检索结果-内蒙古大学图书馆

4th international conference on Sentiment Analysis and Deep Learning, ICSADL 2025

作者： Patel, Prachi Bhushanwar, Kush Patel, Hemlata Piet Parul University Computer Engineering Department Gujarat Vadodara India Piet Parul University Computer Science & Engineering Department Gujarat Vadodara India

ISBN: (纸本)9798331523923

Social media websites have provided rich contextual data but its misuse for criminal activities creates lot of problems. This paper seeks to solve the problem of criminal behavioral analysis on social media platforms using machine learning, deep learning and natural language processing (NLP) methodologies. The goal is to assess the effectiveness of classification algorithms such as SVM, Navier Bayes, Random Forest, Convolutional Neural Network, Recurrent Neural Network, and the natural language processing (NLP) methodology that include sentiment analysis, topic modelling and entity recognition for detecting irregularities. Specifically, this research study discusses about the recent progress in these methods while considering essential problems, including data accessibility, model flexibility, and the topic of ethics. © 2025 IEEE.

关键词： Adversarial machine learning

来源：评论

学校读者我要写书评

暂无评论

Efficient Detection of Toxic Prompts in Large language Models 24

Efficient Detection of Toxic Prompts in Large Language Model...

引用

39th ACM/IEEE international conference on Automated Software engineering (ASE)

作者： Liu, Yi Yu, Junzhe Sun, Huijia Shi, Ling Deng, Gelei Chen, Yuqi Liu, Yang Nanyang Technol Univ Singapore Singapore ShanghaiTech Univ Shanghai Peoples R China

ISBN: (数字)9798400712487

ISBN: (纸本)9798400712487

Large language models (LLMs) like ChatGPT and Gemini have significantly advanced natural language processing, enabling various applications such as chatbots and automated content generation. However, these models can be exploited by malicious individuals who craft toxic prompts to elicit harmful or unethical responses. These individuals often employ jailbreaking techniques to bypass safety mechanisms, highlighting the need for robust toxic prompt detection methods. Existing detection techniques, both blackbox and whitebox, face challenges related to the diversity of toxic prompts, scalability, and computational efficiency. In response, we propose TOXICDETECTOR, a lightweight greybox method designed to efficiently detect toxic prompts in LLMs. TOXICDETECTOR leverages LLMs to create toxic concept prompts, uses embedding vectors to form feature vectors, and employs a Multi-Layer Perceptron (MLP) classifier for prompt classification. Our evaluation on various versions of the LLama models, Gemma-2, and multiple datasets demonstrates that TOXICDETECTOR achieves a high accuracy of 96.39% and a low false positive rate of 2.00%, outperforming state-of-the-art methods. Additionally, ToxicDetector's processing time of 0.0780 seconds per prompt makes it highly suitable for real-time applications. TOXICDETECTOR achieves high accuracy, efficiency, and scalability, making it a practical method for toxic prompt detection in LLMs.

关键词： Accuracy Scalability Large language models Chatbots Feature extraction Vectors Real-time systems Safety Software engineering Glass box

来源：评论

学校读者我要写书评

暂无评论

A Survey and Analysis of Textual Content Based on Exploratory Data Analysis Technique and Opinion Analysis 4

A Survey and Analysis of Textual Content Based on Explorator...

引用

4th international conference on Computation, Automation and knowledge Management, ICCAKM 2023

作者： Puri, Atharva Agrawal, Gaurav Dukare, Akash Jawale, Madhuri Sanjivani College of Engineering Savitribai Phule Pune University Department of Information Technology Pune India

ISBN: (纸本)9798350393248

This research paper presents an Exploratory Data Analysis (EDA) technique, Artificial Intelligence (AI), and natural language processing (NLP) based approach for the analysis of textual content, and opinion analysis. Although technically difficult, the task is very beneficial. Using the natural language processing method, we addressed the problem of detecting violent content in text documents in this work. Researchers have used a variety of techniques, including machine learning techniques such as Naïve Bayes, Lexicon-based approach, and Encoding techniques and we have used datasets to compare the data. Therefore, it provides an opportunity for research to explore and recommend sophisticated techniques. For this paper, we examined a unique dataset. © 2023 IEEE.

关键词： Machine learning

来源：评论

学校读者我要写书评

暂无评论

COLLEGEBOT: Virtual Assistant System for Enquiry Using natural language processing 2

COLLEGEBOT: Virtual Assistant System for Enquiry Using Natur...

引用

2nd international conference on Intelligent Data Communication Technologies and Internet of Things, IDCIoT 2024

作者： Selvi, A. Mounika, V. Rubika, V. Uvadharanee, B. M.Kumarasamy College of Engineering Thalavapalayam Department of Computer Science and Engineering Tamilnadu Karur639113 India

ISBN: (纸本)9798350327533

The Artificial Intelligence (AI) and natural language processing (NLP) software or application helps to interact with people in a human-like fashion by delivering information, answering questions, completing tasks, and offering assistance in different ways is called a virtual assistant. These virtual assistants can be found in many different places, such as smart speakers, websites, apps, and smartphones. Virtual assistants are able to comprehend human language and provide responses. They process and interpret text or voice inputs using NLP algorithms, allowing users to interact with them in a conversational fashion. Therefore, this study analyse a virtual assistant for college inquiry, a sophisticated artificial intelligence program created to offer prospective college students and their families all-inclusive support during the college admissions process. It can comprehend user requests and respond to them in a conversational manner with the use of Artificial Intelligence (AI) and natural language processing (NLP). These virtual assistants provide individualised advice by providing answers that are specific to each user's questions, whether they relate to particular academic programmes, admission requirements, financial aid options, or campus amenities. They are particularly skilled at responding to frequently requested queries, ensuring prompt and accurate responses while also freeing up administrative staff to concentrate on more difficult responsibilities. This research study offers a platform for users to interact with chatbots that have undergone extensive machine learning training on datasets. Machine learning algorithms employ a more natural approach to computing rather than a logical one. The training data affects the outcomes. © 2024 IEEE.

关键词： College enquiry chatbot Machine learning natural language processing Text mining Virtual Assistant system

来源：评论

学校读者我要写书评

暂无评论

Design and Optimization of an Intelligent Question Answering System Integrates natural language processing and knowledge Graph

Design and Optimization of an Intelligent Question Answering...

引用

2024 international conference on Power, Electrical engineering, Electronics and Control, PEEEC 2024

作者： Chou, Shanliang Academic Affairs Office Jiangsu Vocational College of Tourism Jiangsu Yangzhou225100 China

ISBN: (纸本)9798350378917

With the rapid development of information technology, intelligent question answering system (QAS) has become an important tool for users to obtain information. This paper aims to design an intelligent QAS which integrates naturallanguage processing (NLP) and knowledge graph (KG), and optimize its performance. By effectively combining NLP with KG technology, the system can accurately understand the semantic information of users' questions and provide users with accurate and comprehensive answers with the support of KG. Firstly, this paper introduces the development background and technical challenges of intelligent QAS, especially the important role of NLP and KG in the system. Thenthe architecture design of the system is described in detail, including input layer, processing layer and output layer, and the main functions and technical realization of each layer are introduced. In the aspect of system optimization, this paper puts forward query pruning strategy based on heuristic rules, caching mechanism and semantic understanding ability improvement strategy based on BERT (Bidirectional Encoder Representations from Transformers) model, anddesigns a set of regular updating and maintenance mechanism of KG. Through the experimental verification of two public data sets, WikiQA and TREC QA, the system shows high accuracy and recall rate, which proves the reliability and effectiveness of the intelligent QAS designed in this paper in practical application. This research not only enriches the theoretical basis of intelligent QAS, but also provides new ideas and methods for practical application in related fields. © 2024 IEEE.

关键词： knowledge graph

来源：评论

学校读者我要写书评

暂无评论

Editorial to the special issue on JCDL 2022

引用

international JOURNAL ON DIGITAL LIBRARIES 2024年第2期25卷 237-240页

作者： Mayr, Philipp Hinze, Annika Schaer, Philipp GESIS Leibniz Inst Social Sci Knowledge Technol Social Sci KTS Cologne Germany Univ Waikato Comp & Math Sci Hamilton New Zealand TH Koln Univ Appl Sci Inst Informat Sci Cologne Germany

This special issue features the selected works of authors who have presented papers at the 2022 iteration of the Joint conference on Digital Libraries (JCDL) in Cologne, Germany. The motto of the conference was "Bridging Worlds" and was run as a fully hybrid event. Ten papers covering all aspects of Digital Libraries, namely natural language processing, Information Retrieval, User Behavior, Scholarly Communication, Classification, Information Extraction are included in this issue.

关键词： Digital libraries natural language processing Information retrieval User behavior Scholarly communication Classification Information extraction

来源：评论

学校读者我要写书评

暂无评论

Source Prompt: Coordinated pre-training of language Models on Diverse Corpora from Multiple Sources 24

Source Prompt: Coordinated pre-training of Language Models o...

引用

33rd ACM international conference on Information and knowledge Management (CIKM)

作者： Xu, Yipei Lu, Dakuan Liang, Jiaqing Zhao, Jin Wang, Xintao Wu, Hengkui Chen, Ken Liu, Liujiang Xin, Yingsi Liu, Xuepeng Xiao, Yanghua Li, Zhixu Fudan Univ Shanghai Peoples R China Super Symmetry Technol Beijing Peoples R China

ISBN: (纸本)9798400704369

Pre-trained language models (PLMs) have established the new paradigm in the field of NLP. For more powerful PLMs, one of the most popular and successful ways is to continuously scale up sizes of the models and the pre-training corpora. These large corpora, typically obtained by converging smaller ones from multiple sources, are thus growing increasingly diverse. However, colossal converged corpora don't always enhance PLMs' performance. In this paper, we identify the disadvantage of heterogeneous corpora from multiple sources for pre-training PLMs. Towards coordinated pre-training on diverse corpora, we further propose Source Prompt (SP), which explicitly prompt the model with the source of data at the pre-training and fine-tuning stages. Extensive experimental results show that pre-training PLMs with SP on diverse corpora significantly improves performance in various downstream tasks.

关键词： Large language Models Pre-training Data engineering

来源：评论

学校读者我要写书评

暂无评论

Historical Ink: 19th Century Latin American Spanish Newspaper Corpus with LLM OCR Correction 4

Historical Ink: 19th Century Latin American Spanish Newspape...

引用

4th international conference on natural language processing for Digital Humanities, NLP4DH 2024

作者： Manrique-Gómez, Laura Montes, Tony Rodríguez-Herrera, Arturo Manrique, Rubén History and Geography Department Universidad de los Andes D.C. Bogotá Colombia Systems and Computing Engineering Department Universidad de los Andes D.C. Bogotá Colombia Civil and Environmental Engineering Department Rice University HoustonTX United States

ISBN: (纸本)9798891761810

This paper presents two significant contributions: First, it introduces a novel dataset of 19th-century Latin American newspaper texts, addressing a critical gap in specialized corpora for historical and linguistic analysis in this ***, it develops a flexible framework that utilizes a Large language Model for OCR error correction and linguistic surface form detection in digitized *** semi-automated framework is adaptable to various contexts and datasets and is applied to the newly created dataset. © 2024 Association for Computational Linguistics.

关键词： Computational linguistics

来源：评论

学校读者我要写书评

暂无评论

NLP Based Video SparkNotes Assistant 4

NLP Based Video SparkNotes Assistant

引用

4th international conference on Soft Computing for Security Applications, ICSCSA 2024

作者： Bhilare, Amol Gaikwad, Namrata Kurapati, Aditya Dabade, Ronak Saklani, Daksh Vishwakarma Institute of Technology Department of Computer Engineering Maharashtra Pune India

ISBN: (纸本)9798331515720

This research presents a novel method of automatic video summarising and note-generating using natural language processing (NLP) and audio recognition techniques. The exponential rise in internet video footage has increased demand for effective summarising methods. The proposed approach uses speech recognition to transcribe audio content, which is further processed with natural language processing (NLP) algorithms to provide succinct summaries. Furthermore, the system creates notes based on the main ideas that are taken from the transcript of the movie. The spoken words in the video are converted to text using voice recognition technology. After that, this study interpret and condense this material using specialized computer techniques, making it more concise yet nonetheless useful. Furthermore, our technology creates notes that highlight significant scenes or topics from the video. © 2024 IEEE.

关键词： natural language processing systems

来源：评论

学校读者我要写书评

暂无评论

A natural language processing Approach to the Translation of Speech into Indian Sign language 2

A Natural Language Processing Approach to the Translation of...

引用

2nd international conference on Distributed Computing and Optimization Techniques, ICDCOT 2024

作者： Vaishnavi, Katta Akbar, Mohammad Fazlul Balaji, Tedla Reddy, Kesavarapu Vivek Rani, Maddu Sudha Department of Information Technology Seshadri Rao Gudlavalleru Engineering College Andhra Pradesh India

ISBN: (纸本)9798350382952

Communication is a critical aspect of every individual's interaction, and individuals typically exchange information in a variety of languages. However, individuals with hearing and speech impairments may encounter difficulties in communicating on par with the general population. Additionally, hearing-impaired individuals comprehend our thoughts through the use of sign language. Research indicates that acquiring sign language skills facilitates the comprehension of lip-reading as well as one's mother tongue. This is accurate for both literate and illiterate individuals throughout India. The approach involves converting speech into written text, which is going to go through text preprocessing utilizing natural language processing (NLP) techniques to enhance analysis. Text and voice input are accepted by the system, which compares these with the clips in the authors' database. It displays matching sign motions based on Indian Sign language grammatical rules as a result if they match;if not, it proceeds through the lemmatization and tokenization stages. natural language processing, which powers part-of-speech tagging, parsing, lemmatization, and tokenization, is the system's central component. An accuracy of 92% is achieved in the proposed strategy. © 2024 IEEE.

关键词： Audition

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：