IndRT-GCNets: Knowledge Reasoning with Independent Recurrent Temporal Graph Convolutional Representations

Due to the structural dependencies among concurrent events in the knowledge graph and the substantial amount of sequential correlation information carried by temporally adjacent events,we propose an Independent Recurrent Temporal Graph Convolution Networks(IndRT-GCNets)framework to efficiently and accurately capture event attribute information.The framework models the knowledge graph sequences to learn the evolutionary represen-tations of entities and relations within each period.Firstly,by utilizing the temporal graph convolution module in the evolutionary representation unit,the framework captures the structural dependency relationships within the knowledge graph in each period.Meanwhile,to achieve better event representation and establish effective correlations,an independent recurrent neural network is employed to implement auto-regressive modeling.Furthermore,static attributes of entities in the entity-relation events are constrained andmerged using a static graph constraint to obtain optimal entity representations.Finally,the evolution of entity and relation representations is utilized to predict events in the next subsequent step.On multiple real-world datasets such as Freebase13(FB13),Freebase 15k(FB15K),WordNet11(WN11),WordNet18(WN18),FB15K-237,WN18RR,YAGO3-10,and Nell-995,the results of multiple evaluation indicators show that our proposed IndRT-GCNets framework outperforms most existing models on knowledge reasoning tasks,which validates the effectiveness and robustness.

Local-to-Global Causal Reasoning for Cross-Document Relation Extraction

Cross-document relation extraction(RE),as an extension of information extraction,requires integrating information from multiple documents retrieved from open domains with a large number of irrelevant or confusing noisy texts.Previous studies focus on the attention mechanism to construct the connection between different text features through semantic similarity.However,similarity-based methods cannot distinguish valid information from highly similar retrieved documents well.How to design an effective algorithm to implement aggregated reasoning in confusing information with similar features still remains an open issue.To address this problem,we design a novel local-toglobal causal reasoning(LGCR)network for cross-document RE,which enables efficient distinguishing,filtering and global reasoning on complex information from a causal perspective.Specifically,we propose a local causal estimation algorithm to estimate the causal effect,which is the first trial to use the causal reasoning independent of feature similarity to distinguish between confusing and valid information in cross-document RE.Furthermore,based on the causal effect,we propose a causality guided global reasoning algorithm to filter the confusing information and achieve global reasoning.Experimental results under the closed and the open settings of the large-scale dataset Cod RED demonstrate our LGCR network significantly outperforms the state-ofthe-art methods and validate the effectiveness of causal reasoning in confusing information processing.

Text Reasoning Chain Extraction for Multi-Hop Question Answering

With the advent of the information age, it will be more troublesome to search for a lot of relevant knowledge to find the information you need. Text reasoning is a very basic and important part of multi-hop question and answer tasks. This paper aims to study the integrity, uniformity, and speed of computational intelligence inference data capabilities. That is why multi-hop reasoning came into being, but it is still in its infancy, that is, it is far from enough to conduct multi-hop question and answer questions, such as search breadth, process complexity, response speed, comprehensiveness of information, etc. This paper makes a text comparison between traditional information retrieval and computational intelligence through corpus relevancy and other computing methods. The study finds that in the face of multi-hop question and answer reasoning, the reasoning data that traditional retrieval methods lagged behind in intelligence are about 35% worse. It shows that computational intelligence would be more complete, unified, and faster than traditional retrieval methods. This paper also introduces the relevant points of text reasoning and describes the process of the multi-hop question answering system, as well as the subsequent discussions and expectations.

Case-Based Reasoning(CBR) Model for Ultra-Fast Cooling in Plate Mill

New generation thermo-mechanical control process（TMCP） based on ultra-fast cooling is being widely adopted in plate mill to product high-performance steel material at low cost. Ultra-fast cooling system is complex because of optimizing the temperature control error generated by heat transfer mathematical model and process parameters. In order to simplify the system and improve the temperature control precision in ultra-fast cooling process, several existing models of case-based reasoning（CBR） model are reviewed. Combining with ultra-fast cooling process, a developed R5 CBR model is proposed, which mainly improves the case representation, similarity relation and retrieval module. Certainty factor is defined in semantics memory unit of plate case which provides not only internal data reliability but also product performance reliability. Similarity relation is improved by defined power index similarity membership function. Retrieval process is simplified and retrieval efficiency is improved apparently by windmill retrieval algorithm. The proposed CBR model is used for predicting the case of cooling strategy and its capability is superior to traditional process model. In order to perform comprehensive investigations on ultra-fast cooling process, different steel plates are considered for the experiment. The validation experiment and industrial production of proposed CBR model are carried out, which demonstrated that finish cooling temperature（FCT） error is controlled within±25℃ and quality rate of product is more than 97%. The proposed CBR model can simplify ultra-fast cooling system and give quality performance for steel product.

CASE STORAGE BASED ON RELATIONAL DATABASE

This paper focused on the integration of case base and relational database management system (RDBMS). The organizational and commercial impact will be far greater if the case based reasoning (CBR) system is integrated with main stream information system, which is exemplified by RDBMS. The scalability, security and robustness provided by a commercial RDBMS facilitate the CBR system to manage the case base. The virtual table in relational database (RDB) is important for CBR systems to implement the flexibility of case template. It was discussed how to implement a flexible and succinct case template, and a mapping model between case template and RDB was proposed. The key idea is to build the case as the virtual view of underlying data.

A Dialogue System for Coherent Reasoning with Inconsistent Knowledge Bases

Traditionally, the AI community assumes that a knowledge base must be consistent. Despite that, there are many applications where, due to the existence of rules with exceptions, inconsistent knowledge must be considered. One way of restoring consistency is to withdraw conflicting rules;however, this will destroy part of the knowledge. Indeed, a better alternative would be to give precedence to exceptions. This paper proposes a dialogue system for coherent reasoning with inconsistent knowledge, which resolves conflicts by using precedence relations of three kinds: explicit precedence relation, which is synthesized from precedence rules;implicit precedence relation, which is synthesized from defeasible rules;mixed precedence relation, which is synthesized by combining explicit and implicit precedence relations.

Build Evidence Chain Relational Model Based on Chinese Judgment Document

The reasoning of judgment documents is the touchstone of justice. Attaching importance to the reasoning of judgment documents is essentially the embodiment of judiciary civilization. In order to promote the reform of judgment documents reasoning and improve the level of it, the technology of automated judgment documents reasoning evaluation has to be studied on. How to build evidence chain relational model is the basis and key to this technology.An approach is proposed to build evidence chain relational model based on Chinese judgment documents. Using automated text preprocessing for Chinese judgment documents creates semi-structured XML documents and extracts evidence set and fact set. The method of key elements extraction is used to obtain the keywords of evidence and facts. Calculating the degree of association can work out the connection points of evidence chain relational model. Tabular display and graphical display of evidence chain relational model can be realized.

基于跨模态引导和对齐的多模态预训练方法

现有的视觉语言多模态预训练方法仅在图像和文本的全局语义上进行特征对齐,对模态间细粒度特征交互的探索不足.针对这一问题,本文提出了一种基于跨模态引导和对齐的多模态预训练方法.该方法在模态特征提取阶段,采用基于视觉序列压缩的双流特征提取网络,在视觉编码器中联合图像和文本信息逐层引导视觉序列压缩,缓解与文本无关的冗余视觉信息对模态间细粒度交互的干扰;在模态特征对齐阶段,对图像和文本特征进行细粒度关系推理,实现视觉标记与文本标记的局部特征对齐,增强对模态间细粒度对齐关系的理解.实验结果表明,本文方法能够更好地对齐视觉文本的细粒度特征,在图文检索任务中,微调后的图像检索和文本检索的平均召回率分别达到了86.4%和94.88%,且零样本图文检索的整体指标相较于经典图文检索算法CLIP(Contrastive Language-Image Pre-training)提升了5.36%,在视觉问答等分类任务中,准确率也优于目前主流多模态预训练方法.

民法典中意思表示解释规则视野下的重大误解——以意思表示错误为基础

对重大误解的认定是一项意思表示解释的工作。民法典第142条区分有相对人的意思表示和无相对人的意思表示,对前者采表示主义的解释标准,对后者采意思主义的解释标准。解释标准,是对错误应否救济、应否认定为重大误解的标准。通过提出动机他项事实错误和动机同项事实错误两个创新概念,解决立法例尚未解决的动机进入效果意思的问题。不能对抗善意相对人的错误、非重大错误以及可以通过意思表示解释修正的错误,不构成重大误解。

融合实体语义及结构信息的知识图谱推理

目前,图注意力网络(GAT)通过引入注意力机制对目标实体的邻域实体赋予不同权重并进行信息聚合,使得它更关注实体的局部邻域,忽略了图结构中实体和关系之间的拓扑结构;而且在多头注意力后将输出嵌入向量简单拼接或平均,导致注意力头之间相互独立,未能捕捉不同注意力头的重要语义信息。针对GAT应用于知识图谱(KG)推理任务时未充分挖掘实体结构信息和语义信息的问题,提出融合实体语义及结构信息的知识图谱推理(FESSI)模型。首先,使用TransE将实体和关系表示为同一空间的嵌入向量。其次,提出交互注意力机制,将GAT中多头注意力重新融合成多个混合注意力,增强注意力头之间的交互性,以提取目标实体更丰富的语义信息;同时,利用关系图卷积网络(R-GCN)提取实体的结构信息,并通过权重矩阵学习GAT和R-GCN的输出特征向量。最后,使用ConvKB作为解码器进行评分。在知识图谱数据集Kinship、NELL-995和FB15K-237上的实验结果表明,FESSI模型的效果优于多数对比模型,在3个数据集的平均倒数排名(MRR)指标上的结果分别为0.964、0.565和0.562。

多尺度子群体交互关系下的群体行为识别方法

群体行为识别旨在识别包含多个个体的群体行为。在现实场景中,群体行为通常可以被视为从群体到子群体再到个人的层次结构。然而,以前的群体行为识别方法主要侧重于对个体之间的关系进行建模,缺乏对子群体之间关系的深度研究。从该角度出发,该文提出一种基于多尺度子群体交互关系(MSIR)的多层次群体行为识别框架。除对个体关系进行建模外,重点关注了子群体之间的多尺度交互特征。具体优化如下:设计子群体划分模块,通过个体外观特征和其空间位置来聚合可能存在关联的个体,再进一步利用语义信息动态地生成不同尺度大小的子群体;设计子群体交互特征提取模块,通过构建不同子群体之间的交互矩阵以及图神经网络的关系推理能力,提取更具判别力的子群体特征。实验结果表明,与现有12种方法在排球数据集和集体活动数据集这两个群体行为识别基准数据集上对比,该文方法都取得最好的性能结果。作为一个易于扩展和优化的群体行为识别框架,该算法在不同数据集上都具有较好的泛化能力。

表面相似性与呈现方式对关系类比推理的影响:匹配物效应

关系类比推理是将一种情境中的某种关系推导至另一种情境中的过程,被试由“源问题”中隐含的某种结构关系映射到“靶问题”中也存在这种关系,即进行了关系类比推理。本研究采用图片匹配范式,通过两个实验考察了表面相似性与呈现方式对关系类比推理的影响,其中实验1探讨“源问题”和“靶问题”存在相同匹配物的情景,实验2探讨无相同匹配物的情景。实验1结果发现,同时呈现方式更有利于关系推理,低表面相似性更有利于关系推理;实验2发现,虽然同时呈现方式仍利于关系推理,但高表面相似性反而更利于关系推理。两个实验的对比分析揭示了一种“匹配物效应”:有相同匹配物情景下,低表面相似性比高表面相似性做出了更多的关系推理,而无相同匹配物情景下则出现了相反的结果。根据这些结果,讨论了“积极脑”的理论解释。

基于规则提示的知识图谱通用推理预训练模型

知识图谱是存储真实世界海量知识的图数据库,为大量知识驱动的下游任务提供了数据支持.知识图谱往往具有不完备性,存在大量缺失的事实,因此知识图谱推理任务基于已知事实推理新结论来补全知识图谱.随着知识工程及其商业应用的研究与发展,大量通用和领域知识图谱被构建.现有知识图谱推理方法大多面向单一知识图谱的补全,不具备通用推理能力.近年来,受预训练大语言模型通用能力的启发,一些通用的知识图谱推理预训练模型被提出.针对现有预训练模型无法识别高质量推理模式的问题,提出一个基于规则提示的知识图谱通用推理预训练模型——RulePreM,该模型筛选与利用高质量推理规则来提高知识图谱上的推理能力.首先基于推理规则构建关系IO图和一个编码器RuleGNN对关系进行编码,然后将关系编码作为提示来编码知识图谱中的实体,最后对候选实体进行打分预测.还提出一种结合规则置信度的注意力机制,来进一步减少低质量推理模式的影响.实验结果表明,所提出的模型在43个不同设定下的知识图谱上具有良好的通用推理能力,平均性能指标均优于现有的有监督模型和预训练模型.

A Survey of Knowledge Graph Construction Using Machine Learning

Knowledge graph(KG)serves as a specialized semantic network that encapsulates intricate relationships among real-world entities within a structured framework.This framework facilitates a transformation in information retrieval,transitioning it from mere string matching to far more sophisticated entity matching.In this transformative process,the advancement of artificial intelligence and intelligent information services is invigorated.Meanwhile,the role ofmachine learningmethod in the construction of KG is important,and these techniques have already achieved initial success.This article embarks on a comprehensive journey through the last strides in the field of KG via machine learning.With a profound amalgamation of cutting-edge research in machine learning,this article undertakes a systematical exploration of KG construction methods in three distinct phases:entity learning,ontology learning,and knowledge reasoning.Especially,a meticulous dissection of machine learningdriven algorithms is conducted,spotlighting their contributions to critical facets such as entity extraction,relation extraction,entity linking,and link prediction.Moreover,this article also provides an analysis of the unresolved challenges and emerging trajectories that beckon within the expansive application of machine learning-fueled,large-scale KG construction.

基于证据图推理的文档级实体关系抽取

[研究目的]为缓解文档级实体关系抽取任务中存在的句子噪声问题,提高文档级实体关系抽取性能,提出一种基于证据图推理的文档级实体关系抽取方法,为文档级实体关系抽取和知识发现研究提供参考。[研究方法]通过启发式规则捕获实体对间关系推理所需证据句路径信息;引入图结构学习思想将证据句路径信息融入异构文档图;基于关系图卷积网络进行关系推理以提升文档图对证据句信息的聚合能力;采用前馈神经网络对实体关系进行预测,实现文档级实体关系高效抽取。[研究结论]所提出的模型在国际公开文档级评测数据集CDR和GDA上F1值分别达到71.3%和85.4%,较基准模型EIDER提高1.2%与1.1%。实验结果表明该方法能够有效选择实体关系推理所需证据路径,提升文档级实体关系抽取性能。

基于不确定知识图谱嵌入的多关系近似推理模型

针对大规模知识图谱(KG)的不确定性嵌入模型中无法对多种逻辑关系进行近似推理的问题,提出一种基于不确定KG嵌入(UKGE)的多关系近似推理模型UDConEx(Uncertainty DistMult(Distance Multiplicative) and complex Convolution Embedding)。首先,UDConEx结合DistMult和ComplEx(Complex Embedding)模型的特点,使得UDConEx具有推理对称与非对称关系的能力;其次,UDConEx采用卷积神经网络(CNN)捕获不确定性KG中的交互信息,使它具有推理逆关系和传递关系的能力;最后,UDConEx利用神经网络对KG的不确定信息进行置信度学习,在UKGE空间中可以进行近似推理。在CN15k、NL27k和PPI5k这3个公开数据集上的实验结果表明,相较于MUKGE(Multiplex UKGE)模型,UDConEx在CN15k、NL27k和PPI5k的置信度预测任务中平均绝对误差(MAE)分别降低了6.3%,30.1%和44.9%;在关系事实排名任务中,基于线性的归一化折损累计增益(NDCG)在CN15k和NL27k数据集中分别提升了5.8%和2.6%;在多关系近似推理任务中验证了UDConEx具有多种逻辑关系的近似推理能力。UDConEx弥补了传统嵌入模型无法进行置信度预测的不足,实现了对多种逻辑关系的近似推理,具有更精确、具有可解释性的不确定性知识图谱推理能力。

基于高阶图卷积推理网络的任意形状文本检测

通用场景文本检测被广泛应用于地图导航、无人驾驶等多个领域。场景文本方向各异且形状复杂多变,使得文本检测难度大。针对这一问题,提出一种高阶图卷积推理网络。以文本检测框架DRRG为基础,设计高阶图方案,提出高阶图卷积推理网络,扩展了推理范围,有效组合高阶邻居提供的辅助信息。改进一阶邻居的设置,降低无关组件的干扰,提高了反向传播和组件链接的效率。引入SE聚合模块为每个节点独立且自适应地生成聚合方案,进一步提高了对高阶信息的利用率。实验结果表明,改进后的网络在Total-Text、CTW-1500和ICDAR2015数据集上的平均精度(F1)分别提升了1.4、1.05和1.26个百分点。

年龄相关性听力损失与营养的关系

年龄相关性听力损失是老年人群中常见的慢性疾病,其防治问题受到医学界的高度重视。合理的膳食调整能明显地延缓听力衰退,并降低年龄相关性听力损失的发病率。本文旨在探讨年龄相关性听力损失的成因,深入研究营养素与年龄相关性听力损失之间的关系及作用机制,提出科学的饮食建议,为预防和治疗年龄相关性听力损失提供思路。

Hierarchical Detailed Description for Spatial Direction Relations

Direction relation is an important spatial relation. Descriptions and representations for direction relations have different levels of detail because of the varying dimensions of spatial objects and different scales of the embedding spaces. Based on a direction- relation matrix, the hierarchical frame of spatial direction relations which partitions direction relations orderly and thoroughly is built. Interior direction relations are used to perfect the representation of direction relations and the binary-encoding idea is creatively applied to construct an interior detailed matrix describing multiple interior direction relations by a uniform matrix. The model integrates topological information into the description model for direction relations, which will lay the foundations of spatial compositive reasoning.

Determination of reasonable finished state of self-anchored suspension bridges

A systematic and generic procedure for the determination of the reasonable finished state of self-anchored suspension bridges is proposed, the realization of which is mainly through adjustment of the hanger tensions. The initial hanger tensions are first obtained through an iterative analysis by combining the girder-tower-only finite element(FE) model with the analytical program for shape finding of the spatial cable system. These initial hanger tensions, together with the corresponding cable coordinates and internal forces, are then included into the FE model of the total bridge system, the nonlinear analysis of which involves the optimization technique. Calculations are repeated until the optimization algorithm converges to the most optimal hanger tensions(i.e. the desired reasonable finished bridge state). The "temperature rigid arm" is introduced to offset the unavoidable initial deformations of the girder and tower, which are due to the huge axial forces originated from the main cable. Moreover, by changing the stiffness coefficient K in the girder-tower-only FE model, the stiffness proportion of the main girder, the tower or the cable subsystem in the whole structural system could be adjusted according to the design intentions. The effectiveness of the proposed method is examined and demonstrated by one simple tutorial example and one self-anchored suspension bridge.