基于因果推断的冷轧带钢厚度异常根因分析

厚度精度是衡量冷轧带钢质量的重要指标之一，快速诊断带钢厚度异常并定位异常发生的根本原因，对提升带钢质量及生产稳定性具有非常重要的意义。为此，通过轧制机理模型确定厚度影响因素，采用多元线性回归方法构建厚度增量残差模型，对残差进行核密度估计来检测厚度异常；针对厚度异常，基于因果推断计算影响因素的因果效应，辨识出厚度异常的根本原因。实际应用结果表明，该方法能准确地诊断厚度异常和辨识异常的根本原因，与常规方法相比，其对冷连轧过程变量之间具有高度相关性的工况更为有效。     

基于广义归纳逻辑因果模型的知识发现与算法实现

以语言场、广义细胞自动机和广义归纳逻辑因果模型为理论依据,分析了广义因果联系类知识的发现机理,给出了因果联系类知识发现的实现算法.该算法为解决具有随机不确定和模糊不确定性特征的因果联系类知识的发现提供了行之有效的方法.通过算法的运行实例,验证和说明了算法的正确性和有效性.     

因果推断在钢铁企业能耗诊断应用中的探索

以钢铁企业焦化厂工序能耗的管理和诊断为例，运用因果推断并辅以神经网络、线性回归等数据分析方法探索影响焦化工序能耗上升或下降的原因。因果推断分析方法不但能够识别、跟踪影响工序能耗的关键指标，还能够对长尾指标与工序能耗的因果关系进行量化比较分析，通过数据手段结合人的主观经验的方法为企业能耗管理技术人员提供新的分析手段。     

基于商空间的黄金价格SVM模型预测

结合商空间理论和支持向量机方法,根据黄金价格的价格因子对我国黄金价格进行预测。采用Person相关系数法,对现阶段黄金价格的9个价格因子与黄金价格的相关性进行比较,筛选出相关系数较大的5个价格因子,并通过Granger因果检验,得出可能导致黄金价格变化的2个价格因子;将Person相关系数法和Granger因果检验选出的7个因子作为黄金价格预测的主要价格因子,结合商空间理论,按照时间属性,将黄金价格论域划分为年、季、月3个粒度,建立3层商空间,并进行粒度的合成和计算。然后建立基于商空间理论的支持向量机预测模型,预测得出年、季、月粒度的黄金价格预测值分别为8122.4,7947.506和8089.5元/金衡盎司,合成结果为8053.1元/金衡盎司。将预测结果与GM(1,1)预测值和实际黄金价格进行对比,证明该模型的预测结果在误差允许范围内,优于传统的价格预测方法。     

广义细胞自动机与广义归纳逻辑因果模型

在勃克斯(A.W.Burks)构筑的描述因果世界的细胞自动机的基础上(属归纳概率逻辑范畴),提出了可综合处理随机不确定性与模糊不确定性的广义细胞自动机与相应的广义归纳逻辑因果模型,解决了原模型中未曾解决的主因判定与因果扰动响应的问题,并找到了它在智能控制中的应用.     

基于强度理论的岩石损伤弹塑性模型

根据单轴受力特性曲线唯象地考察岩石材料损伤演化,定义弹性应变表示的一维损伤变量及其本构模型,利用双剪强度理论将其推广至三维模型.塑性是潜在破坏面的摩擦滑移,在传统塑性理论的框架中,建立了基于摩尔-库仑强度理论与潜在滑移面摩擦软-硬化特性的各向异性损伤弹塑性本构关系.结果表明,计算的损伤演化与CT观测结果符合很好,用本文的弹塑性模型反映损伤材料的力学特性是可行的.     

知识图谱的最新进展、关键技术和挑战

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.

     

小样本贫信息条件下高炉冶炼烧结终点组合预测法

由于烧结过程具有不确定性、多变量耦合、时变时滞的特点,并且烧结终点受多种因素的影响,采用传统控制方法难以将烧结终点控制在要求的范围内,提出应用支持向量机优良的时序预测性能,以及贝叶斯理论能够利用样本信息和先验知识来简化预测模型和优化参数的特性,建立了贝叶斯支持向量机烧结终点的预报模型.首先对烧结终点的机理分析,后分别叙述贝叶斯框架理论和LS-SVM算法,并将贝叶斯证据框架应用于最小二乘支持向量机模型参数的自动选择,建立起时间序列的烧结终点非线性预测模型.在贝叶斯推断的第一层,进行模型参数的选择;在贝叶斯推断的第二层,进行模型超参数的选择;在贝叶斯推断的第三层,估计模型核参数,然后利用贝叶斯最小二乘支持向量机算法(LS-SVM)对烧结终点进行预测,并在此基础上构造了烧结终点的贝叶斯最小二乘支持向量机模型.仿真结果和多种模型比较表明,本模型能在小样本贫信息条件下对烧结终点做出比较准确的预测,并具有预测精度高、所需样本少、计算简便等优点,取得了令人满意的结果.     

基于盲数理论的机械稳健性优化设计

在机械稳健性优化设计中不确定变量常被假定为服从某种特定分布的随机变量,在随后的优化模型中丢掉了这些不确定变量的初始数据,这种对不确定变量的处理方式往往不能真切地反映不确定变量的本质,使得出的结论偏离事实.为了更加真切地反映这些不确定变量,使用盲数表达机械设计中的不确定变量,运用盲数运算规则表达各不确定变量之间的关系,把盲数理论和稳健设计相结合,建立基于盲数理论的稳健设计优化模型.使用该优化模型对一个气动换向装置进行了优化计算,优化结果好于传统的稳健设计结果.基于盲数的方法是一种离散化的数值计算方法,算例充分展示了其灵活性,证明了该优化方法是合理的和实用的.     

基于VAR模型的氧化钕价格分析及预测

从氧化钕供求关系、宏观经济指标、联产品间的价格联动性三个维度对影响氧化钕价格的因素进行分析,并应用Pearson相关系数、共线性诊断、逐步回归、Johansen检验、Granger因果检验等方法筛选出对预测氧化钕价格有意义的解释变量,构建了VAR(1)模型,得到了2019年1月至2020年12月的氧化钕价格预测值。结果表明,未来两年氧化钕月价格呈缓慢上升的趋势,以每月0.18%的增长率增长。     

Dissociating Processes Supporting Causal Perception and Causal Inference in the Brain.

An understanding of relations between causes and effects is essential for making sense of the dynamic physical world. It has been argued that this understanding of causality depends on both perceptual and inferential components. To investigate whether causal perception and causal inference rely on common or on distinct processes, the authors tested 2 callosotomy (split-brain) patients and a group of neurologically intact participants. The authors show that the direct perception of causality and the ability to infer causality depend on different hemispheres of the divided brain. This finding implies that understanding causality is not a unitary process and that causal perception and causal inference can proceed independently. (PsycINFO Database Record (c) 2010 APA, all rights reserved)     

Judging probable cause.

Argues that people use systematic rules for assessing cause, both in science and everyday inference. By explicating the processes that underlie the judgment of causation, the authors review and integrate various theories of causality proposed by psychologists, philosophers, statisticians, and others. Because causal judgment involves inference and uncertainty, the literature on judgment under uncertainty is also considered. It is suggested that the idea of a "causal field" is central for determining causal relevance, differentiating causes from conditions, determining the salience of alternative explanations, and affecting molar versus molecular explanations. Various "cues-to-causality" such as covariation, temporal order, contiguity in time and space, and similarity of cause and effect are discussed, and it is shown how these cues can conflict with probabilistic ideas. A model for combining the cues and the causal field is outlined that explicates methodological issues such as spurious correlation, "causalation," and causal inference in case studies. The discounting of an explanation by specific alternatives is discussed as a special case of the sequential updating of beliefs. Conjunctive explanations in multiple causation are also considered. (120 ref) (PsycINFO Database Record (c) 2010 APA, all rights reserved)     

Causal judgments about relations between multilevel variables.

Three experiments presented stimulus information about cause and effect variables taking 3 quantitative values. Judgments tended to vary in accordance with considerations of conditions affecting the validity of causal inference from correlational data: whether causal candidates were presented simultaneously or in a temporal order such that one could affect the other and whether candidates were confounded with each other. The results supported a general hypothesis that causal judgments are moderated in accordance with acquired methodological intuitions. The 4th experiment showed that tendencies in correlation judgment were different from those in causal judgment, further supporting the hypothesis that causal judgment from multilevel variable information is, to some extent, determined by processes or conceptual frameworks specific to the domain of causal cognition. (PsycINFO Database Record (c) 2010 APA, all rights reserved)     

Ideas about causation in philosophy and psychology.

Philosophical theories summarized here include regularity and necessity theories from D. Hume (1739 [1978], 1740 [1978]) to the present; manipulability theory; the theory of powerful particulars; causation as connected changes within a defined state of affairs; departures from "normal" events or from some standard for comparison; causation as a transfer of something between objects; and causal propagation and production. Issues found in this literature and of relevance for psychology include whether actual causal relations can be perceived or known; what sorts of things people believe can be causes; different levels of causal analysis; the distinction between the causal relation itself and cues to causal relations; causal frames or fields; internal and external causes; and understanding of causation in different realms of the world, such as the natural and artificial realms. A full theory of causal inference by laypeople should address all of these issues. (PsycINFO Database Record (c) 2010 APA, all rights reserved)     

The Collider Principle in Causal Reasoning: Why the Monty Hall Dilemma Is So Hard.

The authors tested the thesis that people find the Monty Hall dilemma (MHD) hard because they fail to understand the implications of its causal structure, a collider structure in which 2 independent causal factors influence a single outcome. In 4 experiments, participants performed better in versions of the MHD involving competition, which emphasizes causality. This manipulation resulted in more correct responses to questions about the process in the MHD and a counterfactual that changed its causal structure. Correct responses to these questions were associated with solving the MHD regardless of condition. In addition, training on the collider principle transferred to a standard version of the MHD. The MHD taps a deeper question: When is knowing about one thing informative about another? (PsycINFO Database Record (c) 2010 APA, all rights reserved)     

Driven by power? Probe question and presentation format effects on causal judgment.

It has been proposed that causal power (defined as the probability with which a candidate cause would produce an effect in the absence of any other background causes) can be intuitively computed from cause-effect covariation information. Estimation of power is assumed to require a special type of counterfactual probe question, worded to remove potential sources of ambiguity. The present study analyzes the adequacy of such questions to evoke normative causal power estimation. The authors report that judgments to counterfactual probes do not conform to causal power and that they strongly depend on both the probe question wording and the way that covariation information is presented. The data are parsimoniously accounted for by an alternative model of causal judgment, the Evidence Integration rule. (PsycINFO Database Record (c) 2010 APA, all rights reserved)     

Learning a theory of causality.

The very early appearance of abstract knowledge is often taken as evidence for innateness. We explore the relative learning speeds of abstract and specific knowledge within a Bayesian framework and the role for innate structure. We focus on knowledge about causality, seen as a domain-general intuitive theory, and ask whether this knowledge can be learned from co-occurrence of events. We begin by phrasing the causal Bayes nets theory of causality and a range of alternatives in a logical language for relational theories. This allows us to explore simultaneous inductive learning of an abstract theory of causality and a causal model for each of several causal systems. We find that the correct theory of causality can be learned relatively quickly, often becoming available before specific causal theories have been learned—an effect we term the blessing of abstraction. We then explore the effect of providing a variety of auxiliary evidence and find that a collection of simple perceptual input analyzers can help to bootstrap abstract knowledge. Together, these results suggest that the most efficient route to causal knowledge may be to build in not an abstract notion of causality but a powerful inductive learning mechanism and a variety of perceptual supports. While these results are purely computational, they have implications for cognitive development, which we explore in the conclusion. (PsycINFO Database Record (c) 2010 APA, all rights reserved)     

Seeing it happen and knowing how it works: how children understand the relation between perceptual causality and underlying mechanism

Two experiments studied how 5- to 10-year-olds integrate perceptual causality with their knowledge of the underlying causal mechanism. Children learned about 2 devices by which a ball dropped into one end of a box made a bell ring at the other end, either immediately (contiguous mechanism) or after a delay (noncontiguous mechanism). When 1 ball was dropped first and a 2nd ball was dropped after a delay, and then the bell rang immediately, 5- and 7-year-olds chose the contiguous cause regardless of the mechanism inside. This was not due to lack of specific knowledge or problems with salient distractors. The results suggest a link between temporal contiguity and causality in children's understanding. Children also considered causal mechanism, in agreement with previous research, but they may not understand that mechanism is superordinate to perceptual cues for causality.