检索结果-内蒙古大学图书馆

7th international conference on Artificial Intelligence and Security, ICAIS 2021

作者： Wu, Mingli Du, Chunlai Huang, Yafei Cui, Xianwei Duan, Jianyong School of Information Science and Technology North China University of Technology Beijing100144 China

ISBN: (纸本)9783030786144

With the economic development, loan business has rapidly developed in China. the risk that customers can’t repay their loans on time has increased. therefore it is an important problem for financial organizations to approve the customers’ loan application or not. Typical machine learning methods for classification can be employed to mine customers’ financial information and give valuable judgments. However, these learning methods rely on shallow features, and the relationships between these features are not well studied. We investigate the function of Convolutional Neural Network (CNN) in this work, as it is successful in field of image recognition, speech recognition and natural language processing. We investigate four different CNN models. Experiments show that the fourth model with stochastic gradient descent algorithm and momentum achieves the best performance. Its accuracy and recall are 0.95 and 0.26 respectively. © 2021, Springer Nature Switzerland AG.

关键词： Classification (of information)

来源：评论

学校读者我要写书评

暂无评论

SesameBERT: Attention for Anywhere 7

SesameBERT: Attention for Anywhere

引用

7th IEEE international conference on Data Science and advanced Analytics (DSAA)

作者： Su, Ta-Chun Cheng, Hsiang-Chih Cathay Financial Holdings Lab Taipei Taiwan

ISBN: (纸本)9781728182063

Fine-tuning with pre-trained models has achieved exceptional results for many language tasks. In this study, we focused on one such self-attention network model, namely BERT, which has performed well in terms of stacking layers across diverse language-understanding benchmarks. However, in many downstream tasks, information between layers is ignored by BERT for fine-tuning. In addition, although self-attention networks are well-known for their ability to capture global dependencies, room for improvement remains in terms of emphasizing the importance of local contexts. In light of these advantages and disadvantages, this paper proposes SesameBERT, a generalized fine-tuning method that (1) enables the extraction of global information among all layers through Squeeze and Excitation and (2) enriches local information by capturing neighboring contexts via Gaussian blurring. Furthermore, we demonstrated the effectiveness of our approach in the HANS dataset, which is used to determine whether models have adopted shallow heuristics instead of learning underlying generalizations. the experiments revealed that SesameBERT outperformed BERT with respect to GLUE benchmark and the HANS evaluation set.

关键词： Natural language processing Self-Attention

来源：评论

学校读者我要写书评

暂无评论

Engage Learn: An AI-Based English Proficiency Improviser

Engage Learn: An AI-Based English Proficiency Improviser

引用

Ubiquitous Computing and Intelligent information Systems (ICUIS), international conference on

作者： Balasaheb Jadhav Manas Jain Adhip Jajoo Devika Kadam Harshvardhan Kadam Toshish Kakkad Department of Artificial Intelligence and Data Science Vishwakarma Institute of Technology Pune India

ISBN: (数字)9798331529635

ISBN: (纸本)9798331529642

In this research, the new and highly effective web-based English learning and teaching platform called Engage Learn is introduced with the main purpose of avoiding the mistakes of the traditional models and to help users gain the ability to write and speak English more effectively. Engage Learn employs a marked-defined approach, that involves the application of several advanced technologies. Generative models including ChatGPT 3.5 turbo and Gemini 1.5 flash get enhanced presence when trained using Retrieval-Augmented Generation (RAG) technique. First, the ability for users to speak English makes for an almost informal procedure when learning since the user and the computer are communicating with one another by use of words. In addition to this, targeted feedback modules make an assessment of the user's input and give an immediate-feedback of what is required in terms of improving a user's vocabulary, grammar, and fluency. It offers another learning approach since it presents the content in a manner that is suited more to the current ability level of the user and the usage of preferred learning objectives. Engage Learn also provides flexibility of use based on learning modes as encompassed by general conversation practice, focused grammar correction and vocabulary buildup. this paper reviews the system architecture, explaining features of the different parts of the system, including the speech recognition, LLM with RAG, output analyzing, and the feedback-giving modules on grammar, vocabulary, and reading preferences. Further the paper discusses the tuning of the LLMs in a carefully selected set for language acquisition. this fine-tuning makes sure that the LLMs shall have adequate capability and knowledge required so as to get a feel about how spoken English is and about the particular feedbacks that different user may require out of them. In the last section of the abstract, the planned evaluation methods are presented with the focus on the responsibility of user te

关键词： Vocabulary Reviews Computational modeling Systems architecture Speech recognition Ubiquitous computing Grammar Speech processing Tuning Testing

来源：评论

学校读者我要写书评

暂无评论

Interpretation and automatic integration of geospatial data into the Semantic web Towards a process of automatic geospatial data interpretation, classification and integration using semantic technologies

引用

COMPUTING 2020年第2期102卷 365-391页

作者： Prudhomme, Claire Homburg, Timo Ponciano, Jean-Jacques Boochs, Frank Cruz, Christophe Roxin, Ana-Maria Mainz Univ Appl Sci Lucy Hillebrand Str 2 D-55128 Mainz Germany Univ Bourgogne Franche Comte LIB EA 7534 Batiment i3M Rue Sully F-21000 Dijon France

In the context of disaster management, geospatial information plays a crucial role in the decision-making process to protect and save the population. Gathering a maximum of information from different sources to oversee the current situation is a complex task due to the diversity of data formats and structures. Although several approaches have been designed to integrate data from different sources into an ontology, they mainly require background knowledge of the data. However, non-standard data set schema (NSDS) of relational geospatial data retrieved from e.g. web feature services are not always documented. this lack of background knowledge is a major challenge for automatic semantic data integration. Focusing on this problem, this article presents an automatic approach for geospatial data integration in NSDS. this approach does a schema mapping according to the result of an ontology matching corresponding to a semantic interpretation process. this process is based on geocoding and natural language processing. this article extends work done in a previous publication by an improved unit detection algorithm, data quality and provenance enrichments, the detection of feature clusters. It also presents an improved evaluation process to better assess the performance of this approach compared to a manually created ontology. these experiments have shown the automatic approach obtains an error of semantic interpretation around 10% according to a manual approach.

关键词： Semantic interpretation Data quality Natural language processing Ontologies Spatial fusion Semantic web

来源：评论

学校读者我要写书评

暂无评论

Auto-Generating Business Process Model from Heterogeneous Documents: A Comprehensive Literature Survey 9

Auto-Generating Business Process Model from Heterogeneous Do...

引用

9th international conference on Electrical Engineering, Computer Science and Informatics, EECSI 2022

作者： Indahyanti, Uce Djunaidy, Arif Siahaan, Daniel Institut Teknologi Sepuluh Nopember Dept. of Informatics Surabaya Indonesia Universitas Muhammadiyah Sidoarjo Faculty of Science and Technology Indonesia Institut Teknologi Sepuluh Nopember Dept. of Information Systems Surabaya Indonesia

ISBN: (数字)9786239213565

ISBN: (纸本)9781665470438

Automatically generated business processes can be sourced from documents in the form of structured text or natural language. Although there have been several studies discussing generating business processes, there is still no research using systematically reviews as the source documents. this study presents a systematic literature review on document sources, methods, and challenges in generating business processes. We conducted a systematic literature review published from 2017 to early 2022 and identified 24 main studies discussing the sources of documents in generating business processes. We formulated and applied inclusion and exclusion criteria in two stages to determine the most relevant studies for our research goal. this literature review found that the most frequently used document sources were textual business rules (structured rules), using case diagrams (14 main studies), event logs (7 studies), and natural language text (3 studies) including customer feedback. In the aspect of the method, the most widely used method is in the field is natural language processing (NLP), followed by other methods such as semantic knowledge engineering (SKE), fuzzy, graph-based, deep learning, and the combination of NLP with deep learning. Meanwhile, the challenges faced in generating business processes include text preprocessing or document extraction, integration of business rules with business processes, and challenges in the form of time interval constraints, activity sequences, dummy activities, or invisible tasks that are generally found in event logs. © 2022 Institute of advanced Engineering and Science (IAES).

关键词： auto-generating business process document sources type systematic review

来源：评论

学校读者我要写书评

暂无评论

Predictive Analytics of Detection of Speech Disfluency Using GAN 7th

Predictive Analytics of Detection of Speech Disfluency Using...

引用

7th international conference on Soft Computing and Signal processing, ICSCSP 2024

作者： Ravi Kondal, E. Kakulapati, V. Lahari, Karumanchi Srijanya, Singirikonda Sanjana, Kasarla Department of Information Technology Sreenidhi Institute of Science and Technology Yamnampet Ghatkesar Telangana Hyderabad501301 India

ISBN: (纸本)9789819609239

Stuttering, along with other speech disfluencies, affects millions of people worldwide, with a large proportion experiencing mild stutters during heated conversations. this speech condition interrupts communication fluency by causing involuntary repetitions, elongations, or quiet pauses. It often appears in childhood, affecting roughly 2.5% of children under the age of five. Extensive studies in Western nations, notably in British and American English environments, have investigated the frequency of such problems among youngsters. this research also concentrated on identifying and comprehending voice disfluencies within the context of these linguistic changes. India has one of the world's largest Telugu-speaking populations, with almost 75 million people living in Andhra Pradesh and Telangana. Given this large population, investigating differences within the Telugu language is critical. To address volume restrictions in datasets, data augmentation techniques are used, which try to artificially enlarge the training set by modifying existing data. WaveGAN, or waveform generative adversarial network, is a well-known augmentation approach. Using WaveGAN, more training data can be generated, improving the accuracy of audio classification tasks. Various classification techniques, such as logistic regression (LR), random forest (RF), decision tree (DT), K-nearest neighbors (KNN), AdaBoost, and XGBoost, were used to classify audio samples as fluent or diffluent. Furthermore, various assessment measures were used to assess the performance of these machine learning algorithms. © the Author(s), under exclusive license to Springer Nature Singapore Pte Ltd. 2025.

关键词： k-nearest neighbors

来源：评论

学校读者我要写书评

暂无评论

DEX2web – A web-Based Software Implementing the Multiple-Criteria Decision-Making Method DEX 7th

DEX2Web – A Web-Based Software Implementing the Multiple-Cr...

引用

7th international conference on Decision Support System technology, ICDSST 2021

作者： Kikaj, Adem Bohanec, Marko Jožef Stefan International Postgraduate School Jamova 39 Ljubljana1000 Slovenia Department of Knowledge Technologies Jožef Stefan Institute Jamova 39 Ljubljana1000 Slovenia

ISBN: (纸本)9783030739751

DEX2web is an online suite of tools to help individuals and groups with their decision-making. DEX2web implements the qualitative multiple-criteria decision-modelling method DEX. DEX is useful for supporting complex decision-making tasks, where there is a need to select a particular option from a set of possible ones to satisfy the goals of the decision-maker. DEX2web primarily supports interactive development and evaluation of DEX models. Most of the functionality of the first available version of DEX2web is inherited from its desktop ancestor DEXi: development of DEX model structure, editing of attributes and their scales, definition of decision rules, multi-attribute evaluation and analysis of alternatives, and presenting evaluation results with charts. DEX2web has a modern software architecture and employs a newly developed DEX software library. DEX2web is freely available on https://***/. © 2021, Springer Nature Switzerland AG.

关键词： Decision making

来源：评论

学校读者我要写书评

暂无评论

Code Mixed information Retrieval for Gujarati Script News Articles 7th

Code Mixed Information Retrieval for Gujarati Script News Ar...

引用

Proceedings of the 7th international conference on Advances in Computing and Data Sciences, ICACDS 2023

作者： Joshi, Payal Joshi, Dhaval Department of Information and Communication Technology Veer Narmad South Gujarat University Gujarat Surat India

ISBN: (纸本)9783031379390

Large amount of digitized content is being created and made available online in many Indian languages. Searching from such content is difficult because of lack of regional language input tools and interface. Further documents may also contain commonly used English words written using Indian script which we refer as code mixed content. One of the easier methods of query input could be typing query’s words using Roman script mixing words of Indian language and English in single script which we refer as code mixed queries. For various Indian languages work is done on single language (code) transliterated search but code mixed information retrieval is very less explored area. Gujarati is an Indian regional language spoken by 55 million people in the world, but very less work is done for code mixed information retrieval and it lacks transliteration pair corpus. To search code mixed contents we need to identifying language of a word and we have to back transliterate it from Roman to Indian language script. Moreover certain words belong to both the languages. All these issues are major challenges in code mixed IR especially for Indian languages. A novel hybrid approach is proposed in this paper for Gujarati words transliteration which combines the use of Roman script Gujarati Soundex based match with alternative use of machine learning based improved CRF with post processing. A huge transliteration pair corpus of Gujarati script is created for words of both languages. Analysis of approaches for language identification and transliteration is presented in this paper along with compared retrieval result for different transliteration approaches and obtained an overall increase of 17% in MAP as compared to other approaches. © 2023, the Author(s), under exclusive license to Springer Nature Switzerland AG.

关键词： information retrieval

来源：评论

学校读者我要写书评

暂无评论

Sarcasm detection in product reviews using textual entailment approach 23

Sarcasm detection in product reviews using textual entailmen...

引用

Proceedings of the 2023 7th international conference on Natural language processing and information Retrieval

作者： Shubham Sinha Tarun Vijeta Pratik Kishor Kubde Ayush Pradeep Gajbhiye Mansi Anup Radke Christopher Jones Department of Computer Science and Engineering Visvesvaraya National Institute of Technology India School of Computer Science and Informatics Cardiff University UK

ISBN: (纸本)9798400709227

Sarcasm is a form of sentiment characterized by the use of words that express the opposite of what is meant. Sarcasm detection has applications in multiple domains ranging from sentiment analysis in product reviews to user feedback, and online forums. Sarcasm detection is important to understand user opinions and intentions in areas such as sentiment-based classification and opinion mining. this can result in better product development and customer service. Sarcasm detection can be a challenging task because sarcastic sentences may use positive expressions to convey negative meanings or may use negative sentences to convey positive meanings. Also, sarcastic sentences form a very small component of the entire communication. the increasing use of sarcasm in various social media such as Twitter, Reddit, Amazon product reviews, etc. has highlighted the importance of detecting and understanding sarcasm in various contexts. Sarcasm detection is a challenging problem for NLP systems that often rely on statistical models for performing sentiment analysis. In this research, the focus is on the use of a textual entailment approach for detecting sarcasm. Textual entailment is a natural language inference task that involves determining whether one text (hypothesis) can be derived from another text (premise). the underlying assumption behind this approach is that - if there is a contradiction between the premise and hypothesis, we can say that the hypothesis is sarcastic. To test our approach, an annotated corpus of 3000 product reviews was developed methodically from the Amazon Reviews dataset and tested using the textual entailment approach. the proposed approach achieved an F1 score of 0.76 on this dataset. the result is better than the baseline considered which is the BERT binary classifier which gives an F1 score of 0.48 on the same dataset.

关键词： Hypothesis Natural language Inference Premise Sarcasm Detection Textual Entailment

来源：评论

学校读者我要写书评

暂无评论

Development of Intelligent Telegram Chatbot Using Natural language processing

Development of Intelligent Telegram Chatbot Using Natural La...

引用

international conference on Wireless and Telematics (ICWT)

作者： Teddy Surya Gunawan Asaad Balla Falelmula Babiker Nanang Ismail Mufid Ridlo Effendi International Islamic University Malaysia Kuala Lumpur Malaysia UIN Sunan Gunung Djati Bandung Indonesia

ISBN: (纸本)9781665444033

Intelligent chatbots have been gaining interest in the past years due to the advance of artificial intelligence algorithms. As a result, many studies have been conducted on emotional transition and dialog structures. One of the benefits is on the medical applications, in which psychological assessment, clinical counseling, autism diagnostics, and advanced cognitive models could be provided. On the other hand, chatbot knowledge generally comes from a web-based information repository, in which the information is reliable, but it is rather not versatile as it does not contain emotions. the objective of this paper is to develop an intelligent chatbot using natural language processing and Telegram API. First, text processing using Telegram API on Python was developed. Next, emotion recognition was performed on the recorded chats. the appropriate response is then sent to the user. Results showed that our Telegram chatbot could interact smoothly with the users and identify the user's emotions.

关键词： Wireless communication Knowledge engineering Employee welfare Emotion recognition Psychology Medical services Chatbots

来源：评论

学校读者我要写书评

暂无评论

建议与咨询留下您的常用邮箱和电话号码，以便我们向您反馈解决方案和替代方法

时间限定

文献类型

馆藏选择

核心期刊

语言

文献类型

帮助

文字说明：

检索规则说明：

检索范例：

分类表

所选分类

限定检索结果

文献类型

馆藏范围

日期分布

学科分类号

主题

机构

作者

语言

请选择保存的检索档案：

请选择收藏分类：

通借通还

建议与咨询 留下您的常用邮箱和电话号码，以便我们向您反馈解决方案和替代方法

时间限定

文献类型

馆藏选择

核心期刊

语言

文献类型

帮助

文字说明：

检索规则说明：

检索范例：

分类表

所选分类

限定检索结果

文献类型

馆藏范围

日期分布

学科分类号

主题

机构

作者

语言

请选择保存的检索档案： 新增检索档案 确定 取消

请选择收藏分类： 新增自定义分类 确定 取消

通借通还

建议与咨询留下您的常用邮箱和电话号码，以便我们向您反馈解决方案和替代方法

请选择保存的检索档案：

请选择收藏分类：