Directly predicting N_(2) electroreduction reaction free energy using interpretable machine learning with non-DFT calculated features

Electrocatalytic nitrogen reduction to ammonia has garnered significant attention with the blooming of single-atom catalysts(SACs),showcasing their potential for sustainable and energy-efficient ammonia production.However,cost-effectively designing and screening efficient electrocatalysts remains a challenge.In this study,we have successfully established interpretable machine learning(ML)models to evaluate the catalytic activity of SACs by directly and accurately predicting reaction Gibbs free energy.Our models were trained using non-density functional theory(DFT)calculated features from a dataset comprising 90 graphene-supported SACs.Our results underscore the superior prediction accuracy of the gradient boosting regression(GBR)model for bothΔg(N_(2)→NNH)andΔG(NH_(2)→NH_(3)),boasting coefficient of determination(R^(2))score of 0.972 and 0.984,along with root mean square error(RMSE)of 0.051 and 0.085 eV,respectively.Moreover,feature importance analysis elucidates that the high accuracy of GBR model stems from its adept capture of characteristics pertinent to the active center and coordination environment,unveilling the significance of elementary descriptors,with the colvalent radius playing a dominant role.Additionally,Shapley additive explanations(SHAP)analysis provides global and local interpretation of the working mechanism of the GBR model.Our analysis identifies that a pyrrole-type coordination(flag=0),d-orbitals with a moderate occupation(N_(d)=5),and a moderate difference in covalent radius(r_(TM-ave)near 140 pm)are conducive to achieving high activity.Furthermore,we extend the prediction of activity to more catalysts without additional DFT calculations,validating the reliability of our feature engineering,model training,and design strategy.These findings not only highlight new opportunity for accelerating catalyst design using non-DFT calculated features,but also shed light on the working mechanism of"black box"ML model.Moreover,the model provides valuable guidance for catalytic material design in multiple proton-electron coupling reactions,particularly in driving sustainable CO_(2),O_(2),and N_(2) conversion.

The Correlation Between Note Features and Consecutive Interpreting Quality for English Majors

Note-taking skill is a necessary component in interpreter training programs,and previous research has yielded findings such as note-taking training methods or features of interpreter trainees’notes.However,little research has been done to investigate the changes in note features and correlations between note features and interpreting quality concerning Chinese students’C-E(Chinese-English)and E-C(EnglishChinese)interpreting.Using the framework of Daniel Gile’s Effort Model and Interpretive Theory of Translation,this paper examined how 45 English Majors’notes develop within one semester(seventeen weeks)and the relationship between note features(quantity,form,and language choice of notes)and consecutive interpreting quality.The participants of this study were all beginner interpreting trainees,and the note-taking training was introduced in Week 6.The study employed note manuscripts,interpreting tests,and semi-structured interviews to track the features and changes in students’notes.Correlation analyses and T-tests showed that(a)after the note-taking training,the number of notes increased from Week 8 to Week 17,and it was positively correlated with interpreting quality(fidelity and delivery)for both C-E and E-C interpreting;(b)as for forms of notes,participants primarily employ single Chinese words and the percentages of abbreviations and symbols rose prominently from Week 8 to Week 17 for C-E interpreting.Besides,correlation analyses show that interpreting quality improves with fewer single Chinese words and more abbreviations and symbols.For E-C interpreting,notes were mainly in English,especially single English words and abbreviations.The percentages of single Chinese words and abbreviations ascended whereas those of single English words and symbols decreased.Furthermore,results show that the more abbreviations and symbols,the better target-text fidelity,and fewer abbreviations,the better the targettext delivery;(c)concerning language choice,notes were mainly in source language for both C-E and E-C interpreting and the percentage of target language notes went up significantly for C-E interpreting.Consequently,the percentage of target language notes was positively correlated with interpreting quality.Interviews indicate that most participants do not pay much attention to language selection in the first stage,and if the source text a familiar topic with little difficult vocabulary,he or she records the target language.Otherwise,it was safer to use the source language.

The Role and features of Student Peer Feedback in Interpreting Training——Preliminary Findings of a Survey of both trainers and students

The project delves into the preliminary findings of a survey of both trainers and students on the practice of using student peer feedback in interpreting practice.It first explains the theoretical foundation which justifies the use of peer feedback in interpreting practice,the research methodology and data collection.Then it brings forth specific findings concerning the implementation of peer feedback in the interpreting class followed by discussions of the role and features of student peer feedback as a means to help students ready for the booth.Analysis of the results shows that peer feedback in interpreting practice keeps students on-task,attentive and help them spot their own problems.Trainers and students themselves point to similar features of student peer feedback as focusing on comprehension of the original,word choice and numbers.The preliminary findings of the survey demonstrate the roles and features of student peer feedback in interpreting practice and point to the possible way of enhancing student’s learning curve through more effective peer feedback.

On the Verbal and Non-Verbal Features in Chinese-English Interpretation

This paper investigates the verbal and non - verbal features of interpretation from Chinese into English . On the one hand the language of interpretation belongs to the category of oral language, So It determines the path an interpreter should follow while interpreting . On the other hand it is suggested that the non - verbal approach plays an important role in interpretation. Therefore an interpreter can not be a qualified interpreter unless he is, in addition to language techniques, skilled in the application of paralanguage.

Artificial intelligence-driven radiomics study in cancer:the role of feature engineering and modeling

Modern medicine is reliant on various medical imaging technologies for non-invasively observing patients’anatomy.However,the interpretation of medical images can be highly subjective and dependent on the expertise of clinicians.Moreover,some potentially useful quantitative information in medical images,especially that which is not visible to the naked eye,is often ignored during clinical practice.In contrast,radiomics performs high-throughput feature extraction from medical images,which enables quantitative analysis of medical images and prediction of various clinical endpoints.Studies have reported that radiomics exhibits promising performance in diagnosis and predicting treatment responses and prognosis,demonstrating its potential to be a non-invasive auxiliary tool for personalized medicine.However,radiomics remains in a developmental phase as numerous technical challenges have yet to be solved,especially in feature engineering and statistical modeling.In this review,we introduce the current utility of radiomics by summarizing research on its application in the diagnosis,prognosis,and prediction of treatment responses in patients with cancer.We focus on machine learning approaches,for feature extraction and selection during feature engineering and for imbalanced datasets and multi-modality fusion during statistical modeling.Furthermore,we introduce the stability,reproducibility,and interpretability of features,and the generalizability and interpretability of models.Finally,we offer possible solutions to current challenges in radiomics research.

Features of E-C Public Speech Interpreting: A Case Study of the Interpretation of Obama's Speeches

This paper is trying to analyze the E-C interpreting scripts of Inaugural Address, Remarks on Winning the Nobel Prize and Shanghai Speech by the 44th president of United States Barack Obama with a comparative method based on data collected. The analysis will be employed on the lexical, syntactic as well as rhetorical level and the features of E-C public speech interpreting will be achieved accordingly. The features may serve as reference for the interpreters in their interpretation practice in order to improve the interpretation effects.

An improved deep dilated convolutional neural network for seismic facies interpretation

With the successful application and breakthrough of deep learning technology in image segmentation,there has been continuous development in the field of seismic facies interpretation using convolutional neural networks.These intelligent and automated methods significantly reduce manual labor,particularly in the laborious task of manually labeling seismic facies.However,the extensive demand for training data imposes limitations on their wider application.To overcome this challenge,we adopt the UNet architecture as the foundational network structure for seismic facies classification,which has demonstrated effective segmentation results even with small-sample training data.Additionally,we integrate spatial pyramid pooling and dilated convolution modules into the network architecture to enhance the perception of spatial information across a broader range.The seismic facies classification test on the public data from the F3 block verifies the superior performance of our proposed improved network structure in delineating seismic facies boundaries.Comparative analysis against the traditional UNet model reveals that our method achieves more accurate predictive classification results,as evidenced by various evaluation metrics for image segmentation.Obviously,the classification accuracy reaches an impressive 96%.Furthermore,the results of seismic facies classification in the seismic slice dimension provide further confirmation of the superior performance of our proposed method,which accurately defines the range of different seismic facies.This approach holds significant potential for analyzing geological patterns and extracting valuable depositional information.

Analysis on the Relationship among Memorizing Ability, Note-taking, and Psychological Factors in Interpretation Teaching

Since China＇s reform and accession to the world trade organization （WTO）, the international labor division, cooperation, and communication of production have been a direction of the global productivity development. The expansion of the world trade organization and the growth of multinational corporations have promoted China＇s market economy to head for internationalization; the exchanges of all countries＇ science and technology, culture, and education have become increasingly frequent and beyond the borders very early, so opening-up has become a world trend. China is playing an important role in this historical process. In this situation, the importance of language translation is increasing day by day, and an unprecedented prosperous situation has especially appeared to the interpretation. In order to accord with this new historical situation, many colleges and universities have set up language translation program and have offered interpretation course. To promote the economic development of China and train more senior interpreters, the studies of interpretation teaching and the improvement of interpretation teaching quality have been particularly important. In this paper, the relationship among memory, note-taking, and psychological factors are analyzed.

A Preliminary Analysis on the Note-taking in English Interpretation

Interpretation is a very immediate translation practice.To be a qualified interpreter,you must have a good bilingual ability,rich knowledge as well as some relevant skills.The scene of interpretation is very complicated,so interpreters often can not successfully com plete the task of interpretation only by memory.They have to rely on the help of taking some notes.Therefore,note-taking is of essential importance in interpreting and it is very important in interpreting practice.This paper mainly make a brief summary of five kinds of note-taking symbols commonly used and illustrates each one by one.Besides,this paper puts forward practical training methods of note-taking.

The Skills of Note-taking in Consecutive Interpretation from the Perspective of Gile's Effort Models

This paper introduces the method of note-taking based on the Gile's Effort Models,exploring how tokeep the balance of memory and note-aking in consecutive interpreting.The paper also analyses some examples tofind an effective way to balance the memory and note-taking in consecutive interpreting,so as to help interpreters tobetter convey the meaning of speakers accurately and quickly with the help of interpreting notes,thereby improvingthe quality of interpreting.

Analysis of Feature Importance and Interpretation for Malware Classification

This study was conducted to enable prompt classification of malware,which was becoming increasingly sophisticated.To do this,we analyzed the important features of malware and the relative importance of selected features according to a learning model to assess how those important features were identified.Initially,the analysis features were extracted using Cuckoo Sandbox,an open-source malware analysis tool,then the features were divided into five categories using the extracted information.The 804 extracted features were reduced by 70%after selecting only the most suitable ones for malware classification using a learning model-based feature selection method called the recursive feature elimination.Next,these important features were analyzed.The level of contribution from each one was assessed by the Random Forest classifier method.The results showed that System call features were mostly allocated.At the end,it was possible to accurately identify the malware type using only 36 to 76 features for each of the four types of malware with the most analysis samples available.These were the Trojan,Adware,Downloader,and Backdoor malware.

基于FeatureStation的地理国情地表覆盖解译方案研究

分析了FeatureStation用于地理国情要素提取与解译的技术路线,并提出了特殊地貌的解译方法,以及针对选定区域的自动解译方法。在地理国情普查项目中的应用表明,FeatureStation软件具有快速、实用、有针对性等特点。

一种可解释的云平台任务终止状态预测方法

基于特征选择和模型可解释方法构建可解释性强的云平台任务终止状态预测模型,该模型可视化任务/作业的静态和动态属性与终止状态之间的映射关系,进而找出负载特征与任务终止状态之间的映射机理.利用Google公开的工作负载监控日志,并加入云平台中任务的动态信息,采用沙普利加和解释(Shapley additive explain,SHAP)找出静态和动态属性对终止状态影响的重要性,利用变量重要性结合SHAP值和XGBoost模型,对任务终止状态预测模型建模后的结果进行解释,使用可视化技术呈现负载特征如何影响模型对不同任务终止状态的预测.用SHAP值绝对值的平均值衡量特征的重要性,实现任务不同终止状态特征重要性的全局可视化,根据结果筛选出对任务终止状态预测模型影响大的20个变量,作为特征筛选的依据;由可视化的结果可知,任务运行过程中,各特征的不同特征值对任务的终止状态有影响,不同特征值对终止状态的产生有不同的影响.特征选择结合模型可解释性方法运用于任务终止状态预测模型的构建流程中,可辅助构建高分类性能及易于理解的任务终止状态预测模型,通过对负载特征与任务终止状态之间映射机理的探索,可以优化云平台的调度机制.

利用知识图谱的多跳可解释问答

基于知识图谱的多跳问答需要分析和理解自然语言问题并在知识图谱的实体和关系上经过多次推理获取答案,是自然语言处理的重要研究领域.现有的模型一般通过知识图谱与问题嵌入,利用神经网络推断答案;或使用一阶逻辑规则结合概率方法预测答案;前者缺乏可解释性,后者在复杂问题中性能欠佳.为解决上述问题,本文提出一种基于知识图谱的多跳可解释问答方法(MIQA),它通过在实体间的多次跳跃推理来获取答案.MIQA首先使用BERT预训练模型获取自然语言问题表征向量以及问题分词后的词向量矩阵,在每一跳中,结合问题向量提取问题当前时刻的特征向量,根据特征向量的分类结果计算下一跳的关系分数和实体分数,多次跳跃后,综合分数最高的实体被作为答案,而获取该答案所对应的路径为推理路径.该方法推理准确率高,同时具有明显的可解释性.在MetaQA、WebQuestionsSP、ComplexWebQuestions这3个数据集上,通过和其他8个知名算法相比较,仿真结果表明MIQA性能优异,达到了当前的SOTA.

拉普拉斯卷积的双路径特征融合遥感图像智能解译方法

由于遥感图像存在多尺度变化和目标边缘模糊等问题,对其进行智能解译仍然是一项极具挑战性的工作。传统的语义分割方法在处理这些问题时存在局限性,难以有效捕捉全局和局部信息。针对上述问题,文中提出一种双路径特征融合分割方法 DFNet。首先,使用Swin Transformer作为主干提取全局语义特征,以处理像素之间的长距离依赖关系,从而促进对图像中不同区域相关性的理解;其次,将拉普拉斯卷积嵌入到空间分支,以捕获局部细节信息,加强目标地物边缘信息表达;最后,引入多尺度双向特征融合模块,充分利用图像中的全局和局部信息,以增强多尺度信息的获取能力。在实验中,使用了三个公开的高分辨率遥感图像数据集进行验证,并通过消融实验验证了所提模型不同模块的作用。实验结果表明,所提方法在Uavid数据集、Potsdam数据集、LoveDA数据集的mIoU达到了71.32%、85.58%、54.01%,提高了语义分割的性能,使分割结果更为精细。

基于层次分析法的口译质量评估

近年来,评分量表在口译质量评估中得到广泛运用。基于《口译能力特征量表》,采用层次分析法,比较并计算了已有口译质量研究文献中主要参数的排名和权重,从而利用交叉学科方法优化了该量表的评分机制。研究发现:“内容”被视为口译质量评估最重要的参数,其权重占总数的62%;“表达”次之,占24%;“交互”权重最低,仅占14%。研究结果有助于克服以口译经验和思辨总结为导向进行质量评估的不足,提升了《口译能力特征量表》在课堂教学评估、自主学习评价和高风险测试公平三方面的科学性。

法庭口译话语特征研究——以运动员孙杨CAS兴奋剂仲裁案口译为例

孙杨兴奋剂仲裁案的听证会是具有重大国际影响的体育仲裁事件。过程中的口译质量问题一度成为口译和涉外法律研究的焦点话题。本文以法庭口译的质量要求为基准,以听证会的话轮和话语特征为研究对象,剖析法庭质证过程中的口译质量问题与存在原因。本研究旨在帮助译员熟悉范式、降低负荷、提供策略,促进法庭口译研究和人才培养,响应加速涉外法治人才培养的国家政策。

问题导向学习法指导下的英语专业课程思政教育——以“科技英语口译”课程为例

基于课程思政教育思想,认为目前高校专业课程与课程思政融合特色不够突出。以上海理工大学的“科技英语口译”课程为例,基于该课程口译材料内容多为科技知识或事件的特色,以“问题导向学习法”(PBL)为指导理论,引导学生思考口译原文讲述的现实问题及其解决方法。课程实践结果表明,该方法可以促进学生对科技词汇中英文表达的理解和记忆,提高科技口译能力和发现并解决问题的能力,对中外科技史上由现实问题触发的科技进步达到更深刻的认识,从而增强爱国热情,培养国际视野。

基于学习的源代码漏洞检测研究与进展

源代码漏洞自动检测是源代码漏洞修复的前提和基础,对于保障软件安全具有重要意义.传统的方法通常是基于安全专家人工制定的规则检测漏洞,但是人工制定规则的难度较大,且可检测的漏洞类型依赖于安全专家预定义的规则.近年来,人工智能技术的快速发展为实现基于学习的源代码漏洞自动检测提供了机遇.基于学习的漏洞检测方法是指使用基于机器学习或深度学习技术来进行漏洞检测的方法,其中基于深度学习的漏洞检测方法由于能够自动提取代码中漏洞相关的语法和语义特征,避免特征工程,在漏洞检测领域表现出了巨大的潜力,并成为近年来的研究热点.本文主要回顾和总结了现有的基于学习的源代码漏洞检测技术,对其研究和进展进行了系统的分析和综述,重点对漏洞数据挖掘与数据集构建、面向漏洞检测任务的程序表示方法、基于机器学习和深度学习的源代码漏洞检测方法、源代码漏洞检测的可解释方法、细粒度的源代码漏洞检测方法等五个方面的研究工作进行了系统的分析和总结.在此基础上,给出了一种结合层次化语义感知、多粒度漏洞分类和辅助漏洞理解的漏洞检测参考框架.最后对基于学习的源代码漏洞检测技术的未来研究方向进行了展望.

面向电力生产精细化风险解译的高度相似防护工具智能检测技术研究

电力生产通常面临高低电压、强弱电流等复杂工作环境转换,不同作业场景有严格的防护工具使用标准,因此,研究生产作业过程防护工具的精细辨识对保障人员安全及电网安全意义重大。已有研究可实现安全帽、工作服等基础着装类检测,而实际生产中存在大量形态高度相似的实体防护工具,如绝缘手套与线手套、绝缘杆与验电杆等。为此,该文提出一种基于深度代表性度量学习的相似防护工具智能检测方法。将目标类别特征学习转换为以差异化表达不同目标特征距离为目的的嵌入式空间特征学习,得到表征不同目标的深度代表性特征向量,通过计算未知目标与代表性特征向量的距离进行类别判断,最后以现场图像进行试验验证。试验结果表明:所提方法实现了对形态相似防护工具的特征差异表达和精准辨识,相比于常见目标检测模型具有更优越的辨识性能,从而提高电力生产安全风险辨识的精细化水平。