A novel fracture-cavity reservoir outcrop geological knowledge base construction method considering parameter collection and processing,mutual transformation of data-knowledge,application and update

This study endeavors to formulate a comprehensive methodology for establishing a Geological Knowledge Base(GKB)tailored to fracture-cavity reservoir outcrops within the North Tarim Basin.The acquisition of quantitative geological parameters was accomplished through diverse means such as outcrop observations,thin section studies,unmanned aerial vehicle scanning,and high-resolution cameras.Subsequently,a three-dimensional digital outcrop model was generated,and the parameters were standardized.An assessment of traditional geological knowledge was conducted to delineate the knowledge framework,content,and system of the GKB.The basic parameter knowledge was extracted using multiscale fine characterization techniques,including core statistics,field observations,and microscopic thin section analysis.Key mechanism knowledge was identified by integrating trace elements from filling,isotope geochemical tests,and water-rock simulation experiments.Significant representational knowledge was then extracted by employing various methods such as multiple linear regression,neural network technology,and discriminant classification.Subsequently,an analogy study was performed on the karst fracture-cavity system(KFCS)in both outcrop and underground reservoir settings.The results underscored several key findings:(1)Utilization of a diverse range of techniques,including outcrop observations,core statistics,unmanned aerial vehicle scanning,high-resolution cameras,thin section analysis,and electron scanning imaging,enabled the acquisition and standardization of data.This facilitated effective management and integration of geological parameter data from multiple sources and scales.(2)The GKB for fracture-cavity reservoir outcrops,encompassing basic parameter knowledge,key mechanism knowledge,and significant representational knowledge,provides robust data support and systematic geological insights for the intricate and in-depth examination of the genetic mechanisms of fracture-cavity reservoirs.(3)The developmental characteristics of fracturecavities in karst outcrops offer effective,efficient,and accurate guidance for fracture-cavity research in underground karst reservoirs.The outlined construction method of the outcrop geological knowledge base is applicable to various fracture-cavity reservoirs in different layers and regions worldwide.

Application of sparse S transform network with knowledge distillation in seismic attenuation delineation

Time-frequency analysis is a successfully used tool for analyzing the local features of seismic data.However,it suffers from several inevitable limitations,such as the restricted time-frequency resolution,the difficulty in selecting parameters,and the low computational efficiency.Inspired by deep learning,we suggest a deep learning-based workflow for seismic time-frequency analysis.The sparse S transform network(SSTNet)is first built to map the relationship between synthetic traces and sparse S transform spectra,which can be easily pre-trained by using synthetic traces and training labels.Next,we introduce knowledge distillation(KD)based transfer learning to re-train SSTNet by using a field data set without training labels,which is named the sparse S transform network with knowledge distillation(KD-SSTNet).In this way,we can effectively calculate the sparse time-frequency spectra of field data and avoid the use of field training labels.To test the availability of the suggested KD-SSTNet,we apply it to field data to estimate seismic attenuation for reservoir characterization and make detailed comparisons with the traditional time-frequency analysis methods.

Could One Transformational Leader Convert the Organization From Knowledge Based Into Learning Organization, Then Into Innovation？

This paper examines the role of transformational leadership in transforming an organization into a knowledge based, then into learning organization so that it becomes an innovative company. Important features of the leader such and ability to assist in developing and accommodating the implementation of knowledge management programs, learning organization concepts and innovation protocols are discussed in this paper. This paper demonstrates that shifting the organization to become a knowledge based and then to be learning organization and finally to become innovative company could involve some unique attributes of a transformation leadership. In that regards, the paper also demonstrates that organizations need first to create, capture, transfer, and mobilize knowledge before it can be used for learning and then for innovation. The paper will present a method of a studying how successful innovation leaders of companies could found themselves acting in three roles namely： knowledge leader, learning leader and then innovation leader.

Clinical Knowledge-Based Hybrid Swin Transformer for Brain Tumor Segmentation

Accurate tumor segmentation from brain tissues in Magnetic Resonance Imaging(MRI)imaging is crucial in the pre-surgical planning of brain tumor malignancy.MRI images’heterogeneous intensity and fuzzy boundaries make brain tumor segmentation challenging.Furthermore,recent studies have yet to fully employ MRI sequences’considerable and supplementary information,which offers critical a priori knowledge.This paper proposes a clinical knowledge-based hybrid Swin Transformermultimodal brain tumor segmentation algorithmbased on how experts identify malignancies from MRI images.During the encoder phase,a dual backbone network with a Swin Transformer backbone to capture long dependencies from 3D MR images and a Convolutional Neural Network(CNN)-based backbone to represent local features have been constructed.Instead of directly connecting all the MRI sequences,the proposed method re-organizes them and splits them into two groups based on MRI principles and characteristics:T1 and T1ce,T2 and Flair.These aggregated images are received by the dual-stem Swin Transformer-based encoder branch,and the multimodal sequence-interacted cross-attention module(MScAM)captures the interactive information between two sets of linked modalities in each stage.In the CNN-based encoder branch,a triple down-sampling module(TDsM)has been proposed to balance the performance while downsampling.In the final stage of the encoder,the feature maps acquired from two branches are concatenated as input to the decoder,which is constrained by MScAM outputs.The proposed method has been evaluated on datasets from the MICCAI BraTS2021 Challenge.The results of the experiments demonstrate that the method algorithm can precisely segment brain tumors,especially the portions within tumors.

Progress and Knowledge Transfer from Science to Technology in the Research Frontier of CRISPR Based on the LDA Model

Purpose:This study explores the underlying research topics regarding CRISPR based on the LDA model and figures out trends in knowledge transfer from science to technology in this area over the latest 10 years.Design/methodology/approach:We collected publications on CRISPR between 2011 and2020 from the Web of Science,and traced all the patents citing them from lens.org.15,904 articles and 18,985 patents in total are downloaded and analyzed.The LDA model was applied to identify underlying research topics in related research.In addition,some indicators were introduced to measure the knowledge transfer from research topics of scientific publications to IPC-4 classes of patents.Findings:The emerging research topics on CRISPR were identified and their evolution over time displayed.Furthermore,a big picture of knowledge transition from research topics to technological classes of patents was presented.We found that for all topics on CRISPR,the average first transition year,the ratio of articles cited by patents,the NPR transition rate are respectively 1.08,15.57%,and 1.19,extremely shorter and more intensive than those of general fields.Moreover,the transition patterns are different among research topics.Research limitations:Our research is limited to publications retrieved from the Web of Science and their citing patents indexed in lens.org.A limitation inherent with LDA analysis is in the manual interpretation and labeling of"topics".Practical implications:Our study provides good references for policy-makers on allocating scientific resources and regulating financial budgets to face challenges related to the transformative technology of CRISPR.Originality/value:The LDA model here is applied to topic identification in the area of transformative researches for the first time,as exemplified on CRISPR.Additionally,the dataset of all citing patents in this area helps to provide a full picture to detect the knowledge transition between S&T.

一种基于安全多方计算的快速Transformer安全推理方案

Transformer模型在自然语言处理、计算机视觉等众多领域得到了广泛应用,并且有着突出的表现.在Transformer的推理应用中用户的数据会被泄露给模型提供方.随着数据隐私问题愈发得到公众的关注,上述数据泄露问题引发了学者们对Transformer安全推理的研究,使用安全多方计算(secure multi-party computation,MPC)实现Transformer模型的安全推理是当前的一个研究热点.由于Transformer模型中存在大量非线性函数,因此使用MPC技术实现Transformer安全推理会造成巨大的计算和通信开销.针对Transformer安全推理过程中开销较大的Softmax注意力机制,提出了2种MPC友好的注意力机制Softmax freeDiv Attention和2Quad freeDiv Attention.通过将Transformer模型中的Softmax注意力机制替换为新的MPC友好的注意力机制,同时结合激活函数GeLU的替换以及知识蒸馏技术,提出了一个MPC友好的Transformer转换框架,通过将Transformer模型转化为MPC友好的Transformer模型,提高Transformer安全推理的效率.在局域网环境下使用安全处理器(secure processing unit,SPU)提供的隐私计算协议,基于所提出的MPC友好的Transformer转换框架,在SST-2上使用Bert-Base进行安全推理.测试结果表明,在保持推理准确率与无近似模型一致的情况下,安全推理计算效率提高2.26倍.

Influence of Network Embedding on the Transformation Performance of Original Equipment Manufactures

Based on the embeddedness theory,this paper analyzes the impact of network embedding features and types on the transformation performance of foundry enterprises.Through the investigation and analysis of more than 200 foundry enterprises in Zhejiang Province,it is found that network relationship embedding and structure embedding have positive impacts on transformation of foundry enterprises.Professional embedding and technical embedding have a positive effect in the transformation of foundry enterprises,and knowledge absorption ability has a positive adjustment role in the transformation performance of network embedding foundry enterprises.

Fuzzy Methodology for Taxonomy and Knowledge Base Design

This paper summarizes the research results dealing with washer and nut taxonomy and knowledge base design, making the use of fuzzy methodology. In particular, the theory of fuzzy membership functions, similarity matrices, and the operation of fuzzy inference play important roles.A realistic set of 25 washers and nuts are employed to conduct extensive experiments and simulations.The investigation includes a complete demonstration of engineering design. The results obtained from this feasibility study are very encouraging indeed because they represent the lower bound with respect to performance, namely correctrecognition rate, of what fuzzy methodology can do. This lower bound shows high recognition rate even with noisy input patterns, robustness in terms of noise tolerance, and simplicity in hardware implementation. Possible future works are suggested in the conclusion.

Leadership,Knowledge Management,and Human Capital Development

This research highlights the need to develop a framework for leadership,human capital development,and knowledge management by reviewing existing literature in the field of research.The main aim of this research is to propose a model which supports the relationship between leadership(servant leadership,transformational leadership)and human capital development.The study also proposes that knowledge management(knowledge sharing,knowledge acquisition)will moderate the relationship between leadership(servant leadership,transformational leadership)and human capital development.A set of propositions that represent an empirically-driven research agenda,and also describe the relationships between the focal variables are presented to enhance audience’s understanding within a business context.

Basic Elements Knowledge Acquisition Study in the Chinese Character Intelligent Formation System

In the Chinese character intelligent formation system without Chinese character library, it is possible that the same basic element in different Chinese characters is different in position, size and shape. The geometry transformation from basic elements to the components of Chinese characters can be realized by affine transformation, the transformation knowledge acquisition is the premise of Chinese character intelligent formation. A novel algorithm is proposed to ac-quire the affine transformation knowledge of basic elements automatically in this paper. The interested region of Chi-nese character image is determined by the structure of the Chinese character. Scale invariant and location invariant of basic element and Chinese character image are extracted with SIFT features, the matching points of the two images are determined according to the principle of Minimum Euclidean distance of eigenvectors. Using corner points as identifi-cation features, calculating the one-way Hausdorff distance between corner points as the similarity measurement from the affine image to the Chinese character sub-image, affine coefficients are determined by optimal similarity. 70244 Chinese characters in National Standards GB18030-2005 character set are taken as the experimental object, all the characters are performed and the experimental courses and results are presented in this paper.

Successful Digital Transformation and HRM Humanization

This paper investigates a core role of Human Resource Management(HRM)humanization for successful digital transformation in digital economy.The term“humanization”is applied to an iterative method of human relation development for human resources satisfaction and high results of organizational performance.The author summarized the peculiarities of digitalization in Russian companies in the context of the Russian labor market trends.The paper focuses on factors that determine human potential utilization and development in modern condition.The author grounded linkage between HRM humanization and digital transformation projects effect by three examples of Russian companies.The results let us conclude importance of HRM humanization and define core problems and directions in Russian context.

Network Configuration Entity Extraction Method Based on Transformer with Multi-Head Attention Mechanism

Nowadays,ensuring thequality of networkserviceshas become increasingly vital.Experts are turning toknowledge graph technology,with a significant emphasis on entity extraction in the identification of device configurations.This research paper presents a novel entity extraction method that leverages a combination of active learning and attention mechanisms.Initially,an improved active learning approach is employed to select the most valuable unlabeled samples,which are subsequently submitted for expert labeling.This approach successfully addresses the problems of isolated points and sample redundancy within the network configuration sample set.Then the labeled samples are utilized to train the model for network configuration entity extraction.Furthermore,the multi-head self-attention of the transformer model is enhanced by introducing the Adaptive Weighting method based on the Laplace mixture distribution.This enhancement enables the transformer model to dynamically adapt its focus to words in various positions,displaying exceptional adaptability to abnormal data and further elevating the accuracy of the proposed model.Through comparisons with Random Sampling(RANDOM),Maximum Normalized Log-Probability(MNLP),Least Confidence(LC),Token Entrop(TE),and Entropy Query by Bagging(EQB),the proposed method,Entropy Query by Bagging and Maximum Influence Active Learning(EQBMIAL),achieves comparable performance with only 40% of the samples on both datasets,while other algorithms require 50% of the samples.Furthermore,the entity extraction algorithm with the Adaptive Weighted Multi-head Attention mechanism(AW-MHA)is compared with BILSTM-CRF,Mutil_Attention-Bilstm-Crf,Deep_Neural_Model_NER and BERT_Transformer,achieving precision rates of 75.98% and 98.32% on the two datasets,respectively.Statistical tests demonstrate the statistical significance and effectiveness of the proposed algorithms in this paper.

基于知识蒸馏的轻量化Transformer目标检测

在自动驾驶领域,目标检测的高效性和准确性尤为重要,基于Transformer结构的目标检测方法逐渐成为主流,省去了复杂的锚点生成和非极大值抑制。针对现有方法计算成本高和收敛速度慢的问题,设计了一种基于池化操作的轻量化Transformer目标检测模型(LPT),包含了池化主干网络和双池化注意力机制,设计了针对DETR(detection transformer)模型的通用知识蒸馏方法,将预测结果、查询向量和教师提取的特征作为知识传递给轻量化的Transformer模型,帮助其提升精确度性能。通过在MS COCO 2017数据集上的实验,验证经过蒸馏的LPT模型在自动驾驶中的应用潜力,实验结果表明:本文方法具有较好的准确性,与一些先进的方法相比具有一定优势。

基于自蒸馏视觉Transformer的无监督行人重识别

针对视觉Transformer(ViT)缺乏归纳偏置,导致在相对小规模的行人重识别数据上难以学习有意义的视觉表征的问题,提出一种基于自蒸馏视觉Transformer的无监督行人重识别方法。首先,利用ViT的模块化架构,即每个中间块生成的特征维度相同的特性,随机选择一个中间Transformer块并将它送入分类器以得到预测结果;其次,通过最小化随机选择的中间分类器输出与最终分类器输出分布之间的Kullback-Leibler散度,约束中间块的分类预测结果与最终分类器的结果保持一致,据此构建自蒸馏损失函数;最后,通过对聚类级对比损失、实例级对比损失和自蒸馏损失进行联合最小化,对模型进行优化。此外,通过从最终分类器向中间块提供软监督,有效地给ViT模型引入归纳偏置,进而有助于模型学习更鲁棒和通用的视觉表征。与基于TransReID的自监督学习(TransReID-SSL)相比,在Market-1501数据集上,所提方法的平均精度均值(mAP)和Rank-1分别提升1.2和0.8个百分点;在MSMT17数据集上,所提方法的mAP和Rank-1分别提升3.4和3.1个百分点。实验结果表明,所提方法能够有效提高无监督行人重识别的精度。

KDTGAN:基于Transformer-GAN和知识蒸馏的高光谱目标检测

高光谱目标检测在地球观测中至关重要,被广泛应用于军事和民用领域。然而,由于高光谱图像的背景复杂性和目标样本的有限性,该任务面临较大的挑战。本文首先采用CEM (约束能量最小化)粗检测方法提取背景数据。随之,引入了一种新的知识蒸馏模型,即KDTGAN (通过Transformer-GAN实现)。教师模型的生成器采用了Transformer编码器的结构,并结合多尺度数据融合的方法,能够准确地学习背景分布,进而通过重构背景信息实现目标检测。为了克服GAN(生成对抗网络)训练不稳定的挑战,特别是纯背景数据的稀缺性,本文提出了一种新的损失算法,以减小可疑目标样本对模型性能的负面影响。为了降低模型的计算负担,本文引入知识蒸馏,并设计新的蒸馏损失对学生模型加以约束,使模型轻量化的同时提高学生模型检测精度。实验结果表明:KDTGAN相较于当前检测方法表现更优,具有更高的检测精度和鲁棒性。

基于改进Transformer模型的文本摘要生成方法

传统的文本摘要方法,如基于循环神经网络和Encoder-Decoder框架构建的摘要生成模型等,在生成文本摘要时存在并行能力不足或长期依赖的性能缺陷,以及文本摘要生成的准确率和流畅度的问题。对此,提出了一种动态词嵌入摘要生成方法。该方法基于改进的Transformer模型,在文本预处理阶段引入先验知识,将ELMo(Embeddings from Language Models)动态词向量作为训练文本的词表征,结合此词对应当句的文本句向量拼接生成输入文本矩阵,将文本矩阵输入到Encoder生成固定长度的文本向量表达,然后通过Decoder将此向量表达解码生成目标文本摘要。实验采用Rouge值作为摘要的评测指标,与其他方法进行的对比实验结果表明,所提方法所生成的文本摘要的准确率和流畅度更高。

Transformer优化及其在苹果病虫命名实体识别中的应用

为提高苹果生产领域实体识别的准确性,提出一种新的Transformer优化模型。首先,为解决苹果生产数据集的缺失,基于苹果栽培领域园艺专家的知识经验,创建以苹果病虫害为主的产业数据集。通过字向量与词向量的拼接,提高文本语义表征的准确性;随后,为防止位置信息缺失,引入具有方向和距离感知的注意力机制,平均集成BiLSTM的上下文长距离依赖特征;最后,结合条件随机场(Conditional random fields, CRF)约束上下文标注结果,最终得到Transformer优化模型。实验结果表明,所提方法在苹果病虫命名实体识别中的F1值可达92.66%,可为农业命名实体的准确智能识别提供技术手段。

Knowledge Graph Enhanced Transformers for Diagnosis Generation of Chinese Medicine

Chinese medicine(CM)diagnosis intellectualization is one of the hotspots in the research of CM modernization.The traditional CM intelligent diagnosis models transform the CM diagnosis issues into classification issues,however,it is difficult to solve the problems such as excessive or similar categories.With the development of natural language processing techniques,text generation technique has become increasingly mature.In this study,we aimed to establish the CM diagnosis generation model by transforming the CM diagnosis issues into text generation issues.The semantic context characteristic learning capacity was enhanced referring to Bidirectional Long Short-Term Memory(BILSTM)with Transformer as the backbone network.Meanwhile,the CM diagnosis generation model Knowledge Graph Enhanced Transformer(KGET)was established by introducing the knowledge in medical field to enhance the inferential capability.The KGET model was established based on 566 CM case texts,and was compared with the classic text generation models including Long Short-Term Memory sequence-to-sequence(LSTM-seq2seq),Bidirectional and Auto-Regression Transformer(BART),and Chinese Pre-trained Unbalanced Transformer(CPT),so as to analyze the model manifestations.Finally,the ablation experiments were performed to explore the influence of the optimized part on the KGET model.The results of Bilingual Evaluation Understudy(BLEU),Recall-Oriented Understudy for Gisting Evaluation 1(ROUGE1),ROUGE2 and Edit distance of KGET model were 45.85,73.93,54.59 and 7.12,respectively in this study.Compared with LSTM-seq2seq,BART and CPT models,the KGET model was higher in BLEU,ROUGE1 and ROUGE2 by 6.00–17.09,1.65–9.39 and 0.51–17.62,respectively,and lower in Edit distance by 0.47–3.21.The ablation experiment results revealed that introduction of BILSTM model and prior knowledge could significantly increase the model performance.Additionally,the manual assessment indicated that the CM diagnosis results of the KGET model used in this study were highly consistent with the practical diagnosis results.In conclusion,text generation technology can be effectively applied to CM diagnostic modeling.It can effectively avoid the problem of poor diagnostic performance caused by excessive and similar categories in traditional CM diagnostic classification models.CM diagnostic text generation technology has broad application prospects in the future.

Efficient knowledge distillation for hybrid models:A vision transformer‐convolutional neural network to convolutional neural network approach for classifying remote sensing images

In various fields,knowledge distillation(KD)techniques that combine vision transformers(ViTs)and convolutional neural networks(CNNs)as a hybrid teacher have shown remarkable results in classification.However,in the realm of remote sensing images(RSIs),existing KD research studies are not only scarce but also lack competitiveness.This issue significantly impedes the deployment of the notable advantages of ViTs and CNNs.To tackle this,the authors introduce a novel hybrid‐model KD approach named HMKD‐Net,which comprises a CNN‐ViT ensemble teacher and a CNN student.Contrary to popular opinion,the authors posit that the sparsity in RSI data distribution limits the effectiveness and efficiency of hybrid‐model knowledge transfer.As a solution,a simple yet innovative method to handle variances during the KD phase is suggested,leading to substantial enhancements in the effectiveness and efficiency of hybrid knowledge transfer.The authors assessed the performance of HMKD‐Net on three RSI datasets.The findings indicate that HMKD‐Net significantly outperforms other cuttingedge methods while maintaining a significantly smaller size.Specifically,HMKD‐Net exceeds other KD‐based methods with a maximum accuracy improvement of 22.8%across various datasets.As ablation experiments indicated,HMKD‐Net has cut down on time expenses by about 80%in the KD process.This research study validates that the hybrid‐model KD technique can be more effective and efficient if the data distribution sparsity in RSIs is well handled.

基于Transformer融合词性特征的中文语法纠错模型

针对中文同一个词的不同词性在句子中所代表的关系不同的问题,提出基于Transformer融合词性特征的中文语法纠错(CGEC)模型,所提模型将语言学知识作为辅助信息融入中文语法纠错任务。首先,在不改变句子序列长度的基础上,在原始词嵌入层中以不同方式拼接词性向量,得到全差异词嵌入、词差异词嵌入和词性差异词嵌入三种不同的词嵌入方式;然后,将新的词嵌入方式与Transformer模型相结合,对错误语句进行语法纠错。实验结果表明,三种词嵌入方式均不同程度地提高了F0.5值,且全差异词嵌入方式的效果最好:与Transformer模型相比,F0.5提升了2.73个百分点,BLEU提升了6.27个百分点;与基于Transformer增强架构的中文语法纠错模型相比,F0.5提升了1.88个百分点。所提模型在对词性特征提取时可以侧重源语句与目标语句的语法差异,更好地捕捉句子的语法特征。