检索结果-内蒙古大学图书馆

Opportunities and Challenges of Generative Artificial Intelligence: Research, Education, Industry Engagement, and Social Impact

引用

IEEE INDUSTRIAL ELECTRONICS MAGAZINE 2025年第1期19卷 30-45页

作者： De Silva, Daswin Kaynak, Okyay El-Ayoubi, Mona Mills, Nishan Alahakoon, Damminda Manic, Milos La Trobe Univ Res Ctr Data Analyt & Cognit Melbourne Vic 3083 Australia Bogazici Univ Dept Elect & Elect Engn TR-34342 Istanbul Turkiye La Trobe Univ Educ Serv Dept Melbourne Vic 3083 Australia Virginia Commonwealth Univ Dept Comp Sci Richmond VA 23284 USA

Generative artificial intelligence (Generative AI) is transforming the way we live and work. Following several decades of artificial narrow intelligence, Generative AI is signaling a paradigm shift in the intelligence of machines, an increased generalization capability with increased accessibility and equity for nontechnical users. Large language models (LLMs) are leading this charge, specifically conversational interfaces, such as ChatGPT, Gemini, Claude, and Llama (large language model meta AI). Besides language and text, robust and effective Generative AI models have emerged for all other modalities of digital data, image, video, audio, code, and combinations thereof. This article presents the opportunities and challenges of Generative AI in advancing industrial systems and technologies. The article begins with an introduction to Generative AI, which includes its rapid progression to state-of-the-art, the deep learning algorithms, large training datasets, and computing infrastructure used to build Generative AI models, as well as the technical limitations. The contribution, value, and utility of Generative AI is presented in terms of its four capabilities of accelerating academic research, augmenting the learning and teaching experience, supporting industry practice, and increasing social impact. The article concludes with an expeditious message to the academic research and industry practitioner communities to invest time and effort in the training, adoption, and application of Generative AI, with consideration for AI literacy for all stakeholders, human-centricity, and the responsible development and use of AI in industrial settings.

关键词： Generative AI Artificial intelligence Training Data models computational modeling Solid modeling Regulation

来源：评论

学校读者我要写书评

暂无评论

Onboard Person Retrieval System With Model Compression: A Case Study on Nvidia Jetson Orin AGX

引用

IEEE ACCESS 2025年 13卷 8257-8269页

作者： Chaudhari, Jay N. Galiyawala, Hiren Sharma, Paawan Shukla, Pancham Raval, Mehul S. Ahmedabad Univ Sch Engn & Appl Sci Ahmadabad 380009 India RyDOT Infotech Pvt Ltd Ahmadabad 380027 India Pandit Deendayal Energy Univ Sch Technol Gandhinagar 382007 India Imperial Coll London London SW7 2AZ England

A person retrieval system (PRS) in video surveillance identifies an individual based on descriptive attributes, a task that employs several computationally intensive deep learning models. We implement and analyse a PRS for pre-recorded videos on a graphics processing unit (GPU) and Nvidia Jetson Orin AGX. This paper presents a new Person Attribute Recognition (PAR) architecture, CorPAR, using three backbone networks, ConvNext, ResNet-50, and EfficientNet-B0. It enhances the F1-score by 4.1% with ConvNeXT-Base, 1.63% with the ResNet, and by 8.07% with EfficientNet-B0, surpassing the performance of the state-of-the-art Weighted-PAR method. The proposed method uses model compression techniques like quantisation and pruning with L1 regularisation to assess their impact on person retrieval. The study reveals that the PRS utilising EfficientNet-B0, with 32-bit quantisation, achieves the best performance, delivering a throughput of 22 frames per second and a True Positive Rate of 71% on Nvidia Jetson Orin AGX matching the performance of a model implemented using GPU.

关键词： Clothing Image edge detection Surveillance Quantization (signal) Real-time systems Performance evaluation Image color analysis Graphics processing units Videos computational modeling Edge device model compression person attribute recognition person retrieval pruning quantization surveillance

来源：评论

学校读者我要写书评

暂无评论

Federated Learning With Sailfish-Optimized Ensemble Models for Anomaly Detection in IoT Edge Computing Environment

引用

IEEE ACCESS 2025年 13卷 53171-53187页

作者： Babu, Aravam Bagubali, A. Vellore Inst Technol Sch Elect Engn Vellore 632014 Tamil Nadu India

The Internet of Things (IoT) has transformed cyber-physical systems by enabling seamless connectivity and automation. However, IoT devices face resource constraints, making anomaly detection challenging. Traditional centralized approaches suffer from computational inefficiencies, increased latency, and privacy concerns, making them unsuitable for real-time anomaly detection in distributed IoT environments. To address these challenges, this paper proposes a privacy-preserving anomaly detection framework that integrates Federated Learning (FL) with an optimized Isolation Forest model. FL enables decentralized training on IoT devices, reducing the risk of data breaches. However, anomaly detection performance is often hindered by suboptimal parameter selection. To overcome this, the Sailfish Optimization Algorithm (SFO) is incorporated to fine-tune the Isolation Forest model's parameters dynamically, balancing exploration and exploitation. This optimization enhances accuracy while maintaining data confidentiality. Additionally, the framework is evaluated against leading FL-based and traditional anomaly detection models, including Local Outlier Factor (LOF), Generative Adversaria (GAN), and Variational autoencoder (VAE), demonstrating superior performance in recall and F1-score. Extensive experiments on benchmark datasets confirm that the proposed method achieves higher anomaly detection efficiency with a lower error rate than existing methods. The results establish this framework as a scalable, privacy-preserving, and computationally efficient solution for anomaly detection in IoT edge environments, addressing critical limitations in security, latency, and data privacy in real-world applications.

关键词： Internet of Things Anomaly detection Forestry computational modeling Edge computing Data models Real-time systems Data privacy Accuracy Security federated learning IoT environment isolation forest sailfish optimizer

来源：评论

学校读者我要写书评

暂无评论

Improved Lightweight Saliency Model Based on Neural Network for Noncooperative Spacecraft Detection

引用

IEEE TRANSACTIONS ON AEROSPACE AND ELECTRONIC SYSTEMS 2025年第1期61卷 642-654页

作者： Wang, Menghan Feng, Dongzhu Wang, Hui Guo, Hehe Dai, Pei Cui, Jiashan Wang, Xiaoming Xidian Univ Sch Aerosp Sci & Technol Xian 710126 Peoples R China Xidian Univ Key Lab Equipment Efficiency Extreme Environm Minist Educ Xian 710126 Peoples R China Xidian Univ Sch Aerosp Sci & Technol Xian 710126 Peoples R China Beijing Inst Spacecraft Syst Engn Beijing Peoples R China Beijing Inst Tracking & Telecommun Technol Beijing 100094 Peoples R China

With the expansion of space exploration and activities, the space environment becomes increasingly complex, and thus, the importance of noncooperative spacecraft detection (NCSD) becomes more significant for enhancing space situational awareness capabilities. However, the implementation of NCSD in space faces two challenges: the resource constraints of spaceborne embedded systems and the object scale variations caused by complex imaging environments in space. Accordingly, an improved lightweight saliency model is proposed in this article for NCSD in the space environment. In this model, a lightweight feature extractor that aims to extract multilevel multiscale features is designed as the backbone network using the in-layer multiscale block and the intrafrequency multiscale block. Integrity channel attention feature fusion modules are then applied to enhance the channel discrimination of multilevel features. Furthermore, a part-whole verification module is introduced to strengthen the learned integrity feature by measuring the consistency between the target parts and the complete target region. Extensive contrast experiments were conducted on the SwissCube dataset, and the results clearly demonstrate that the proposed lightweight saliency model achieves relatively ideal segmentation effect of noncooperative object spacecraft and competitive performance in terms of a wide range of metrics with only 1.99M parameters.

关键词： Feature extraction Space vehicles Object detection Convolutional neural networks computational modeling Space missions Satellites Lightweight feature extractor neural network noncooperative spacecraft detection (NCSD) saliency model space situational awareness

来源：评论

学校读者我要写书评

暂无评论

Performance Evaluation and Application Potential of Small Large Language Models in Complex Sentiment Analysis Tasks

引用

IEEE ACCESS 2025年 13卷 49007-49017页

作者： Yang, Yunchu Li, Jiaxuan Guo, Jielong Pang, Patrick Cheong-Iao Wang, Yapeng Yang, Xu Im, Sio-Kei Macao Polytech Univ Fac Appl Sci Macau Peoples R China Macao Polytech Univ Engn Res Ctr Appl Technol Machine Translat & Artif Minist Educ Macau Peoples R China

Sentiment analysis using Large Language Models (LLMs) has gained significant attention in recent research due to its outstanding performance and ability to understand complex texts. However, popular LLMs, such as ChatGPT, are typically closed-source and come with substantial API costs, posing challenges for resource-limited scenarios and raising concerns about privacy. To address this, our study evaluates the feasibility of using small LLMs (sLLMs) as alternatives to GPT for aspect-based sentiment analysis in Chinese healthcare reviews. We compared several Chinese sLLMs of varying sizes with GPT-3.5, using GPT-4o's results as the benchmark, and assessed their classification accuracy by computing F1 scores for each individual aspect as well as an overall F1 score. Additionally, we examined sLLMs' instruction-following capabilities, VRAM requirements, generation times, and the impact of temperature settings on the performance of top-performing sLLMs. The results demonstrate that several sLLMs can effectively follow instructions and even surpass GPT-3.5 in accuracy. For instance, InternLM2.5 achieved an F1 score of 0.85 with zero-shot prompting, while the smaller Qwen2.5-3B model performed well despite its minimal size. Prompt strategies significantly influenced smaller and older models like Qwen2.5-1.5B and ChatGLM3.5 but had limited impact on newer models. Temperature settings showed minimal effect, while older models generated responses faster, and newer models offered higher accuracy. This study underscores the potential of sLLMs as resource-efficient, privacy-preserving alternatives to closed-source LLMs in specialized domains. Our work demonstrates versatility, with potential applications across domains such as finance and education, and tasks like sentiment analysis, credit risk assessment, and learning behavior analysis, offering valuable insights for real-world use cases.

关键词： Sentiment analysis Analytical models Reviews Load modeling Large language models Data privacy Accuracy Performance evaluation Data models computational modeling Small large language model aspect-based sentiment analysis sentiment analysis natural language processing resource-constrained environments data privacy

来源：评论

学校读者我要写书评

暂无评论

An Improved Stochastic Configuration Networks With Compact Structure and Parameter Adaptation

引用

IEEE ACCESS 2025年 13卷 18141-18151页

作者： Li, Sanyi Guan, Hongyu Liu, Peng Yue, Weichao Wang, Qian Zhengzhou Univ Light Ind Sch Elect & Informat Engn Zhengzhou 450002 Peoples R China

Stochastic Configuration Networks (SCNs) perform well in machine learning and data mining tasks in complex data environments. However, traditional SCNs have limitations in network size and computation time. To address these issues, this paper proposes an improved version of SCNs. There are two key improvements: First, the stopping condition for generating neurons is optimized to improve the effectiveness of new neurons. Second, the regularization parameter r is adjusted dynamically to speed up the learning process. These improvements are trying to increase the efficiency of SCN construction, reduce the number of redundant neurons, and shorten the overall computation time. Experiments comparing this method with existing ones show that the proposed approach not only reduces network complexity but also effectively decreases training time. In addition, experimental results using baseline datasets and UCI databases show that the number of nodes required for oscn is reduced by approximately 50% and the computation time is reduced by approximately 40% compared to traditional algorithms.

关键词： Neurons Training Complexity theory Biological neural networks Adaptation models Vectors Stochastic processes computational modeling Accuracy Predictive models Stochastic configuration networks parameter optimization memory efficiency dynamic regularization training efficiency

来源：评论

学校读者我要写书评

暂无评论

Content Caching-Assisted Vehicular Edge Computing Using Multi-Agent Graph Attention Reinforcement Learning

引用

IEEE TRANSACTIONS ON VEHICULAR TECHNOLOGY 2025年第2期74卷 3509-3514页

作者： Shen, Jinjin Lin, Yan Zhang, Yijin Zhang, Weibin Shu, Feng Li, Jun Nanjing Univ Sci & Technol Sch Elect & Opt Engn Nanjing 210094 Peoples R China Hainan Univ Sch Informat & Commun Engn Haikou 570228 Peoples R China Southeast Univ Sch Informat Sci & Engn Nanjing 210096 Peoples R China

In order to avoid repeated task offloading and realize the reuse of popular task computing results, we construct a novel content caching-assisted vehicular edge computing (VEC) framework. In the face of irregular network topology and unknown environmental dynamics, we further propose a multi-agent graph attention reinforcement learning (MGARL) based edge caching scheme, which utilizes the graph attention convolution kernel to integrate the neighboring nodes' features of each agent and further enhance the cooperation among agents. Our simulation results show that our proposed scheme is capable of improving the utilization of caching resources while reducing the long-term task computing latency compared to the baselines.

关键词： Vehicle dynamics Network topology Edge computing computational modeling Reinforcement learning Decision making Topology Indexes Simulation Roads Edge caching graph attention reinforcement learning multi-agent vehicular edge computing

来源：评论

学校读者我要写书评

暂无评论

Mobility and Cost Aware Inference Accelerating Algorithm for Edge Intelligence

引用

IEEE TRANSACTIONS ON MOBILE COMPUTING 2025年第3期24卷 1530-1549页

作者： Yuan, Xin Li, Ning Wei, Kang Xu, Wenchao Chen, Quan Chen, Hao Guo, Song Harbin Inst Technol Sch Ocean Engn Harbin 150001 Peoples R China Harbin Inst Technol Sch Comp Sci & Technol Harbin 150001 Peoples R China TheHong Kong Polytech Univ Dept Comp Hong Kong 999077 Peoples R China Guangdong Univ Technol Sch Comp Sci & Technol Guangzhou 510006 Peoples R China Hong Kong Univ Sci & Technol Dept Comp Sci & Engn Hong Kong Peoples R China

The edge intelligence (EI) has been widely applied recently. Splitting the model between device, edge server, and cloud can significantly improve the performance of EI. The model segmentation without user mobility has been investigated in detail in previous studies. However, in most EI use cases, the end devices are mobile. Few studies have been conducted on this topic. These works still have many issues, such as ignoring the energy consumption of mobile device, inappropriate network assumption, and low effectiveness on adapting user mobility, etc. Therefore, to address the disadvantages of model segmentation and resource allocation in previous studies, we propose mobility and cost aware model segmentation and resource allocation algorithm for accelerating the inference at edge (MCSA). Specifically, in the scenario without user mobility, the loop iteration gradient descent (Li-GD) algorithm is provided. When the mobile user has a large model inference task that needs to be calculated, it will take the energy consumption of mobile user, the communication and computing resource renting cost, and the inference delay into account to find the optimal model segmentation and resource allocation strategy. In the scenario with user mobility, the mobility aware Li-GD (MLi-GD) algorithm is proposed to calculate the optimal strategy. Then, the properties of the proposed algorithms are investigated, including convergence, complexity, and approximation ratio. The experimental results demonstrate the effectiveness of the proposed algorithms.

关键词： Servers computational modeling Inference algorithms Resource management Delays Costs Energy consumption Complexity theory Partitioning algorithms Mobile handsets Edge intelligence model segmentation inference accelerating mobility

来源：评论

学校读者我要写书评

暂无评论

On-the-Fly Syntax Highlighting: Generalisation and Speed-Ups

引用

IEEE TRANSACTIONS ON SOFTWARE ENGINEERING 2025年第2期51卷 355-370页

作者： Palma, Marco Edoardo Wolf, Alex Salza, Pasquale Gall, Harald C. Univ Zurich Dept Informat CH-8006 Zurich Switzerland

On-the-fly syntax highlighting involves the rapid association of visual secondary notation with each character of a language derivation. This task has grown in importance due to the widespread use of online software development tools, which frequently display source code and heavily rely on efficient syntax highlighting mechanisms. In this context, resolvers must address three key demands: speed, accuracy, and development costs. Speed constraints are crucial for ensuring usability, providing responsive feedback for end users and minimizing system overhead. At the same time, precise syntax highlighting is essential for improving code comprehension. Achieving such accuracy, however, requires the ability to perform grammatical analysis, even in cases of varying correctness. Additionally, the development costs associated with supporting multiple programming languages pose a significant challenge. The technical challenges in balancing these three aspects explain why developers today experience significantly worse code syntax highlighting online compared to what they have locally. The current state-of-the-art relies on leveraging programming languages' original lexers and parsers to generate syntax highlighting oracles, which are used to train base Recurrent Neural Network models. However, questions of generalisation remain. This paper addresses this gap by extending previous work validation dataset to six mainstream programming languages thus providing a more thorough evaluation. In response to limitations related to evaluation performance and training costs, this work introduces a novel Convolutional Neural Network (CNN) based model, specifically designed to mitigate these issues. Furthermore, this work addresses an area previously unexplored performance gains when deploying such models on GPUs. The evaluation demonstrates that the new CNN-based implementation is significantly faster than existing state-of-the-art methods, while still delivering the same near-perfect ac

关键词： Codes Accuracy Syntactics Grammar computational modeling Java Convolutional neural networks Analytical models Source coding Predictive models Syntax highlighting neural networks deep learning regular expressions

来源：评论

学校读者我要写书评

暂无评论

CBKG-DTI: Multi-Level Knowledge Distillation and Biomedical Knowledge Graph for Drug-Target Interaction Prediction

引用

IEEE JOURNAL OF BIOMEDICAL AND HEALTH INFORMATICS 2025年第3期29卷 2284-2296页

作者： Zhao, Xiaosa Wang, Qixian Zhang, Ye He, Chenglong Yin, Minghao Zhao, Xiaowei Northeast Normal Univ Sch Informat Sci & Technol Changchun 130117 Peoples R China Northeast Normal Univ Sch Phys Changchun 130024 Peoples R China

The prediction of drug-target interactions (DTIs) has emerged as a vital step in drug discovery. Recently, biomedical knowledge graph enables the utilization of multi-omics resources for modelling complex biological systems and further improves overall performance of specific predictive task. However, due to the scale and generalization of biomedical knowledge graph, it is necessary to capture task-specific knowledge from biomedical knowledge graph for DTI prediction. Moreover, although biomedical knowledge graph has rich interactions between biological entities, there still needs to contain unignorable structural information of drugs or targets in the multi-modal fusion manner. To this end, we develop a novel DTI identification framework, CBKG-DTI, which aims to distill task-specific knowledge from the complex knowledge graph to the lightweight DTI prediction model. Specifically, CBKG-DTI first introduces a hierarchy-aware knowledge graph embedding as teacher model to capture semantic hierarchy information of biomedical knowledge graph. Then, to further improve model performance, CBKG-DTI integrates information from multiple aspects such as relational information and structural information by constructing a heterogeneous network and then employs a heterogeneous graph attention network framework as the lightweight student model. Moreover, we design a multi-level distillation mechanism to improve the representation and prediction ability of the lightweight student model via capturing the representation and logit distribution of the teacher model. Finally, we conduct the extensive comparison experiments and can reach the AUC of 0.9751 and the AUPR of 0.6310 under 5-fold cross validation. This not only demonstrates the superiority of CBKG-DTI in DTI prediction, but also, more importantly, validate the effectiveness of the framework capturing task-specific knowledge from biomedical knowledge graph.

关键词： Diffusion tensor imaging Biological system modeling Knowledge graphs Drugs Protein engineering Predictive models Semantics Feature extraction Data models computational modeling Drug-target interaction knowledge distillation knowledge graph embedding graph neural network

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：