检索结果-内蒙古大学图书馆

Advancements in Electrical, Electronics, Communication, Computing and Automation (ICAECA), International conference on

作者： Salabh Shashank Neha Tyagi Department of Computer Science and Engineering Amity University Noida Uttar Pradesh India

ISBN: (数字)9798331525439

ISBN: (纸本)9798331525446

The year 2023 saw an enormous rise in patent applications, with around 9700 patents filed daily, resulting in around 3.55 million annually. While this number saw a spike from last year’s statistics, the problem is the classification of patents. Currently, the process is not fully automated. There is some human intervention, which leads to delays in the process. With the help of this paper, we presented an approach using the techniques of Machine Learning and natural language processing to classify these patents into their respective categories. In natural language processing, we used methods like stopword removal, lowercasing, and vectorization. In contrast, Machine Learning is used to apply various models and test their accuracy and other performance metrics. The models used are Random Forest, Decision Tree, Support Vector, Logistic Regression, and Naïve Bayes, while the performance metrics are accuracy, precision, recall, F1-score, true positives, true negatives, false positives, and false negatives. These metrics are computed for each model, and the results are compared accordingly. Random Forest has outperformed all the models in all performance parameters. It has an accuracy of 0.94, precision of 0.9410, recall of 0.94, and F1-score of 0.9390. It also offers what further steps can be taken to improve the model’s performance. This research effort contributes significantly to classifying patents based on their aspects.

关键词： Measurement Patents Logistic regression Accuracy Computational modeling natural language processing Vectors Telecommunication computing Random forests Text processing

来源：评论

学校读者我要写书评

暂无评论

Evaluating Large language Models: Challenges, Limitations, and Future Directions

Evaluating Large Language Models: Challenges, Limitations, a...

引用

International ECTI Northern Section conference on Electrical, Electronics, Computer and Telecommunications Engineering (ECTI-NCON)

作者： Li-Hua Li Agus Cahyo Nugroho Department of Information Management College of Informatics Chaoyang University of Technology Taichung Taiwan

ISBN: (数字)9798331543273

ISBN: (纸本)9798331543280

Large language Models (LLMs) have transformed natural language processing (NLP) by providing previously unheard-of capabilities in text production, translation, and comprehension. Large language Models have advanced rapidly. Nevertheless, there are several obstacles and restrictions associated with the implementation and assessment of these models. This study offers a thorough analysis of the situation of LLMs today, with an emphasis on the methods for assessing their effectiveness and significance. Key issues like prejudice, moral dilemmas, scalability, and transparency are noted, and the shortcomings of the current frameworks for evaluation are also covered. Our goal in examining these problems is to identify the areas that require improvement. In addition, we suggest possible direction for future investigation, stressing the importance of inclusive datasets, improved interpretability, and stronger assessment criteria. In addition to guiding the creation of more morally and effectively constructed AI systems, our mission is to promote a greater understanding of LLMs.

关键词： Measurement Ethics Translation Large language models Scalability Training data Production natural language processing Safety Testing

来源：评论

学校读者我要写书评

暂无评论

Adversarial natural language processing: Overview, Challenges and Future Directions 58

Adversarial Natural Language Processing: Overview, Challenge...

引用

58th Hawaii International conference on System Sciences, HICSS 2025

作者： Shaw, Laxmi Ansari, Mohammed Wasim Ekin, Tahir Texas State University United States

ISBN: (纸本)9780998133188

natural language processing (NLP) has gained wider utilization with the emergence of large language models. However, adversarial attacks threaten their reliability. We present an overview of adversarial NLP with an emphasis on challenges, emerging areas and future directions. First, we review attack methods and evaluate the vulnerabilities of popular NLP models. Then, we review defense strategies including adversarial training. We identify key trends and suggest future directions such as the use of Bayesian methods to improve the security and robustness of NLP systems. © 2025 IEEE Computer Society. All rights reserved.

关键词： Adversarial machine learning

来源：评论

学校读者我要写书评

暂无评论

AI Personalized Mental Health Monitoring System using Machine Learning, and natural language processing

AI Personalized Mental Health Monitoring System using Machin...

引用

Data Science, Agents & Artificial Intelligence (ICDSAAI), International conference on

作者： S. Mohamed Jamal Irfan G. Dhivya N. Raghavendran M. C. Babu T. Jasperline R. Saravanakumar Department of Computer Applications Karpagam Academy of Higher Education Coimbatore Tamil Nadu India Department of Artificial Intelligence and Date Science RMK College of Engineering and Technology Chennai Tamil Nadu India Artificial Intelligence and Data Sciences Chennai Institute of Technology Chennai Tamil Nadu India Department of Computer Science and Engineering Dr. G. U. Pope College of Engineering Tuticorin Tamil Nadu India Iconix Software Solution Tirunelveli Tamil Nadu India

ISBN: (数字)9798331537555

ISBN: (纸本)9798331537562

This study introduces an AI-driven customized mental health monitoring system that uses machine learning (ML) and natural language processing (NLP) to provide ongoing mental health evaluation and assistance. The system uses user-generated textual data, including journal entries, chat logs, and social media postings, to assess emotional states, stress levels, and mental health issues. The system uses sophisticated natural language processing methods to identify linguistic patterns associated with mental health problems such as depression, anxiety, and others. The machine learning algorithms customize monitoring by adjusting to the unique characteristics of each user and use past data to provide customized suggestions and preemptive alerts. The technology aims to enhance human interaction and function as a proactive tool in mental health treatment by providing real-time information. The system extracts emotional indicators from textual input using sentiment analysis methods and pre-trained natural language processing models such as BERT. The results are used with user demographics and history to build machine learning classifiers, such as Random Forest (RF) and Support Vector Machine (SVM), which predict mental health disorders. Supervised learning is used to develop the model, and suggestions are progressively improved using a tailored feedback system. The SVM and RF algorithms demonstrated an 88% and 92%accuracy rate in forecasting mood swings and early warning indicators of mental health disorders when evaluated on a set of anonymized mental health information. The 30% increase in user involvement demonstrated the efficacy of tailored suggestions in proactive mental health treatment.

关键词： Support vector machines Machine learning algorithms Accuracy Social networking (online) Supervised learning Mental health natural language processing Monitoring Random forests Classification tree analysis

来源：评论

学校读者我要写书评

暂无评论

LLM-Based Generation of Solidity Smart Contracts from System Requirements in natural language: The AstraKode Case

LLM-Based Generation of Solidity Smart Contracts from System...

引用

IEEE International conference on Software Analysis, Evolution and Reengineering (SANER)

作者： Gabriele De Vito Damiano D'Amici Fabiano Izzo Filomena Ferrucci Dario Di Nucci Software Engineering (SeSa) Lab - University of Salerno Salerno Italy AstraKode S.r.l. Pescara Italy

ISBN: (数字)9798331535100

ISBN: (纸本)9798331535117

As blockchain technology continues to evolve, the need for accessible solutions for developing smart contracts has grown, especially for non-technical users. This paper addresses practitioners' challenges in generating Solidity smart contracts from natural language requirements within the AstraKode Blockchain no-code platform (AKB). Our goal is to lower the barrier of entry into smart contract development, making it more accessible to users with limited technical expertise. We propose three methods, i.e., Naive Generation, Augmented Generation, and Enhanced Generation, each utilizing large language models to streamline the code generation process. These methods cater to different user needs, from rapid prototyping to handling complex business scenarios, improving accessibility and usability within AKB. We demonstrate their practical relevance, potential, and limitations in addressing real-world challenges in smart contract development through empirical evaluations and practitioner feedback. Thanks to collaboration with academia and effective knowledge transfer, these methods provide innovative solutions to the challenges of smart contract generation. Furthermore, they have been integrated into AKB to enhance user services, ultimately promoting the development and deployment of secure and efficient smart contracts in the industry.

关键词： Industries Codes Large language models Smart contracts natural languages Rapid prototyping Software Blockchains Usability Knowledge transfer

来源：评论

学校读者我要写书评

暂无评论

基于大语言模型的中文实体链接实证研究

引用

自动化学报 2025年第2期51卷 327-342页

作者：徐正斐辛欣北京理工大学计算机学院北京100081 北京理工大学北京市海量语言信息处理与云计算应用工程技术研究中心北京100081

近年来,大语言模型(Large language model,LLM)在自然语言处理中取得重大进展.在模型足够大时,大语言模型涌现出传统的预训练语言模型(Pre-trained language model,PLM)不具备的推理能力.为了探究如何将大语言模型的涌现能力应用于中文... 详细信息

近年来,大语言模型(Large language model,LLM)在自然语言处理中取得重大进展.在模型足够大时,大语言模型涌现出传统的预训练语言模型(Pre-trained language model,PLM)不具备的推理能力.为了探究如何将大语言模型的涌现能力应用于中文实体链接任务,适配了以下四种方法:知识增强、适配器微调、提示学习和语境学习(In-context learning,ICL).在Hansel和CLEEK数据集上的实证研究表明,基于Qwen-7B/ChatGLM3-6B的监督学习方法超过基于小模型的方法,在Hansel-FS数据集上提升3.9%~11.8%,在Hansel-ZS数据集上提升0.7%~4.1%,在CLEEK数据集上提升0.6%~3.7%.而当模型参数量达到720亿时,Qwen-72B的无监督方法实现与监督微调Qwen-7B相近的结果(-2.4%~+1.4%).此外,大语言模型Qwen在长尾实体场景下有明显的优势(11.8%),且随着参数量的增加,优势会更加明显(13.2%).对错误案例进行分析(以下简称错误分析)发现,实体粒度和实体类别相关错误占比较高,分别为36%和25%.这表明在实体链接任务中,准确划分实体边界以及正确判断实体类别是提高系统性能的关键.

关键词：实体链接大语言模型知识增强适配器微调提示学习语境学习

来源：评论

学校读者我要写书评

暂无评论

From Pixels to Voice: A Simple and Efficient End-to-End Spoken Image Description Approach via Vision Codec language Models

From Pixels to Voice: A Simple and Efficient End-to-End Spok...

引用

International conference on Acoustics, Speech, and Signal processing (ICASSP)

作者： Chung Tran Sakriani Sakti Graduate School of Science and Technology Nara Institute of Science and Technology Ikoma Japan

ISBN: (数字)9798350368741

ISBN: (纸本)9798350368758

Neural audio codecs provide a powerful tool for compressing audio signals into discrete codec representations. This compact discrete representation has made it possible to successfully apply a natural language processing (NLP) model to various audio and speech processing tasks, including text-to-speech (e.g., VALL-E, VALL-E X) and multimodal audio-text generation (e.g., LauraGPT, VioLA). While these models excel at handling sequential data like text and speech, their potential for processing non-sequential data, such as images, remains unexplored. In this paper, we introduce PixVoxLM, a simple and efficient end-to-end framework that combines vision-language models with neural audio codecs to tackle the Image-to-Speech (I2S) problem. Experiments on the Flickr8k dataset demonstrate that PixVoxLM delivers promising results compared to existing I2S methods. Furthermore, this research is the first to explore a new capability: visual-guided speech completion in I2S model, paving the way for new practical applications in everyday communication, such as speech prompt-based instruction.

关键词： Codecs Speech coding Signal processing natural language processing Data models Acoustics Text to speech Speech processing Synthetic aperture sonar

来源：评论

学校读者我要写书评

暂无评论

SACR: Self-training with Saliency-Augmented Consistency Regularization for Few-Shot Learners

SACR: Self-training with Saliency-Augmented Consistency Regu...

引用

International conference on Acoustics, Speech, and Signal processing (ICASSP)

作者： Yanyan Feng Yue Zhou Yun Xue Fenghuan Li Zehong Lin School of Electronics and Information Engineering South China Normal University Foshan China Department of Artiffcial Intelligence School of Informatics Xiamen University Xiamen China School of Computer Science and Technology Guangdong University of Technology Guangzhou China

ISBN: (数字)9798350368741

ISBN: (纸本)9798350368758

Pre-trained language models have made significant strides in natural language processing tasks, enabling flexible fine-tuning for downstream applications. However, in few-shot learning scenarios, pre-trained models face challenges related to overfitting due to limited training samples, which hinders their ability to capture data diversity and robustly handle input variations. To overcome these limitations, we propose the Saliency-Augmented Consistency Regularization (SACR) framework, a novel self-training strategy designed to improve few-shot learning performance and robustness. SACR consists of three key components: (1) Saliency-guided Data Perturbation, which uses saliency analysis to identify and perturb words that significantly influence model predictions, generating semantically consistent pseudo-samples; (2) Semantic-equivalent Sample Mining, which employs K-means clustering to select semantically similar samples and prevent semantic shift; and (3) Consistency Training, where regularization ensures consistency between the semantics and prediction distributions of the original and perturbed samples. We conducted extensive experiments on 15 public datasets to evaluate our approach. SACR significantly outperforms strong baseline models and demonstrates superior generalization capabilities compared to state-of-the-art methods, achieving an average 1.0% improvement in classification accuracy.

关键词： Training Adaptation models Sensitivity Perturbation methods Semantics Signal processing Data models natural language processing Few shot learning Speech processing

来源：评论

学校读者我要写书评

暂无评论

Larger language Models Don’t Care How You Think: Why Chain-of-Thought Prompting Fails in Subjective Tasks

Larger Language Models Don’t Care How You Think: Why Chain-...

引用

International conference on Acoustics, Speech, and Signal processing (ICASSP)

作者： Georgios Chochlakis Niyantha Maruthu Pandiyan Kristina Lerman Shrikanth Narayanan CS and ECE - School of Advanced Computing University of Southern California

ISBN: (数字)9798350368741

ISBN: (纸本)9798350368758

In-Context Learning (ICL) in Large language Models (LLM) has emerged as the dominant technique for performing natural language tasks, as it does not require updating the model parameters with gradient-based methods. ICL promises to "adapt" the LLM to perform the present task at a competitive or state-of-the-art level at a fraction of the computational cost. ICL can be augmented by incorporating the reasoning process to arrive at the final label explicitly in the prompt, a technique called Chain-of-Thought (CoT) prompting. However, recent work has found that ICL relies mostly on the retrieval of task priors and less so on "learning" to perform tasks, especially for complex subjective domains like emotion and morality, where priors ossify posterior predictions. In this work, we examine whether "enabling" reasoning also creates the same behavior in LLMs, wherein the format of CoT retrieves reasoning priors that remain relatively unchanged despite the evidence in the prompt. We find that, surprisingly, CoT indeed suffers from the same posterior collapse as ICL for larger language models. Code is available at https://***/gchochla/cot-priors.

关键词： Codes Large language models Computational modeling natural languages Signal processing Predictive models Benchmark testing Cognition Computational efficiency Speech processing

来源：评论

学校读者我要写书评

暂无评论

LLMs for Generation of Architectural Components: An Exploratory empirical Study in the Serverless World

LLMs for Generation of Architectural Components: An Explorat...

引用

Working IEEE/IFIP conference on Software Architecture

作者： Shrikara Arun Meghana Tedla Karthik Vaidhyanathan SERC IIIT Hyderabad India

ISBN: (数字)9798331520908

ISBN: (纸本)9798331520915

Recently, the exponential growth in capability and pervasiveness of Large language Models (LLMs) has led to significant work done in the field of code generation. However, this generation has been limited to code snippets. Going one step further, our desideratum is to automatically generate architectural components. This would not only speed up development time, but would also enable us to eventually completely skip the development phase, moving directly from design decisions to deployment. To this end, we conduct an exploratory study on the capability of LLMs to generate architectural components for Functions as a Service (FaaS), commonly known as serverless functions. The small size of their architectural components make this architectural style amenable for generation using current LLMs compared to other styles like monoliths and microservices. We perform the study by systematically selecting open source serverless repositories, masking a serverless function and utilizing state of the art LLMs provided with varying levels of context information about the overall system to generate the masked function. We evaluate correctness through existing tests present in the repositories and use metrics from the Software Engineering (SE) and natural language processing (NLP) domains to evaluate code quality and the degree of similarity between human and LLM generated code respectively. Along with our findings, we also present a discussion on the path forward for using GenAI in architectural component generation.

关键词： Measurement Codes Software architecture Large language models Microservice architectures natural language processing

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：