Causal temporal graph attention network for fault diagnosis of chemical processes

Fault detection and diagnosis(FDD)plays a significant role in ensuring the safety and stability of chemical processes.With the development of artificial intelligence(AI)and big data technologies,data-driven approaches with excellent performance are widely used for FDD in chemical processes.However,improved predictive accuracy has often been achieved through increased model complexity,which turns models into black-box methods and causes uncertainty regarding their decisions.In this study,a causal temporal graph attention network(CTGAN)is proposed for fault diagnosis of chemical processes.A chemical causal graph is built by causal inference to represent the propagation path of faults.The attention mechanism and chemical causal graph were combined to help us notice the key variables relating to fault fluctuations.Experiments in the Tennessee Eastman(TE)process and the green ammonia(GA)process showed that CTGAN achieved high performance and good explainability.

Abnormal behavior detection by causality analysis and sparse reconstruction

A new approach for abnormal behavior detection was proposed using causality analysis and sparse reconstruction. To effectively represent multiple-object behavior, low level visual features and causality features were adopted. The low level visual features, which included trajectory shape descriptor, speeded up robust features and histograms of optical flow, were used to describe properties of individual behavior, and causality features obtained by causality analysis were introduced to depict the interaction information among a set of objects. In order to cope with feature noisy and uncertainty, a method for multiple-object anomaly detection was presented via a sparse reconstruction. The abnormality of the testing sample was decided by the sparse reconstruction cost from an atomically learned dictionary. Experiment results show the effectiveness of the proposed method in comparison with other state-of-the-art methods on the public databases for abnormal behavior detection.

Efficiency of Stock Exchange Markets in G7 Countries： Bootstrap Causality Approach

Market efficiency is based on efficient market hypothesis （EMH）. EMH claims that market totally contains the available information. In case of EMH, valid investors who take position will not gain abnormal profits. If the efficiency can not be established, that is, if markets are not efficient, investors will have the opportunity of abnormal profits. This paper investigates the causality relations to determine validity of EMH among G7 （Canada, France, Germany, Italy, Japan, United Kingdom, and United States） countries＇ stock exchange markets for the period from July 2003 to October 2014. To find out whether the variables cause each other or not provides knowledge about the market efficiency. The implication of this analysis is twofold. One implication is that if the markets are informationally efficient, the possibility of abnormal returns through arbitrage is ruled out and investors can reduce the risk of their investment for the same expected returns, if they establish portfolios that consist of both markets rather than consisting of only one market. Based on this, Hacker-Hatemi-J. bootstrap causality test that is newer and has many advantages contrary to other tests was used. Results showed that EMH is valid among each G7 countries＇ stock exchange markets. Also portfolio diversification benefits exist among these markets.

A Causal Fusion Inference Method for Industrial Alarm Root Cause Analysis Based on Process Topology and Alarm Event Data

Modern industrial systems are usually in large scale,consisting of massive components and variables that form a complex system topology.Owing to the interconnections among devices,a fault may occur and propagate to exert widespread influences and lead to a variety of alarms.Obtaining the root causes of alarms is beneficial to the decision supports in making corrective alarm responses.Existing data-driven methods for alarm root cause analysis detect causal relations among alarms mainly based on historical alarm event data.To improve the accuracy,this paper proposes a causal fusion inference method for industrial alarm root cause analysis based on process topology and alarm events.A Granger causality inference method considering process topology is exploited to find out the causal relations among alarms.The topological nodes are used as the inputs of the model,and the alarm causal adjacency matrix between alarm variables is obtained by calculating the likelihood of the topological Hawkes process.The root cause is then obtained from the directed acyclic graph(DAG)among alarm variables.The effectiveness of the proposed method is verified by simulations based on both a numerical example and the Tennessee Eastman process(TEP)model.

Is Evolution a Causal, Yet Not-Predetermined Process?

It is demonstrated that “survival of the fittest” approach suffers fundamental flaw planted in its very goal: reaching a uniform state starting from a minor random event. Simple considerations prove that a generic property of any such state is its global instability. That is why a new approach to the evolution is put forward. It conjectures equilibrium for systems put in an ever-changing environment. The importance of this issue lies in the view that an ever-changing environment is much closer to the natural environment where the biological species live in. The major goal of the present paper is to demonstrate that a specific form of dynamical equilibrium among certain mutations is established in each and every stable in a long-run system. Major result of our considerations is that neither mutation nor either kind dominates forever because a temporary dynamical equilibrium is replaced with another one in the time course. It will be demonstrated that the evolution of those pieces of equilibrium is causal, yet not predetermined process.

Causal Circular Narrative and Time-Space Construction of the Movie Little Big Woman

Chinese film Little Big Woman takes the father’s funeral as the main line to tell the story of family affection and real-life emotional entanglements.In the narrative with multiple clues,we go back to find the hidden reasons,and the gradually clear reflection of the past is related to the real situation of the characters.Causal cycle narrative is not only a narrative strategy,but also a narrative logic with deep Chinese traditional cultural and philosophical connotations.

Studying History Education of High School in Taiwan by Causal Layered Analysis

After the 21st century,high school history learning will focus on teachers promoting the twelve-year state education.In recent years,in line with the changes in the new 108-year social curriculum,supporting strategies have been proposed:such as literacy orientation,inquiry and practice,learning process archives,and the structural direction of the controversial Chinese history into East Asian history.Historical learning has indeed had a great impact on the people’s national spiritual education and the development of historical consciousness in Taiwan’s education policy.This is the reason Taiwan’s Ministry of Education strives to improve students’historical literacy and connotation application abilities.When developing a learning policy,both external and internal learning factors need to be considered.The external aspect deals with the reasons for learning:Is learning for the purpose of using or accumulating historical wisdom in daily life to learn from the past and the present,on the other hand,to test the content of the course and the degree of absorption;or is it specifically for exams or other enlightenment purposes.The internal aspect involves those most affected by the policy:students and teachers.After studying and observing high school history learning policies for decades,some alternative future visions for history learning were found in the method of reflection on future research-the conclusion is that history is interestingly revitalized,and the preferred future is thematic history.According to the famous futurology scholar Sohail Inayatuallah’s proposal:the causal layering model.It helps understand how Taiwan’s historical policies operate.And how teachers and students on the front line respond to changes and take future actions.The key is to change the future:in the process of building an alternative future,whether the internal and external mix has changed or whether you want to try new things and expand your horizons.In fact,the difficulty of teaching lies in students’cooperation and conscious learning.Therefore,in the analysis of learning through alternative futures,is it possible to distinguish between internal and external situations and methods such as:1.Internal:Is education centered on teachers?Or is it student-centered?2.External:Does the Ministry of Education prioritize testing,or encourage teachers to adopt interactive communication and integrate education into the curriculum?Therefore,what is the function and inspiration of studying high school history and life?If thematic history teaching is used:teachers can use thematic learning methods to help students focus on causal relationships,the causes of turning points,or the evolution process of the beginning and end of events.This is more advantageous for testing based on the application topic,and it is easy to test how much understanding and understanding of history?Has an activating effect.By studying history in high school,using the“CLA(Causal layered analysis)”method of future studies,you can enter the stage of worldview exploration with the goal of improving professional depth and emotional level,and use it in your own understanding and utilization of history.Based on research,some insights into the prospects and thinking of learning history in high schools are provided:1.Facing the impact of declining birthrate,Taiwan needs a macro perspective to improve its future competitiveness and look forward to a new perspective on world history,using futuristic cause-and-effect level analysis to combine world changes with daily life applications.2.The study of history in high schools should go into a systematic construction:understand its cause-and-effect relationships and global trends,so teachers play a professional and future role in controlling the use of new information and technology.3.In the future,humans may develop more“intelligent”needs.As a reference from history or to explore the preferred path for the future,there will also be a greater need to innovate and meet challenges.4.Studying high school history has entered the professional field.Through self-exploration,it can be transformed into life affairs and establish the concept and value of lifelong learning.5.In studying the“history of high school learning”,have new prospects for the future of education.Through professional knowledge such as“trend theory and causal hierarchy analysis”of futurology,pursue new horizons and visions,making future education full of hope and possibility.

融合关系标签和位置信息的中文医疗文本因果关系抽取方法研究

目的/意义利用因果关系词相对位置辅助深度学习模型,提高因果关系预测能力,挖掘医疗文本增益信息。方法/过程将医疗文本因果关系词相对位置信息表示为关系特征层嵌入预训练语言模型,融合基线模型进行实体识别及关系抽取。结果/结论嵌入关系特征层的模型F1值较基线模型BERT-BiLSTM-CRF和CasRel分别提升2.92个百分点和6.41个百分点,因果关系预测能力较好。

基于改进动态因果图的湿法冶金浸出过程故障诊断方法研究

针对现有湿法冶金工艺机制复杂和浸出工艺故障诊断方法存在精度较差等问题,研究了用改进的动态因果图诊断湿法冶金浸出工艺故障,主要通过二元区间数对动态因果图进行优化改进,并对条件概率上下限进行加权,实现湿法冶金浸出工艺故障诊断。结果表明:该故障诊断方法相较于常规方法,故障诊断准确率更高,能对浸出工艺的故障进行较准确的诊断。

基于改进STPA-DEMATEL的智能航电系统致因要素分析

针对智能航电系统在非线性耦合运行场景下产生的预期功能安全(safety of the intended functionality,SOTIF)问题,提出一种将系统理论过程分析(systematic theory process analysis,STPA)与决策试验与评价实验法(decision-making trial and evaluation laboratory,DEMATEL)相结合的致因分析框架。首先,在定义系统级危险的基础上构建安全控制结构,识别其不安全控制行为并提取与智能化缺陷相关的STPA致因要素。接下来,引入毕达哥拉斯模糊加权平均算子和闵可夫斯基距离对传统DEMATEL方法进行优化,专家根据控制反馈回路对致因要素进行评价并计算其中心度与原因度。最后,分析STPA致因要素与SOTIF致因属性之间的映射关系,给出关键致因要素的风险减缓措施。以单一飞行员驾驶(single-pilot operation,SPO)模式下的虚拟驾驶员助理系统为例说明了所提方法的可行性与有效性。研究结果表明,改进的STPA-DEMATEL方法可以有效识别关键致因要素,且能够克服专家评价的模糊性与不确定性,为智能航电系统的安全性设计提供了参考依据。

基于自适应阈值学习的时序因果推断方法

时序数据存在近因性特点,即变量值普遍依赖近期的历史信息,而现有时序因果推断方法没有充分考虑时序数据的这种特性,在通过假设检验推断不同延迟的因果关系时使用统一的阈值,难以有效推断较弱的因果关系。针对上述问题,提出基于自适应阈值学习的时序因果推断方法:首先提取数据特性,其次根据不同延迟下数据呈现的性质,自动地学习假设检验过程中使用的阈值组合,最后将该阈值组合用于PC(Peter-Clark)算法、PCMCI(Peter-Clark and Momentary Conditional Independence)算法和VAR-LINGAM(Vector AutoRegressive LINear non-Gaussian Acyclic Model)算法的假设检验过程,以得到更准确的因果关系结构。在仿真数据集上的实验结果表明,采用所提方法的自适应PC算法、自适应PCMCI算法和自适应VAR-LINGAM算法的F1值都有所提高。

嵌入共识知识的因果图文检索方法

跨模态图像-文本检索是一项在给定一种模态(如文本)的查询条件下检索另一种模态(如图像)的任务.该任务的关键问题在于如何准确地测量图文两种模态之间的相似性,在减少视觉和语言这两种异构模态之间的视觉语义差异中起着至关重要的作用.传统的检索范式依靠深度学习提取图像和文本的特征表示,并将其映射到一个公共表示空间中进行匹配.然而,这种方法更多地依赖数据表面的相关关系,无法挖掘数据背后真实的因果关系,在高层语义信息的表示和可解释性方面面临着挑战.为此,在深度学习的基础上引入因果推断和嵌入共识知识,提出嵌入共识知识的因果图文检索方法.具体而言,将因果干预引入视觉特征提取模块,通过因果关系替换相关关系学习常识因果视觉特征,并与原始视觉特征进行连接得到最终的视觉特征表示.为解决本方法文本特征表示不足的问题,采用更强大的文本特征提取模型BERT(Bidirectional encoder representations from transformers,双向编码器表示),并且嵌入两种模态数据之间共享的共识知识对图文特征进行共识级的表示学习.在MS-COCO数据集以及MS-COCO到Flickr30k上的跨数据集实验,证明了本文方法可以在双向图文检索任务上实现召回率和平均召回率的一致性改进.

反事实分析+过程追踪:社会科学因果推断的整合进路

因果推断存在多种途径,然而近来兴起的因果推断讨论却明显倒向了“反事实分析”。为了帮助形成更加完整的因果推断认识,论文从因果本体论、认识论出发,结合因果推断的形式与类型,梳理出反事实分析与过程追踪两种因果推断途径,论证它们的因果推断比较优势及其互补性,指出反事实分析和过程追踪分别在识别因果关系和挖掘因果机制上占据比较优势,只有将二者结合,才能对因果形成较全面的理解。对此,文章以中国的治理场景为例,选取制度创新与制度效能转化、政策创新生成两个研究情境,论述了反事实分析和过程追踪的结合何以帮助理解中国之治。论文有助于完善因果推断三角研究,讲好中国故事,同时对于推进社会科学研究方法的纵深思考也具有参考价值。

智能手机组装线质量-可靠性耦合建模与评估方法

为研究多装配阶段因素因果关联致使产线良率传导与形成的机理,以实验室手机组装数字孪生产线为原型测试平台,围绕该平台锁螺丝、点胶两大工艺段良率问题,从因果关联关系出发,提出智能手机组装工艺过程产品质量-设备可靠性耦合建模方法,构建考虑多阶段因果要素关联的动态贝叶斯网络(dynamic Bayesian network,DBN)模型,并进行影响最终良率形成的溯因分析与重要度评价。所提方法的可用性和有效性在手机组装数字孪生产线上获得了测试和验证,所提方法可为产线良率损失阻隔机制、产线预防性维护决策提供性能评估支撑。

知识与句法融合的因果关系抽取网络

因果关系抽取作为关系抽取的一个重要任务,近年来得到了广泛关注。现有的因果关系抽取方法大多将句法结构和背景知识割裂开进行研究,早期的因果关系抽取方法偏重于从句法结构层面进行分析,随着深度学习技术的发展,预训练模型结合背景知识的方法成为主流。然而上述两种方法均未完全融合句内信息和外部知识,带来了不同程度的信息缺失。为了解决这一问题,提出了结合句法结构和背景知识的因果关系抽取模型。该模型将句子解析为同时包含句法和知识的知识句法图结构,使用图卷积网络进行信息融合。模型同时考虑了句法和知识两部分信息,从而进一步丰富了实体嵌入,达到了良好的因果关系抽取效果。本模型在EventStoryLine数据集上取得了良好效果,F1值达到0.445,与现有方法相比提高了2.3%。

基于分位数处理效应的面板数据政策评价方法

学者们在估计不同群体受到同一政策的影响时,通常需要使用协变量或者工具变量。为了在缺少协变量和工具变量的情况下估计一项政策对不同群体产生的影响,文章将分位数处理效应与面板数据政策评价方法相结合,得到一种新的政策影响效应估计方法——基于分位数处理效应的面板数据政策评价方法,并进一步研究了政策影响估计量的方差缩减和分解方法。新方法得到的政策影响估计量具有一致性和渐近正态性,交叉验证方法能够缩减估计量的方差,政策影响估计量可以分解为共有影响和特有影响。新的政策影响估计方法为学者们在缺少工具变量和协变量的情况下研究一项政策对不同群体产生的差异化影响提供了帮助。

An Integrated Approach to Oscillation Propagation Identification and Source Locating in Process Multi-loop Systems

An integrated method for identifying the propagation of multi-loop process oscillations and their source location is proposed in this paper. Oscillatory process loop variables are automatically selected based on the component-related ratio index and a mixing matrix, both of which are obtained in data preprocessing by spectral independent component analysis. The complex causality among oscillatory process variables is then revealed by Granger causality test and is visualized in the form of causality diagram. The simplification of causal connectivity in the diagram is performed according to the understanding of process knowledge and the final simplest causality diagram, which represents the main oscillation propagation paths, is achieved by the automated cutting-off thresh-old search, with which less significant causality pathways are filtered out. The source of the oscillation disturbance can be identified intuitively through the final causality diagram. Both simulated and real plant data tests are presented to demonstrate the effectiveness and feasibility of the proposed method.

中文医疗因果关系抽取数据集CMedCausal

介绍医学领域实体、关系抽取相关研究情况,详细阐述中文医疗因果关系抽取数据集CMedCausal构建方法及实验情况,提出利用数据集定义3类关键的医学因果推理关系:因果关系、条件关系和上下位关系,研究人员可基于CMedCausal开展医疗因果关系挖掘、医疗因果解释图谱构建等方向的研究。

A Basis for Causal Scattering Waves, Relativistic Diffraction in Time Functions

Relativistic diffraction in time wave functions can be used as a basis for causal scattering waves. We derive such exact wave function for a beam of Dirac and Klein-Gordon particles. The transient Dirac spinors are expressed in terms of integral defined functions which are the relativistic equivalent of the Fresnel integrals. When plotted versus time the exact relativistic densities show transient oscillations which resemble a diffraction pattern. The Dirac and Klein-Gordon time oscillations look different, hence relativistic diffraction in time depends strongly on the particle spin.

多主体协同视域下短视频网络舆情导控机理及因果机制研究

[目的/意义]短视频的日益普及与应用给网民的沟通交流带来便利的同时也使得短视频网络舆情甚嚣尘上,给网络舆情导控工作带来了新的挑战。[方法/过程]在对短视频网络舆情相关主体及其影响因素分析的基础上,运用Vensim PLE软件构建包含网民、短视频平台和政府3个子系统的系统动力学模型,并结合“郑州7·20特大暴雨灾害事件”进行仿真分析,进而采用过程追踪法对短视频网络舆情导控的因果机制进行分析,构建理论分析模型。[结果/结论]结果发现,短视频网络舆情受导控需求、主体参与和协同机制的共同影响,可分别从制定导控预案、强化融媒体矩阵建设、加强主体间的有效协同等方面实现短视频网络舆情的科学导控。