以ELISA方法检测细胞培养基中转化生长因子-β(transforming growth factor-β,TGF-β)1的释放;通过免疫印迹和实时定量荧光PCR方法检测上皮-间充质转化(epithelial-mesenchymal transition,EMT)和TGF-β信号通路分子的表达量并最终在裸鼠肺转移模型中验证。结果:CLIP170在PTC中的表达量比在正常甲状腺组织中的表达量低。功能方面,CLIP170KD在体外和体内均显著增强了PTC细胞的转移;机制方面,CLIP170KD触发了TGF-β通路的激活,促进肿瘤细胞的迁移和侵袭。TGF-β的抑制剂有效抑制了TGF-β活性,并显著逆转CLIP170KD所诱导的肿瘤转移。结论:CLIP170有望成为一种缓解有转移倾向的PTC的治疗靶点。

Novelty of Different Distance Approach for Multi-Criteria Decision-Making Challenges Using q-Rung Vague Sets

In this article,multiple attribute decision-making problems are solved using the vague normal set(VNS).It is possible to generalize the vague set(VS)and q-rung fuzzy set(FS)into the q-rung vague set(VS).A log q-rung normal vague weighted averaging(log q-rung NVWA),a log q-rung normal vague weighted geometric(log q-rung NVWG),a log generalized q-rung normal vague weighted averaging(log Gq-rung NVWA),and a log generalized q-rungnormal vagueweightedgeometric(logGq-rungNVWG)operator are discussed in this article.Adescription is provided of the scoring function,accuracy function and operational laws of the log q-rung VS.The algorithms underlying these functions are also described.A numerical example is provided to extend the Euclidean distance and the Humming distance.Additionally,idempotency,boundedness,commutativity,and monotonicity of the log q-rung VS are examined as they facilitate recognizing the optimal alternative more quickly and help clarify conceptualization.We chose five anemia patients with four types of symptoms including seizures,emotional shock or hysteria,brain cause,and high fever,who had either retrograde amnesia,anterograde amnesia,transient global amnesia,post-traumatic amnesia,or infantile amnesia.Natural numbers q are used to express the results of the models.To demonstrate the effectiveness and accuracy of the models we are investigating,we compare several existing models with those that have been developed.

基于CLIP和双空间自适应归一化的图像翻译

现有的图像翻译方法大多依赖数据集域标签来完成翻译任务,这种依赖往往限制了它们的应用范围。针对完全无监督图像翻译任务的方法能够解决域标签的限制问题,但是普遍存在源域信息丢失的现象。为了解决上述2个问题,提出一种基于对比学习语言-图像预训练(CLIP)的无监督图像翻译模型。首先,引入CLIP相似性损失对图像的风格特征施加约束,以在不使用数据集域标签的情况下增强模型传递图像风格信息的能力和准确性;其次,对自适应实例归一化(AdaIN)进行改进,设计一个新的双空间自适应归一化(DSAdaIN)模块,在特征的风格化阶段添加网络的学习和自适应交互过程,以加强对内容源域信息的保留;最后,设计一个鉴别器对比损失来平衡对抗网络损失的训练和优化过程。在多个公开数据集上的实验结果表明,与Star GANv2、Style DIS等模型相比,该模型可在准确传递图像风格信息的同时保留一定的源域信息,且在定量评估指标FID分数和KID分数上分别提升了近3.35和0.57×102,实现了较好的图像翻译性能。

Combined Optimal Dispatch of Thermal Power Generators and Energy Storage Considering Thermal Power Deep Peak Clipping and Wind Energy Emission Grading Punishment

Peak load and wind energy emission pressure rise more as wind energy penetration keeps growing,which affects the stabilization of the PS(power system).This paper suggests integrated optimal dispatching of thermal power generators and BESS(battery energy storage system)taking wind energy emission grading punishment and deep peak clipping into consideration.Firstly,in order to minimize wind abandonment,a hierarchical wind abandonment penalty strategy based on fuzzy control is designed and introduced,and the optimal grid-connected power of wind energy is determined as a result of minimizing the peak cutting cost of the system.Secondly,considering BESS and thermal power,the management approach of BESS-assisted virtual peak clipping of thermal power generators is aimed at reducing the degree of deep peak clipping of thermal power generators and optimizing the output of thermal power generators and the charging and discharging power of BESS.Finally,Give an example of how this strategy has been effective in reducing abandonment rates by 0.66% and 7.46% individually for different wind penetration programs,and the daily average can reduce the peak clipping power output of thermal power generators by 42.97 and 72.31 MWh and enhances the effect and economy of system peak clipping.

Legendre-Jacobi’s Elliptic Integrals Shed Light on the Luminosity Distance in Cosmology

This article concerns the integral related to the transverse comoving distance and, in turn, to the luminosity distance both in the standard non-flat and flat cosmology. The purpose is to determine a straightforward mathematical formulation for the luminosity distance as function of the transverse comoving distance for all cosmology cases with a non-zero cosmological constant by adopting a different mindset. The applied method deals with incomplete elliptical integrals of the first kind associated with the polynomial roots admitted in the comoving distance integral according to the scientific literature. The outcome shows that the luminosity distance can be obtained by the combination of an analytical solution followed by a numerical integration in order to account for the redshift. This solution is solely compared to the current Gaussian quadrature method used as basic recognized algorithm in standard cosmology.

Effect of long reaction distance on gas composition from organic-rich shale pyrolysis under high-temperature steam environment

When high-temperature steam is used as a medium to pyrolyze organic-rich shale,water steam not only acts as heat transfer but also participates in the chemical reaction of organic matter pyrolysis,thus affecting the generation law and release characteristics of gas products.In this study,based on a long-distance reaction system of organic-rich shale pyrolysis via steam injection,the effects of steam temperature and reaction distance on gas product composition are analyzed in depth and compared with other pyrolysis processes.The advantages of organic-rich shale pyrolysis via steam injection are then evaluated.The volume concentration of hydrogen in the gas product obtained via the steam injection pyrolysis of organic-rich shale is the highest,which is more than 60%.The hydrogen content increases as the reaction distance is extended;however,the rate of increase changes gradually.Increasing the reaction distance from 800 to 4000 mm increases the hydrogen content from 34.91%to 69.68%and from 63.13%to 78.61%when the steam temperature is 500℃ and 555℃,respectively.However,the higher the heat injection temperature,the smaller the reaction distance required to form a high concentration hydrogen pyrolysis environment(hydrogen concentration>60%).When the steam pyrolysis temperature is increased from 500℃ to 555℃,the reaction distance required to form a high concentration of hydrogen is reduced from 3800 to 800 mm.Compared with the direct retorting process,the volume concentration of hydrogen obtained from high-temperature steam pyrolysis of organic-rich shale is 8.82 and 10.72 times that of the commonly used Fushun and Kivite furnaces,respectively.The pyrolysis of organic-rich shale via steam injection is a pyrolysis process in a hydrogen-rich environment.

Responses of soil stoichiometry and soil enzyme activities in the different distance around opencast coal mine of the Hulun Buir Grassland of China

The objectives of this study were to explore the changes in soil stoichiometry and enzyme activities at different distances from an opencast coal mine in the Hulun Buir Grassland of China. Four transects were established on north and east sides of the opencast coal mining area, and samples were collected at 50 m, 550 m, and 1550 m from the pit on each transect. Control samples were collected from a grassland station 8 km from the opencast coal mining area that was not disturbed by mining. Four replicate soil samples were collected at each point on the four transects. Soil physicochemical properties and enzyme activities were determined, and correlations between soil properties and stoichiometric ratios and enzyme activities were explored using redundancy analysis. The increase in distance from mining did not significantly affect soil properties, although soil urease activity was significantly lower than that of the control area. Soil properties 1550 m from the mine pit were similar to those at the grassland control. In addition, soil total nitrogen had the greatest effect on soil stoichiometry, and soil total potassium had the greatest effect on soil enzyme activities. Coal dust from opencast mining might be the main factor affecting soil stoichiometry and enzyme activities. The results of this study provide direction for the next step in studying the influence of mining areas on soil properties and processes.

Robust free-space optical frequency transfer in time-varying link distances conditions

Future inter-satellite clock comparison on high orbit will require optical time and frequency transmission technology between moving objects.Here,we demonstrate robust optical frequency transmission under the condition of variable link distance.This variable link is accomplished by the relative motion of a single telescope fixed on the experimental platform to a corner-cube reflector(CCR)installed on a sliding guide.Two acousto–optic modulators with different frequencies are used to separate forward signal from backward signal.With active phase noise suppression,when the CCR moves back and forth at a constant velocity of 20 cm/s and an acceleration of 20 cm/s^(2),we achieve the best frequency stability of 1.9×10^(-16) at 1 s and 7.9×10^(-19) at 1000 s indoors.This work paves the way for future studying optical frequency transfer between ultra-high-orbit satellites.

Extended Lorenz majorization and frequencies of distances in an undirected network

Purpose:To contribute to the study of networks and graphs.Design/methodology/approach:We apply standard mathematical thinking.Findings:We show that the distance distribution in an undirected network Lorenz majorizes the one of a chain.As a consequence,the average and median distances in any such network are smaller than or equal to those of a chain.Research limitations:We restricted our investigations to undirected,unweighted networks.Practical implications:We are convinced that these results are useful in the study of small worlds and the so-called six degrees of separation property.Originality/value:To the best of our knowledge our research contains new network results,especially those related to frequencies of distances.

Source localization in signed networks with effective distance

While progress has been made in information source localization,it has overlooked the prevalent friend and adversarial relationships in social networks.This paper addresses this gap by focusing on source localization in signed network models.Leveraging the topological characteristics of signed networks and transforming the propagation probability into effective distance,we propose an optimization method for observer selection.Additionally,by using the reverse propagation algorithm we present a method for information source localization in signed networks.Extensive experimental results demonstrate that a higher proportion of positive edges within signed networks contributes to more favorable source localization,and the higher the ratio of propagation rates between positive and negative edges,the more accurate the source localization becomes.Interestingly,this aligns with our observation that,in reality,the number of friends tends to be greater than the number of adversaries,and the likelihood of information propagation among friends is often higher than among adversaries.In addition,the source located at the periphery of the network is not easy to identify.Furthermore,our proposed observer selection method based on effective distance achieves higher operational efficiency and exhibits higher accuracy in information source localization,compared with three strategies for observer selection based on the classical full-order neighbor coverage.

Numerical Analysis of the Influence of the Impinging Distance on the Scouring Efficiency of Submerged Jets

Water jet technology is widely used in submerged buried pipes as a non-traditional trenching process,often invol-ving a complex sediment response.An important adjustable and influential engineering variable in this technol-ogy is represented by the impinging distance.In this study,the FLOW-3D software was used to simulate the jet scouring of sand beds in a submerged environment.In particular,four sets of experimental conditions were con-sidered to discern the relationship between the maximum scour depth and mass and the impinging distance.As shown by the results,a critical impinging distance h0 exists by which the static scour depth can be maximized;the scour mass ratio between dynamic and static conditions decreases as the impinging distance increases.Moreover,the profile contours are similar when the erosion parameter Ec is in the range 0.35<Ec<2.Empirical equations applicable for predicting the jet trenching contour under both dynamic and static scour modes are also provided in this study.

Relationship between Inter Canine Width and Inter Alar Width, Inter Commissural Width, and Inter Canthal Distance: A Pilot Study

The aim of this study was to determine the relationship between inter canine width (ICW) and inter alar width (IAW), inter commissural width (ICoW), and inter canthal distance (ICaD) in Bantu population. Materials and Methods: It was a pilot study included all participants who were aged at least 18 years, without a history of orthodontics treatment from Prosthodontics Department, Faculty of Dental Medicine, University of Kinshasa. The Ethics committee of the National Center of Research for Dental Sciences in DR Congo approved the study protocol (CNRSB 1504.218). The participants medical records were obtained from the interview and dental examination. Nature of the study was explained with participant information sheet, and an informed consent of all participants was obtained. The inclusion criteria were no missing maxillary and mandibular teeth, no diastema, and no anterior restorations, 18 years old before enrollment in the study. The exclusion criteria were inability to meet the above requirements, pregnancy, and refusal to participate in the study. The inter-canine width (ICW), inter alar width (IAW), inter canthal distance (ICaD) and inter-commissural width (ICoW) of each participant were measured with a digital Vernier caliper (Mitutoyo, UK Ltd.,) (0.01 mm) three times and the average value was recorded. The inter canthal distance (ICaD) was measured without applying pressure by bringing the recording parts of the caliper just in contact with the medial angle. The inter alar width (IAW) was marked with a fine marked pencil on the widest outer surface of the alae of the nose on either side or width. Those two points were measured without applying pressure by bringing the recording parts of the caliper just in contact with the outer surface. The participant was told to stop breathing shortly to avoid any change in shape of the nose. The inter-commissural width (ICoW) was determined by measuring the maxillary lip vermilion from commissure to commissure without the application of pressure on the tissue in the relaxed state. The inter-canine width (ICW) was measured indirectly using a dental floss. A dental floss was marked on one end prior to placement in the mouth. Using that point as reference, the dental floss was circumference along the curvature of the anterior dentition such that it passed along the contact point of all the teeth. The distal end of the canine teeth on both sides was then marked on the floss while it was stretched in the patients mouth. Floss was marked on both sides with the marker pencil. The distance between the two proximal contact points was measured and recorded. Data, analysis and validation were performed by the SPSS software (version 22.0, IBM SPSS Statistics, Chicago, IL, USA). Unpaired t-test was used, of Kolmogorov-Smirnov test. Analysis of variance (ANOVA) test was used to test for comparability between socio-demographic characteristics dental measurements. Pearsons correlation coefficients test was calculated to determine the relationship between facial and dental parameters. Significance was set at P 0.01. Results: Of 314 participants enrolled, 202 were included. The age ranged from 18 to 68 years, with a mean age of 40.62 12.99 years. Although the Pearsons correlation coefficients were negative. Ninety-three participants (46%) were men and one hundred and nine (54%) were women. The overall mean age was 40.62 12.99 years. Facial and dental measurements were greater in women than men with significant differences for ICW (p = 0.04). However, no significant difference was seen between men and women for IAW, ICaD and ICoW (p = 0.44, p = 0.23, p = 0.31 respectively). The correlation including Pearsons correlation coefficient and P-values for all participants is not demonstrated. Conclusion: Within the limitations of this study, the results suggest that IAW, ICaD and ICoW cannot be used as a preliminary method for determining the width of the maxillary for anterior teeth for edentulous patients.

Hem-o-lok clip migration to duodenal bulb post-cholecystectomy:A case report

BACKGROUND Hem-o-lok clips are typically used to control the cystic duct and vessels during laparoscopic cholecystectomy(LC)and common bile duct exploration for stones in the bile duct and gallbladder.Here,we report a unique example of Hem-o-lok clip movement towards the duodenal bulb after LC,appearing as a submucosal tumor(SMT).Additionally,we provide initial evidence of gradual and evolving endoscopic manifestations of Hem-o-lok clip migration to the duodenal bulb wall and review the available literature.CASE SUMMARY A 72-year-old man underwent LC for gallstones,and Hem-o-lok clips were used to ligate both the cystic duct and cystic artery.Esophagogastroduodenoscopy(EGD)2 years later revealed an SMT-like lesion in the duodenal bulb.Due to the symptomatology,the clinical examination did not reveal any major abnormalities,and the patient was followed up as an outpatient.A repeat EGD performed 5 months later revealed an SMT-like lesion in the duodenal bulb with raised edges and a central depression.A third EGD was conducted,during which a Hem-o-lok clip was discovered connected to the front side of the duodenum.The clip was extracted easily using biopsy forceps,and no complications occurred.Two months after the fourth EGD,the scar was surrounded by normal mucosa.CONCLUSION Clinicians should be aware of potential post-LC complications.Hem-o-lok clips should be removed if symptomatic.

Interrelationships between Length of the Day, Moon Distance, Phanerozoic Geodynamic Cycles, Tidal Dissipation and Earth’s Core: Review and Analysis

The rotation of the Earth and the related length of the day (LOD) are predominantly affected by tidal dissipation through the Moon and the growth of the Earth’s core. Due to the increased concentration of mass around the rotation axis of the spinning Earth during the growth of the core the rotation should have been accelerated. Controversially the tidal dissipation by the Moon, which is mainly dependent on the availability of open shallow seas and the kind of Moon escape from a nearby position, acts towards a deceleration of the rotating Earth. Measurements of LOD for Phanerozoic and Precambrian times open ways to solve questions concerning the geodynamical history of the Earth. These measurements encompass investigations of growth patterns in fossils and depositional patterns in sediments (Cyclostratigraphy, Tidalites, Stromatolites, Rhythmites). These patterns contain information on the LOD and on the changing distance between Earth and Moon and can be used as well for a discussion about the growth of the Earth’s core. By updating an older paper with its simple approach as well as incorporating newly published results provided by the geoscientific community, a moderate to fast growth of the core in a hot early Earth will be favored controversially to the assumption of a delayed development of the core in an originally cold Earth. Core development with acceleration of Earth’s rotation and the contemporaneous slowing down due to tidal dissipation during the filling of the ocean may significantly interrelate.

Reciprocal Distance Laplacian Eigenvalue Distribution Based on Graph Parameters

Let G be a connected graph of order n and m_(RD)^(L)_(G)I denote the number of reciprocal distance Laplacian eigenvaluesof G in an interval I.For a given interval I,we mainly present several bounds on m_(RD)^(L)_(G)I in terms of various structuralparameters of the graph G,including vertex-connectivity,independence number and pendant vertices.

Leading directions and effective distance of larch offspring dispersal at the upper treeline in the Northern and Polar Urals, Russia

Climate has changed sufficiently over the last 150 years and forced out upper treeline advance at the most studied sites around the world.The rate of advance has been extremely variable–from tens to hundreds meters in altitude.This is because the degree at which tree frontal populations respond to climate change depends on the complex interaction of biological and physical factors.The resulting stand pattern is the consequence of the interaction between dispersal and survival functions.A few publications have addressed the question of how this pattern is generated.In order to understand how the spatial structure of tree stands was formed at the upper limit of their distribution in the Ural Mountains,we assessed the distance and direction of dispersal of offspring from maternal individuals.We found that in frontal Larix sibirica Ledeb.populations,‘effective’dispersal of offspring ranges from 3 to 758 m(with a median of 20–33 m in open forest and 219 m in single-tree tundra in the Polar Urals and 107 m in open forest in the Northern Urals).We revealed that most of the offspring effectively dispersed not only in the direction of the prevailing winds,but also in the opposite direction up the slope,and the distance can reach 500–760 m.The data obtained can be used to develop an individual-based model which is capable of simulating in detail the dynamics of tree stands at the upper limit of their growth and reliably predicting the future position and pattern of treeline ecotone as growth conditions continue to improve in the face of observed climate change.