Explainable Artificial Intelligence(XAI)Model for Cancer Image Classification

The use of Explainable Artificial Intelligence(XAI)models becomes increasingly important for making decisions in smart healthcare environments.It is to make sure that decisions are based on trustworthy algorithms and that healthcare workers understand the decisions made by these algorithms.These models can potentially enhance interpretability and explainability in decision-making processes that rely on artificial intelligence.Nevertheless,the intricate nature of the healthcare field necessitates the utilization of sophisticated models to classify cancer images.This research presents an advanced investigation of XAI models to classify cancer images.It describes the different levels of explainability and interpretability associated with XAI models and the challenges faced in deploying them in healthcare applications.In addition,this study proposes a novel framework for cancer image classification that incorporates XAI models with deep learning and advanced medical imaging techniques.The proposed model integrates several techniques,including end-to-end explainable evaluation,rule-based explanation,and useradaptive explanation.The proposed XAI reaches 97.72%accuracy,90.72%precision,93.72%recall,96.72%F1-score,9.55%FDR,9.66%FOR,and 91.18%DOR.It will discuss the potential applications of the proposed XAI models in the smart healthcare environment.It will help ensure trust and accountability in AI-based decisions,which is essential for achieving a safe and reliable smart healthcare environment.

A Study on the Explainability of Thyroid Cancer Prediction:SHAP Values and Association-Rule Based Feature Integration Framework

In the era of advanced machine learning techniques,the development of accurate predictive models for complex medical conditions,such as thyroid cancer,has shown remarkable progress.Accurate predictivemodels for thyroid cancer enhance early detection,improve resource allocation,and reduce overtreatment.However,the widespread adoption of these models in clinical practice demands predictive performance along with interpretability and transparency.This paper proposes a novel association-rule based feature-integratedmachine learning model which shows better classification and prediction accuracy than present state-of-the-artmodels.Our study also focuses on the application of SHapley Additive exPlanations(SHAP)values as a powerful tool for explaining thyroid cancer prediction models.In the proposed method,the association-rule based feature integration framework identifies frequently occurring attribute combinations in the dataset.The original dataset is used in trainingmachine learning models,and further used in generating SHAP values fromthesemodels.In the next phase,the dataset is integrated with the dominant feature sets identified through association-rule based analysis.This new integrated dataset is used in re-training the machine learning models.The new SHAP values generated from these models help in validating the contributions of feature sets in predicting malignancy.The conventional machine learning models lack interpretability,which can hinder their integration into clinical decision-making systems.In this study,the SHAP values are introduced along with association-rule based feature integration as a comprehensive framework for understanding the contributions of feature sets inmodelling the predictions.The study discusses the importance of reliable predictive models for early diagnosis of thyroid cancer,and a validation framework of explainability.The proposed model shows an accuracy of 93.48%.Performance metrics such as precision,recall,F1-score,and the area under the receiver operating characteristic(AUROC)are also higher than the baseline models.The results of the proposed model help us identify the dominant feature sets that impact thyroid cancer classification and prediction.The features{calcification}and{shape}consistently emerged as the top-ranked features associated with thyroid malignancy,in both association-rule based interestingnessmetric values and SHAPmethods.The paper highlights the potential of the rule-based integrated models with SHAP in bridging the gap between the machine learning predictions and the interpretability of this prediction which is required for real-world medical applications.

英语近义词缀的语义倾向研究及可视化呈现——以后缀-proof和-resistant在COHA和GloWbE语料库中的用法为例

本研究主要基于COHA和GloWbE语料库,使用语言动态图和共变词项分析的方法,对与近义后缀-proof和-resistant连用的名词词干的频数分布及语义倾向进行探究。研究结果表明:N-proof的名词词干涵盖了bullet、water、fire等各类与人们的日常生活息息相关的名词,该构式后面主要接服饰类名词;而N-resistant的名词词干主要为drug、antibiotic、methicillin等生物医药类名词,该构式后面主要接细菌类和材料类名词。在频数分布方面,在过去的二三十年间,N-resistant在美国英语中的使用频数显著上升,大有未来逐渐取代N-proof之势。在澳大利亚英语、新西兰英语等其他英语变体中,由于受到地理环境和社会文化因素的影响,与后缀-proof和-resistant连用的名词词干有时会呈现出鲜明的地域特征。就相互竞争的后缀而言,本研究表明类符阻碍和形符阻碍这两种机制很有可能同时存在。

A Study on the Corpus-driven Teaching and Learning of English Near Synonyms

Corpus-based studies show that near-synonyms differ in their collocational behavior and semantic prosody which causes deviation in Chinese EFL learners’English using.This paper attempts to investigate the effectiveness of a corpus-based approach to the teaching and learning of near synonyms and their collocations.By using internet accessible computers to log on to websites of BNC and COCA,learners can view real language data and study the usage of targeted near synonyms.This empirical study tends to show that a corpus based approach is positively effective to the learning of near synonyms.It also provides some pedagogical implications that learners can play a major and active role in DDL and teachers only act as assistants.

The Differnces Beween English Synonyms

ＴｈｅＤｉｆｆｅｒｎｃｅｓＢｅｗｅｅｎＥｎｇｌｉｓｈＳｙｎｏｎｙｍｓＤｅｎｇＹａｎｊｕａｎ（ＤｅｐａｒｔｍｅｎｔｏｆＦｏｒｅｉｇｎＬａｎｇｕａｇｅ，ＫｕｎｍｉｎｇＩｎｓｔｉｔｕｔｅｏｆＴｅｃｈｎｏｌｏｇｙ，Ｋｕｎｍｉｎｇ６５００９３）ＴｈｅＤｉｆｆｅ...

Enhancing Embedding-Based Chinese Word Similarity Evaluation with Concepts and Synonyms Knowledge

Word similarity(WS)is a fundamental and critical task in natural language processing.Existing approaches to WS are mainly to calculate the similarity or relatedness of word pairs based on word embedding obtained by massive and high-quality corpus.However,it may suffer from poor performance for insufficient corpus in some specific fields,and cannot capture rich semantic and sentimental information.To address these above problems,we propose an enhancing embedding-based word similarity evaluation with character-word concepts and synonyms knowledge,namely EWS-CS model,which can provide extra semantic information to enhance word similarity evaluation.The core of our approach contains knowledge encoder and word encoder.In knowledge encoder,we incorporate the semantic knowledge extracted from knowledge resources,including character-word concepts,synonyms and sentiment lexicons,to obtain knowledge representation.Word encoder is to learn enhancing embedding-based word representation from pre-trained model and knowledge representation based on similarity task.Finally,compared with baseline models,the experiments on four similarity evaluation datasets validate the effectiveness of our EWS-CS model in WS task.

THE PETROLOGIC CHARACTERISTICS AND THE FORMATION ENVIRONMENT OF THE PALEOKARST BRECCIAS IN THE LOWER REACHES OF THE QINGJIANG RIVER,WESTERN HUBEI

In the Cambrian-Ordoviclan strata in the lower reaches of the Qingjiang River (western Hubei province), many well-consolidated paleokarst breccia bodies of different sizes and occurrences have been observed. By studying them in detail and by introducing new ideas of modern sedimentology into the study, the authors described and classified these paleokarst breccias. Their forming conditions and mechanisms are also explained on the basis of our research

A Corpus-Based Approach to the Vocabulary Teaching and Learning--Implications From the Corpus-Based Collocational Research on Two Pairs of Synonyms

Languages embody a lot of words that are considered as synonyms, and people just take it for granted that such words are identical in meaning without any discrimination. However, a corpus-based approach to the study of the collocational behavior of the two frequently-used pairs of synonyms （selection and option, ill and sick） reveals significant discrepancies in the use of these two pairs of synonyms by Chinese English learners and native speakers. According to the analysis, the major problems lie in the current ways of vocabulary teaching and learning. This paper aims to highlight the important role of the corpus-based collocational research in English vocabulary teaching and learning. In the end, some suggestions concerning vocabulary teaching and leaming are put forward on the basis of corpus-based research.

可解释人工智能在电力系统中的应用综述与展望

可解释人工智能(XAI)作为新型人工智能(AI)技术,具有呈现AI过程逻辑、揭示AI黑箱知识、提高AI结果可信程度的能力。XAI与电力系统的深度耦合将加速AI技术在电力系统的落地应用,在人机交互的过程中为电力系统的安全、稳定提供助力。文中梳理了电力系统XAI的历史脉络、发展需求及热点技术,总结了XAI在源荷预测、运行控制、故障诊断、电力市场等方面的电力应用,并围绕解释含义、迭代框架、数模融合等方面展望了电力系统XAI的应用前景,可为推动电力系统智能化转型与人机交互迭代提供理论参考与实践思路。

Type-2 Diabetes Mellitus and Glucagon-Like Peptide-1 Receptor toward Predicting Possible Association

Aim: This study aimed to investigate the effect of non-synonymous SNPs (nsSNPs) of the Glucagon-like peptide-1 Receptor (GLP-1R) gene in protein function and structure using different computational software. Introduction: The GLP1R gene provides the necessary instruction for the synthesis of the insulin hormones which is needed for glucose catabolism. Polymorphisms in this gene are associated with diabetes. The protein is an important drug target for the treatment of type-2 diabetes and stroke. Material and Methods: Different nsSNPs and protein-related sequences were obtained from NCBI and ExPASY database. Gene associations and interactions were predicted using GeneMANIA software. Deleterious and damaging effects of nsSNPs were analyzed using SIFT, Provean, and Polyphen-2. The association of the nsSNPs with the disease was predicted using SNPs & GO software. Protein stability was investigated using I-Mutant and MUpro software. The structural and functional impact of point mutations was predicted using Project Hope software. Project Hope analyzes the mutations according to their size, charge, hydrophobicity, and conservancy. Results: The GLP1R gene was found to have an association with 20 other different genes. Among the most important ones is the GCG (glucagon) gene which is also a trans membrane protein. Overall 7229 variants were seen, and the missense variants or nsSNPs (146) were selected for further analysis. The total number of nsSNPs obtained in this study was 146. After being subjected to SIFT software (27 Deleterious and 119 Tolerated) were predicted. Analysis with Provean showed that (20 deleterious and 7 neutral). Analysis using Polyphen-2 revealed 17 probably damaging, 2 possibly damaging and 1 benign nsSNPs. Using two additional software SNPs & GO and PHD-SNPs showed that 14 and 17 nsSNPs had a disease effect, respectively. Project Hope software predicts the effect of the 14 nsSNPs on the protein function due to differences in charge, size, hydrophobicity, and conservancy between the wild and mutant types. Conclusion: In this study, the 14 nsSNPs which were highly affected the protein function. This protein is providing the necessary instruction for the synthesis of the insulin hormones which is needed for glucose catabolism. Polymorphisms in this gene are associated with diabetes and also affect the treatment of diabetic patients due to the fact that the protein acts as an important drug target.

A Comparative Study of the Usage of Synonyms Based on BNC-Taking Assessment and Evaluation as Examples

The discrimination of synonyms has always been one of the great challenges for English learners.Taking assessment and evaluation as examples,this study analyses the similarities and differences of the two words,as well as their usage from the perspectives of frequency,stylistics,collocation and semantic prosody with the help of British National Corpus,and demonstrates the importance of corpus retrieval tools in synonyms discrimination.Furthermore,this paper will give some suggestions for English learners and teachers in English vocabulary teaching.

Mysteries of Solar System Explained by WUM

E. Stone in the article “18 Mysteries and Unanswered Questions About Our Solar System. Little Astronomy” wrote: One of the great things about astronomy is that there is still so much out there for us to discover. There are so many unanswered questions and mysteries about the universe. There is always a puzzle to solve and that is part of beauty. Even in our own neighborhood, the Solar System, there are many questions we still have not been able to answer [1]. In the present paper, we explain the majority of these Mysteries and some other unexplained phenomena in the Solar System (SS) in frames of the developed Hypersphere World-Universe Model (WUM) [2].

交互失误情境下交互式人工智能拟人化的负面影响——基于ChatGPT和搜索引擎的实验证据

[目的/意义]对人机交互失误情境下用户产生的情绪和态度展开研究,旨在验证人工智能拟人化理论的负面效应并探索交互式人工智能拟人化产生负面影响的缓解机制。[方法/过程]以人机交互失误为情境,基于期望一致性理论和拟人化理论,通过实验法分别从交互的结果失误和调用失误两个角度,探索了交互式人工智能不同的拟人化程度对用户产生负面态度的影响机制。[结果/结论]研究发现,人机交互失误时,高拟人化的交互式人工智能会加剧用户对其的厌恶感和负面态度。验证了算法可解释性能够缓解厌恶感和负面态度。综合上述分析,从人机交互健康长远发展角度,提出了人工智能应采用有限拟人化的设计框架和具有可解释性的算法架构等建议。

基于累积置信规则库推理的台风灾害直接经济损失预测

针对台风灾害直接经济损失预测问题,现有的解决方法大多是基于时间序列或评估数据的预测模型,忽略了在建模过程中对历史数据的应用和模型的可解释性。鉴于此,该文将扩展置信规则库模型(EBRB)应用于台风灾害直接经济损失预测,并针对可能存在规则过量和组合爆炸问题,提出基于聚类方法与证据推理(ER)相结合的累积置信规则库(C-BRB)台风灾害经济损失预测模型。最后基于收集到的台风灾害数据进行直接经济损失预测,并通过与已有研究成果进行比较,验证基于C-BRB的台风灾害直接经济损失预测模型的有效性和可行性。

《仪礼》饮食相关类动词

《仪礼》中表示吃喝及与之关系密切的动词包括吃、喝、烹饪、盛放、舀取、置放、切割、除去等类,共有61个动词,其中超过100次的有4词,50—100次之间的有2词,30—50次之间的有8词,20—30次之间的有5词,10—20次之间的有6词,10次以下的有36词,虽然个别词出现频次很高,但总体出现频次并不高。词义方面,各类大都既有用于一般意义的通用词,又有特定含义的特殊词,前者代表了《仪礼》与其他文献的一致性,后者体现了《仪礼》该类动词的特殊性。与上古汉语其他文献相比,《仪礼》该类动词一方面表现出与其他文献的一致性,主要是各义场核心词高度相同。一方面也展现出一定的独特性,一是一些动词在其他文献中不见或很少见;二是一些同义聚合成员少于其他文献,显示了其用词的局限性。

基于视觉Transformer飞行员姿态估计

人体姿态估计是行为感知领域中的一个重要环节,也是民用飞机驾驶舱智能交互方式的一项关键技术。为建立民用飞机驾驶舱复杂光照环境与飞行员姿态估计模型性能的可解释联系,提出基于视觉Transformer飞行员姿态(ViTPPose)估计模型,该模型在卷积神经网络(CNN)主干网络末端使用包含多层编码层的双支路Transformer模块,编码层联合Transformer和空洞卷积,在增大感受野的同时捕捉后期高阶特征的全局相关性。基于飞行机组标准操作程序,建立飞行模拟场景下的飞行员操纵行为关键点检测数据集,ViTPPose估计模型在此数据集上完成飞行员坐姿估计,并通过与基准模型对比,验证了其有效性。在驾驶舱复杂光照的背景下,构建坐姿估计热图,分析ViTPPose估计模型对光照强度的偏好,测试其在不同光照等级下的性能,揭示其对不同光照强度的依赖关系。

语义通信的数学理论

自从1948年经典信息论诞生以来,在其指导下,现代通信技术已经逼近了理论性能极限,例如信息熵H(U)、信道容量C=max_(p(x))I(X;Y)以及率失真函数R(D)=min_(p(x|x):Ed(x,x)≤D)I(X;X)。长期以来,由于经典信息论只研究语法信息,限制了通信科学的进一步发展。近年来,研究语义信息处理与传输的通信技术获得了学术界的普遍关注,语义通信开辟了未来通信技术发展的新方向,但还缺乏一般性的数学指导理论。为了解决这一难题,构建了语义信息论的理论框架,对语义信息的度量体系与语义通信的理论极限进行了系统性阐述。首先,通过深入分析各类信源的数据特征,以及各种下游任务的需求,总结归纳出语义信息的普遍属性——同义性。由此指出语义信息是语法信息的上级概念,是许多等效或相似语法信息的抽象特征,表征隐藏在数据或消息背后的含义或内容。将语义信息与语法信息之间的关系命名为同义映射,这是一种“一对多”映射,即一个语义符号可以由许多不同的语法符号表示。基于同义映射f这一核心概念,引入语义熵H_(s)(U)作为语义信息的基本度量指标,表示为信源概率分布与同义映射的泛函。在此基础上,引入上/下语义互信息I^(s)(X;Y)(I_(s)(X;Y)),语义信道容量C_(s)=max_(f_(xy))max_(p_((x)))I^(s)(X;Y)以及语义率失真函数R_(s)(D)=min_({f_(x),f_(x)})min_(p(x|x):Ed_(s)(x,x)≤D)I_(s)(X;X),从而构建了完整的语义信息度量体系。这些语义信息度量是经典信息度量的自然延伸,都由同义映射约束,如果采用“一对一”映射,则可以退化为传统的信息度量。由此可见,语义信息度量体系包含语法信息度量,前者与后者具有兼容性。其次,证明了3个重要的语义编码定理,以揭示语义通信的性能优势。基于同义映射,引入新的数学工具——语义渐近均分(AEP),详细探讨了同义典型序列的数学性质,并应用随机编码和同义典型序列译码/编码,证明了语义无失真信源编码定理、语义信道编码定理和语义限失真信源编码定理。类似于经典信息论,这些基本编码定理也都是存在性定理,但它们指出了语义通信系统的性能极限,在语义信息论中起着关键作用。由同义映射和这些基本编码定理可以推断,语义通信系统的性能优于经典通信系统,即语义熵小于信息熵H_(s)(U)≤H(U),语义信道容量大于经典信道容量C_(s)≥C,以及语义率失真函数小于经典率失真函数R_(s)(D)≤R(D)。最后,讨论了连续条件下的语义信息度量。此时,同义映射转换为连续随机变量分布区间的划分方式。相应地,划分后的子区间被命名为同义区间,其平均长度定义为同义长度S。特别是对于限带高斯信道,得到了一个新的信道容量公式C_(s)=B log[S^(4)(1+P/N_(0)B)],其中,平均同义长度S表征了信息的辨识能力。这一容量公式是经典信道容量的重要扩展,当S=1时,该公式退化为著名的香农信道容量公式。综上所述,语义信息论依据同义映射这一语义信息的本质特征,构建了语义信息的度量体系,引入新的数学工具,证明了语义编码的基本定理,论证了语义通信系统的性能极限,揭示了未来语义通信的巨大性能潜力。

可信学习分析构建:风险桎梏、理论纾解与实践解蔽

可信学习分析技术的发展已成为全球研究者共同关注的焦点。国内外学者已就其内涵意蕴、价值逻辑等开展广泛探讨,但也存在关键问题不明、理论边界不清和实践路径模糊等问题。为此,研究以关键问题提炼和理论框架构建为先导,以实践路径的探索和实施为主线,对可信学习分析技术的教育应用进行分析。在实践问题方面,研究分析了阻碍可信学习分析实现的关键因素,包括数据漂移现象所导致的算法失准、样本分布失衡现象导致的算法偏见以及算法黑盒现象导致的决策结果失信等一系列问题。在理论研究方面,研究结合理论与实践问题分析,建立起面向学习环境、数据治理、建模分析和干预反馈四要素的可信学习分析理论框架。随后,研究依托该理论框架,结合相关研究方法和技术,构建了可信学习分析的实践路径,如结合价值敏感设计的伦理调适等。研究为可信学习分析研究从理论到实践提供系统化的解决思路。

基于人在回路的纵向联邦学习模型可解释性研究

纵向联邦学习(vertical federated learning,VFL)常用于高风险场景中的跨领域数据共享,用户需要理解并信任模型决策以推动模型应用。现有研究主要关注VFL中可解释性与隐私之间的权衡,未充分满足用户对模型建立信任及调优的需求。为此,提出了一种基于人在回路(human-in-the-loop,HITL)的纵向联邦学习解释方法(explainable vertical federated learning based on human-in-the-loop,XVFL-HITL),通过构建分布式HITL结构将用户反馈纳入VFL的基于Shapley值的解释方法中,利用各参与方的知识校正训练数据来提高模型性能。进一步,考虑到隐私问题,基于Shapley值的可加性原理,将非当前参与方的特征贡献值整合为一个整体展示,从而有效保护了各参与方的特征隐私。实验结果表明,在基准数据上,XVFL-HITL的解释结果具有有效性,并保护了用户的特征隐私;同时,XVFL-HITL对比VFL-Random和直接使用SHAP的VFL-Shapley进行特征选择的方法,模型准确率分别提高了约14%和11%。

融合深度神经网络和方面感知的可解释推荐方法

为提升推荐结果的准确性和可解释性,提出一种融合深度神经网络和方面感知的可解释推荐方法。针对评分数据的稀疏性问题,综合考虑显式和隐式评分数据,通过深度神经网络的矩阵分解模型学习用户和物品的潜在特征;通过无监督的方面提取模块来学习用户和物品的方面特征;将潜在特征和方面特征统一到预测层进行评分预测;针对生成解释质量低且缺乏个性化的问题,在评分预测的基础上,采用提取的主题词和预定义的神经模板相结合生成推荐理由,提高解释的生成质量。实验表明,提出的方法不仅能准确预测用户对物品的评分,还能够生成具有解释性的推荐理由,且生成的解释质量优于对比方法。