The Next Breakthroughs of Artificial Intelligence:The Interdisciplinary Nature of AI

1.Introduction Artificial intelligence(AI)is committed to the realization of machine-borne intelligence.The technologies underpinning AI have made huge leaps in the past decade,bringing exciting applications such as language understanding,vision recognition,and intelligent digital assistants.However,contemporary AI systems are good at specific predefined tasks and are unable to learn by themselves from data or from experience,intuitive reasoning,and adaptation.

Analysis of fracture propagation and shale gas production by intensive volume fracturing

This paper presents an integrated study from fracture propagation modeling to gas flow modeling and a correlation analysis to explore the key controlling factors of intensive volume fracturing.The fracture propagation model takes into account the interaction between hydraulic fracture and natural fracture by means of the displacement discontinuity method(DDM)and the Picard iterative method.The shale gas flow considers multiple transport mechanisms,and the flow in the fracture network is handled by the embedded discrete fracture model(EDFM).A series of numerical simulations are conducted to analyze the effects of the cluster number,stage spacing,stress difference coefficient,and natural fracture distribution on the stimulated fracture area,fractal dimension,and cumulative gas production,and their correlation coefficients are obtained.The results show that the most influential factors to the stimulated fracture area are the stress difference ratio,stage spacing,and natural fracture density,while those to the cumulative gas production are the stress difference ratio,natural fracture density,and cluster number.This indicates that the stress condition dominates the gas production,and employing intensive volume fracturing(by properly increasing the cluster number)is beneficial for improving the final cumulative gas production.

Noncooperative Model Predictive Game With Markov Jump Graph

In this paper,the distributed stochastic model predictive control(MPC)is proposed for the noncooperative game problem of the discrete-time multi-player systems(MPSs)with the undirected Markov jump graph.To reflect the reality,the state and input constraints have been considered along with the external disturbances.An iterative algorithm is designed such that model predictive noncooperative game could converge to the socalledε-Nash equilibrium in a distributed manner.Sufficient conditions are established to guarantee the convergence of the proposed algorithm.In addition,a set of easy-to-check conditions are provided to ensure the mean-square uniform bounded stability of the underlying MPSs.Finally,a numerical example on a group of spacecrafts is studied to verify the effectiveness of the proposed method.

Numerical simulation of seismic wave in elastic and viscoelastic TTI media

Based on the two-dimensional(2D)three-component first-order velocity-stress equation,the high order staggered mesh finite difference numerical simulation method was used to simulate the elastic and viscoelastic tilted transversely isotropic(TTI)media.The perfect matched layer(PML)absorption boundary condition was selected to eliminate the boundary effect.The results show that:(①)Under the condition of fixed elastic parameters of elastic TTI medium,when the polarization angle and azimuth are 60°and 45°respectively,the degree of shear wave splitting is significantly greater than the angle of 0°;②The influence of viscoelasticity on TTI medium is mainly reflected in the amplitude.If the quality factor decreases,the attenuation of the seismic wave amplitude increases,causing the waveform to become wider and distorted.If the quality factor increases,the viscoelastic medium becomes closer to elastic medium;③For TTI medium with different polarization angle and azimuth angle in the upper and lower layers,the shear wave can multiple splits at the interface of medium.The symmetry of seismograms is affected by the polarization angle and azimuth angle of TTI medium;④Viscoelasticity has a great influence on reflected wave,transmitted wave and converted wave in the low-velocity model.When the viscoelasticity is strong,the weaker waves may not be shown.

图像信息量度量

图像不仅承载着丰富的视觉内容,同时还蕴含潜在的高级语义,是重要的信息传递媒介.度量图像信息量则是对图像信息进行定性描述和定量计算的抽象过程,属于对图像处理领域中的科学问题进行理论构建的关键环节.然而,在现有的研究工作中,虽然对“图像信息量”的度量有所提及,但实际上关注的往往是在图像质量或图像特征的层级,而忽略了图像所具有的高级语义以及人对图像的认知.因此,为了衡量图像的信息提供价值,本文基于“图像信息量因人、因任务而异”这一常识,通过建模“先验知识”相关的图像记忆并提取“给定任务”相关的图像特征,尝试对图像信息量进行度量.首先,提出图像信息量度量的理论框架,具体表现为:在“能够获得全世界所有的图像”和“能找到一种能够准确表达图像语义的特征”两个假设条件下,以信息论为基础,通过构建超球来描述样本点的邻域稠密度,由此建模语义信息的概率,进而得到理论假设条件下的图像信息量.其次,探讨了理论假设无法满足的实际情况,将“全世界所有的图像”这一假设条件退化为“给定具体的数据集”,并将“一种能够准确表达图像语义的特征”这一假设退化为“给定任务相关的特征”,进一步通过实验展示了在实际情况下如何有效获得图像信息量的数值结果.最后,指出了本文度量方法的限定对象和度量结果的边界,并展望了未来对相关体系进行丰富和完善的可行方向.

Robust polarimetric dehazing algorithm based on low-rank approximation and multiple virtual-exposure fusion

Polarimetric dehazing is an effective way to enhance the quality of images captured in foggy weather.However,images of essential polarization parameters are vulnerable to noise,and the brightness of dehazed images is usually unstable due to different environmental illuminations.These two weaknesses reveal that current polarimetric dehazing algorithms are not robust enough to deal with different scenarios.This paper proposes a novel,to our knowledge,and robust polarimetric dehazing algorithm to enhance the quality of hazy images,where a low-rank approximation method is used to obtain low-noise polarization parameter images.Besides,in order to improve the brightness stability of the dehazed image and thus keep the image have more details within the standard dynamic range,this study proposes a multiple virtual-exposure fusion(MVEF)scheme to process the dehazed image(usually having a high dynamic range)obtained through polarimetric dehazing.Comparative experiments show that the proposed dehazing algorithm is robust and effective,which can significantly improve overall quality of hazy images captured under different environments.

视频萃取

视频数据是人们日常生活中最重要的信息载体之一.视频萃取(video distillation)通过研究视频数据的时空和语义特性,探索简洁高效的数据展示形式和信息感知模态,是计算机视觉和人工智能的重点研究内容.近年来,随着视频获取方式的快速革新和拍摄需求的多样化发展,视频数据的智能化分析任务面临着新的机遇与挑战,涌现出众多的视频萃取方法.本文创新性地从信息论的角度,解释了数据、信息和知识之间的关系,确立了视频萃取的核心是提高单位数据量的信息提供能力这一基本原则,并依据数据信容(information capacity)分析,从理论上对视频萃取中的各项任务进行了统一.进一步地,分类讨论了视频时空表征中的关键问题与解决方案,系统地分析了从内容、目标和语义角度进行视频萃取的方法,结合视频摘要、浓缩和描述任务,梳理出三条发展主线,展现了视频萃取的发展态势.更重要的是,本文对现有方法的优势与缺陷进行了深入的思考与讨论,指出了尚未解决的若干关键科学问题,并对解决方案进行了初步探讨.同时,本文对视频萃取研究所面临的挑战与未来发展趋势进行了系统的分析与展望.

光学影像序列中基于多视角聚类的群组行为分析

群组行为分析是光学影像序列分析中的一项重要课题,在近年来引起了人工智能领域研究人员的广泛关注.与行人个体相比,群组能提供更高层的语义表示,为分析人群运动模式提供基础.本文将人群影像序列中的影像块作为研究对象,提出了一种基于多视角聚类的群组行为分析方法,对运动模式不同的群组进行区分,主要研究内容有:(1)提出了基于特征点的影像块构图方法,从交互关系、空间位置、运动方向分布,以及运动规律等方面衡量影像块之间的关系;(2)提出了一种多视角聚类方法,通过融合多种特征对每个影像块分配类标签,并引入图多样性正则项以避免特征冗余;(3)提出了一种类合并方法,根据类内特征点的运动方向和类中心位置坐标,对关联度较高的类别进行合并,自动确定最终群组数目.CUHK人群数据集上的实验结果证明了该方法能够准确划分出影像数据中的群组.另外,与现有方法相比,本文提出的多视角聚类方法也在不同数据集上取得了较好的实验结果.

一种基于荧光qPCR检测新型冠状病毒核酸的优化反应体系的建立及相关测试

为实时并快速地检出SARS-CoV-2冠状病毒RNA,对基于荧光定量聚合酶链式反应(Polymerasechain reaction,PCR)的反应体系进行了优化。结果表明,按照本实验提供的方法操作后,检测SARS-CoV-2冠状病毒所用的RNA样本的最小浓度稀释度可调至1/10000(初始值设为10ng/μL)。且用于检测COVID-19临床阳性样本所测得循环值(Cycle threshold,Ct)均低于35或40。其灵敏性测试结果也表明该方法的敏感性较好。同时在同等条件下,与目前市场上的COVID-19试剂盒的检测结果基本一致,并且检测循环数缩短2个单位。因此,本实验所建立的体系适用于前期临床诊断的筛查工作,为在医学上实现快速诊断提供了工具。

Collective Representation for Abnormal Event Detection

Abnormal event detection in crowded scenes is a hot topic in computer vision and information retrieval community. In this paper, we study the problems of detecting anomalous behaviors within the video, and propose a robust collective representation with multi-feature descriptors for abnormal event detection. The proposed method represents different features in an identical representation, in which different features of the same topic will show more common properties. Then, we build the intrinsic relation between different feature descriptors and capture concept drift in the video sequence, which can robustly discriminate between abnormal events and normal events. Experimental results on two benchmark datasets and the comparison with the state-of-the-art methods validate the effectiveness of our method.

Super-resolution reconstruction of a 3 arc-second global DEM dataset

Obtaining multi-source satellite geodesy and ocean observation data to map the high-resolution global digital elevation models(DEMs)is a formidable task at present.The key theoretical and technological obstacles to fine modeling must be overcome before the geographical ditribution of ocean topography and plate move-ment patterns can be explored.Geodesy.

像素级语义理解:从分类到回归

近年来,随着科学技术的快速发展和硬件设备的不断迭代,人工智能在各种领域(如安防监控、医疗辅助、健康诊断、智能推荐、遥感监测、目标定位等)都得到了广泛的应用.随着人们对智能处理任务的要求逐步提高,人工智能算法所需要理解的语义信息层次和输出数据精准度要求也步步攀升.因此,像素级语义理解任务也因其精准度要求远高于图像级理解而越来越受到重视.与图像级理解相比,像素级语义理解具有输出数据量大、逐像素输出精度高的优点,相应地其难度也更大,内部成因更值得关注与研究.为此本文从信息度量的角度出发,结合像素级语义理解任务的特有属性,给出了像素级语义理解任务的定义与优化目标,进一步依据实际任务的特性从初始定义衍生出像素级语义分类和像素级语义回归两类任务;随后分别讨论了在这两类任务中优化目标的退化和演变,并通过详尽的调研总结了常见像素级语义理解任务的发展现状;紧接着探究了当前像素级语义理解的难点和未来发展方向,针对亟待解决的问题给出了深入的分析思考以及可行的解决方案;最后重点反思了后深度学习时代像素级语义理解乃至人工智能领域所面对的机遇与挑战,提出知识的方向指导和数据的优化驱动是未来人工智能发展的重点关注目标.本文意图从像素级语义理解的定义与发展现状出发,延伸出对当前工作的思考以及对整个领域的反思,强调整个领域面临的风险;在介绍像素级语义理解基础认知的同时对相关技术的发展方向和路径进行深入的思考与深度的展望.