检索结果-内蒙古大学图书馆

Improving Machine Translation Formality with Large Language Models

computers, Materials & Continua 2025年第2期82卷 2061-2075页

作者： Murun Yang Fuxue Li School of Computer Science and Engineering Northeastern UniversityShenyang110819China College of Electrical Engineering Yingkou Institute of TechnologyYingkou115014China

Preserving formal style in neural machine translation (NMT) is essential, yet often overlooked as an optimization objective of the training processes. This oversight can lead to translations that, though accurate, lack formality. In this paper, we propose how to improve NMT formality with large language models (LLMs), which combines the style transfer and evaluation capabilities of an LLM and the high-quality translation generation ability of NMT models to improve NMT formality. The proposed method (namely INMTF) encompasses two approaches. The first involves a revision approach using an LLM to revise the NMT-generated translation, ensuring a formal translation style. The second approach employs an LLM as a reward model for scoring translation formality, and then uses reinforcement learning algorithms to fine-tune the NMT model to maximize the reward score, thereby enhancing the formality of the generated translations. Considering the substantial parameter size of LLMs, we also explore methods to reduce the computational cost of INMTF. Experimental results demonstrate that INMTF significantly outperforms baselines in terms of translation formality and translation quality, with an improvement of +9.19 style accuracy points in the German-to-English task and +2.16 COMET score in the Russian-to-English task. Furthermore, our work demonstrates the potential of integrating LLMs within NMT frameworks to bridge the gap between NMT outputs and the formality required in various real-world translation scenarios.

关键词： Neural machine translation formality large language model text style transfer style evaluation reinforcement learning

来源：评论

学校读者我要写书评

暂无评论

Unguided Machine Learning-Based Computation Offloading for Near-Memory Processing 38

Unguided Machine Learning-Based Computation Offloading for N...

引用

38th International Conference on VLSI Design, VLSID 2025

作者： Maity, Satanu Ghose, Manojit Kumar, Avinash Chakraborty, Anol Chakraborty, Ankit Indian Institute of Information Technology Department of Computer Science and Engineering Assam Guwahati India Jorhat Engineering College Department of Computer Science and Engineering Assam India

ISBN: (纸本)9798331522445

Recent data-intensive applications encounter memory wall bottlenecks in the traditional processor-centric computing architecture due to the need for frequent and extensive off-chip data movement. The emerging 3D-enabled near-memory processing (NMP) is a potential solution where computation can be offloaded and performed near memory, eliminating the need for expansive off-chip data transfer. Recently, some crucial works in this context have been proposed that use a supervised machine learning approach that requires a well-labeled dataset, which necessitates creating a time-consuming simulation process. Furthermore, their major focus was on profiling and offloading the entire application to the NMP side, causing a sub-optimal performance because of the low processing power on the NMP cores. This paper presents an unguided machine-learning-based approach (UCO) for computation offloading in NMP architecture that offloads only the memory-intensive regions of an application. The approach extracts several application performance characteristics and creates two clusters using them - one for the CPU side and another for the NMP side. The proposed approach's effectiveness is demonstrated through a series of comprehensive experiments with standard simulators and dataintensive benchmark applications. The results indicate that UCO improves performance (1.31x and 1.10x), saves energy (16% and 12%), and reduces off-chip data transfer (78% and 73%) compared to conventional and state-of-the-art approaches. © 2025 IEEE.

关键词： Unsupervised learning

来源：评论

学校读者我要写书评

暂无评论

SAMFNet: Scene-aware sampling and multi-stage fusion for multimodal 3D object detection

引用

Alexandria engineering Journal 2025年 126卷 90-104页

作者： Wang, Baotong Xia, Chenxing Gao, Xiuju Ge, Bin Li, Kuan-Ching Fang, Xianjin Zhang, Yan Yang, Yuan College of Computer Science and Engineering Anhui University of Science and Technology Huainan232001 China College of Electrical and Information Engineering Anhui University of Science and Technology Huainan232001 China Department of Computer Science and Information Engineering Providence University Taichung43301 China The School of Electronics and Information Engineering Anhui University Hefei230039 China School of Mathematics and Big Data Anhui University of Science and Technology Huainan232001 China

Recently, multimodal 3D object detection (M3OD) that fuses the complementary information from LiDAR data and RGB images has gained significant attention. However, the inherent structural differences between point clouds and images pose fusion challenges, significantly hindering the exploration of correlations within multimodal data. To address this issue, this paper introduces an enhanced multimodal 3D object detection framework (SAMFNet), which leverages virtual point clouds generated from depth completion. Specifically, we design a scene-aware sampling module (SASM) that employs tailored sampling strategies for different bins based on the density distribution of point clouds. This effectively alleviates the detection bias problem while ensuring the key information of virtual points, significantly reducing the computational cost. In addition, we introduce a multi-stage feature fusion module (MSFFM) that embeds point-level and regional-adaptive feature fusion strategies to generate more informative multimodal features by fusing features with different granularities. To further improve the accuracy of model detection, we also introduce a confidence prediction branch unit (CPBU), which improves the detection accuracy by predicting the confidence of feature classification in the intermediate stage. Extensive experiments on the challenging KITTI dataset demonstrate the validity of our model. © 2025 The Authors

关键词： Image fusion

来源：评论

学校读者我要写书评

暂无评论

Thermal-field boosting continuous white light generation in lanthanide doped Y_(2)SiO_(5)phosphor

引用

Journal of Rare Earths 2025年第4期43卷 669-675,I0001页

作者： Qiongyao Han Jianhong Wu Yu Hou Bining Tian Xiaofeng Liu Chunli Guo Yue Tian Institute of New Carbon Materials College of Materials Science and EngineeringTaiyuan University of TechnologyTaiyuan030024China College of Electronic Information and Optical Engineering Taiyuan University of TechnologyTaiyuan030024China College of Materials Science and Engineering Taiyuan University of TechnologyTaiyuan030024China School of Materials Science and Engineering Zhejiang UniversityHangzhou310027China

The continuous white light(CWL)covering the visible and near-infrared(NIR)regions can be observed in various absorptive media excited by continuous-wave(CW)*** is valuable to stimulate more efforts in unravelling the involved photophysical processes and exploring its potential applications in diverse ***,we proved that the enhanced thermal-field can boost the CWL *** rare earth(RE)ions(Pr^(3+),Er^(3+)and Yb^(3+))as the photothermally active centers in Y_(2)SiO_(5)phosphor,we reveal that absorbing more excitation energy and isolating the heat conduction can lead to rapid thermal field accumulation inside the material,thereby significantly reducing the excitation threshold and enhancing white light *** results might have important implications for the understanding of thermally enhanced radiation and may facilitate the CWL commercial application in night vision,bioimaging,and non-destructive detection.

关键词： Continuous white light Rare earths Upconversion Thermally enhanced radiation

来源：评论

学校读者我要写书评

暂无评论

Prediction of banana leaf disease by intelligent algorithm through digital image using neural network technique

引用

Multimedia Tools and Applications 2025年 1-17页

作者： Nagu, Bharathiraja Kaur, Gaganpreet Shanmugavelu, Murugesan M.R.M, Veeramanickam Chitkara University Institute of Engineering and Technology Chitkara University Punjab India Department of Information Technology Tagore Engineering College Tamil Nadu Chennai India Apex Institute of Technology-Computer Science and Engineering Chandigarh University Punjab India

A lot of research shows that there could be several reasons why the duality of agricultural products has been reduced. Plant diseases make up one of the most important components of this quality. Therefore, the reduction of plant diseases allows for much higher product quality. The automated diagnosis of plant diseases using CNN techniques is presented in this article. The simple approach alters the 25 color channels and 24H histograms. The Banana Leaf Disease Detection Technique (BLDDT), which uses two attention mechanisms to assess, identify and categorize traits, was also suggested in this research. In the analysis phase, the testing carried out using the reference information reveals that the disease, detection system may be reasonably more appropriate than the current techniques because it both identifies and diagnoses the diseased/infected areas. The researchers were able to identify Esca, black decay, and isariopsis ––––with a 99.93% accuracy level using the suggested disease detector. © The Author(s), under exclusive licence to Springer science+Business Media, LLC, part of Springer Nature 2024.

关键词： Convolutional neural networks

来源：评论

学校读者我要写书评

暂无评论

An enhanced framework for smart automated evaluations of answer scripts using NLP and deep learning methods

引用

Multimedia Tools and Applications 2025年第11期84卷 8491-8513页

作者： G, Mohanraj R.K, Nadesh M, Marimuthu V, Sathiyapriya School of Computer Science Engineering and Information Systems Vellore Institute of Technology Vellore632014 India School of Computer Science and Engineering Vellore Institute of Technology Chennai600127 India Knowledge Institute of Technology Salem637504 India

The manual process of evaluating answer scripts is strenuous. Evaluators use the answer key to assess the answers in the answer scripts. Advancements in technology and the introduction of new learning paradigms need automation of the evaluation process. This work aims to develop an enhanced novel hybrid framework that can evaluate answer scripts and automatically assign marks for different type of questions based on keywords, grammar, symbols, special keywords, and the given factors. First, the proposed system uses Optical Character Recognition (OCR) to convert image answer scripts into an editable text format. Second, the sentence transformers, the Natural Language Processing (NLP) technique flips the answer script and answers key texts into word embedding vectors. To find similarity measures, these vectors are matched using BERT encoding, spearmanś rank-order correlation, and fuzzy search. At last, the proposed model is trained using Deep Columnar Convolutional Neural Network (DCCNN) in the third step with MINST and Kaggle handwritten mathematical symbols and tested with the segmented mathematical equations to find the similarity. The performance of proposed model is measured using precision, recall, accuracy, and F1-score, and its gives highest accuracy of 93% and 95% when compared to the existing methodologies. © The Author(s), under exclusive licence to Springer science+Business Media, LLC, part of Springer Nature 2024.

关键词： Optical character recognition

来源：评论

学校读者我要写书评

暂无评论

Movable-Antenna-Assisted Covert Communications With Reconfigurable Intelligent Surfaces

引用

IEEE Internet of Things Journal 2025年第9期12卷 12369-12382页

作者： Xie, Wenwu Li, Zilong Yu, Chao Xu, Hongbo Wang, Ji Wu, Weimin Li, Xingwang Yang, Liang Hunan Institute of Science and Technology School of Information Science and Engineering Yueyang414006 China Central China Normal University College of Physical Science and Technology Department of Electronics and Information Engineering Wuhan430079 China Huazhong University of Science and Technology School of Electronic Information and Communications Wuhan430074 China Henan Polytechnic University School of Physics and Electronic Information Engineering Jiaozuo454003 China Hunan University College of Computer Science and Electronic Engineering Changsha410082 China

This article proposes a novel covert communication framework utilizing movable antennas (MAs) to enable covert communications in which the evading detection eavesdropper aided by a reconfigurable intelligent surface (RIS). The trajectories of the MAs over the entire time slot, transmit beamforming, and the phase shift of the RIS in each time slot are jointly optimized to improve the covert rate. Specifically, the movement trajectories of the MAs are modeled as a Markov decision process (MDP), optimized by developing a novel deep reinforcement learning (DRL) approach. Furthermore, an alternating optimization (AO) algorithm is designed to jointly optimize the beamforming and phase. In particular, penalty-based two-layer iterative algorithm is proposed to guarantee that the solution satisfies the rank-one constraints. Numerical results show that the proposed MA-assisted covert communications system significantly outperforms conventional fixed-position antenna (FPA) schemes in terms of covert rate. © 2014 IEEE.

关键词： Markov processes

来源：评论

学校读者我要写书评

暂无评论

Decentralized multi-agent federated and reinforcement learning for smart water management and disaster response

引用

Alexandria engineering Journal 2025年 126卷 8-29页

作者： Mancy, H. Ghannam, Naglaa E. Abozeid, Amr Taloba, Ahmed I. Department of Computer Science College of Engineering and Computer Sciences Prince Sattam Bin Abdulaziz University Al-kharj11942 Saudi Arabia Department of Computer Engineering and Information College of Engineering in Wadi Alddawasir Prince Sattam Bin Abdulaziz University Wadi Alddawasir Saudi Arabia Department of Computer Science College of Computer and Information Sciences Jouf University Saudi Arabia Department of Mathematics Faculty of Science Al-Azhar University Cario Egypt Information System Department Faculty of Computers and Information Assiut University Egypt

Water resource management and disaster response have become some of the most challenging tasks, especially when disasters pose a threat, as delays could lead to more impacts. The centralized system used for water dynamics and disaster control usually presents itself as a scalability problem since more clients present a problem, the system's latency is high, and the system is always prone to a single-point failure. The previous approach lacks flexibility and does not synchronously guarantee the integration of several subjects in real time, especially during unpredictable disaster conditions. The proposed FL-MAPPO model surpasses current methods by facilitating decentralized, privacy-protecting decision-making minimizing latency and single-point failures. In contrast to LSTM, Bi-LSTM, and DRNN, which are based on centralized data processing, FL-MAPPO provides real-time adaptability and effective resource management. Experimental results validate that it has lower MSE, higher R² scores, and quicker response times, making it better suited for flood prediction and disaster response. To this end, this study advances a solution through a Decentralized Learning-Driven Multi-Agent Autonomous System (DL-MAAS). The new feature is a Decentralized Cooperation environment in which intelligent and self-managing agents learn utilizing Reinforcement Learning (RL) and Federated Learning (FL) algorithms for enhancing smart water management and real-time disaster relief. IoT devices are adopted for sensing and data acquisition, adaptive learning for decision-making, and optimization of energy use among the agents in the system through metaheuristic algorithms. The research methodology for implementing the proposed solution involves the design of a multi-layered architecture, including data acquisition, decentralized learning, and real-time execution. With a Mean Squared Error (MSE) of 0.112, R-squared (R²) of 0.953, and Mean Absolute Error (MAE) of 0.207, the proposed method is better

关键词： Multi agent systems

来源：评论

学校读者我要写书评

暂无评论

Deep transfer learning technique to detect white blood cell classification in regular clinical practice using histopathological images

引用

Multimedia Tools and Applications 2025年第9期84卷 5699-5723页

作者： Davamani, K. Anita Jawahar, Malathy Anbarasi, L. Jani Ravi, Vinayakumar Al Mazroa, Alanoud Robin, C. R. Rene Computer Science and Engineering Sathyabama Institute of Science and Technology Chennai India Leather Process Technology Division CSIR-Central Leather Research Institute Chennai India School of Computing Science and Engineering Vellore Institute of Technology Vellore India Computer Science and Engineering Sri Sairam Engineering College Sirukalathur India Center for Artificial Intelligence Prince Mohammad Bin Fahd University Khobar Saudi Arabia Department of Information Systems College of Computer and Information Sciences Princess Nourah bint Abdulrahman University Riyadh11671 Saudi Arabia

Leukemia, a malignant disease characterized by the rapid proliferation of specific types of white blood cells (WBC), has prompted increased interest in leveraging automatic WBC classification system. This study presents the design of a robust WBC classification system, integrating deep learning with microscopic image analysis. The proposed classification model is designed to categorize blood cells into four distinct groups: eosinophil, lymphocyte, monocyte, and neutrophil. The study employs a Kaggle image dataset for comprehensive learning. The dataset comprises 12,500 microscopic images used to categorize the four blood cell groups. To enhance performance and reduce the likelihood of overfitting, transfer learning techniques are employed. Six deep transfer learning models namely VGG16, ResNet50, AlexNet, MobileNet, GoogLeNet, and EfficientNetBO are used for feature extraction and classification. The results of the experiments show that VGG16 and MobileNet demonstrated enhancements compared to other pre-trained models when evaluated on both augmented and original datasets. A diverse set of evaluation criteria was employed to thoroughly assess the effectiveness of medical image classification. Notably, the VGG16 and MobileNet models accurately predicted Lymphocytes and Monocytes with a 100% accuracy rate, and Eosinophil and Neutrophil with rates of 99.35% and 99.81%, respectively. This research work empower medical practitioners, particularly in the field of pathology and hematology by providing advanced tools for accurate blood cell classification. The proposed VGG16 and MobileNet models hold promise for simplifying WBC classification, improving decision-making processes, and ultimately advancing patient care in the medical domain. © The Author(s), under exclusive licence to Springer science+Business Media, LLC, part of Springer Nature 2024.

关键词： Deep learning

来源：评论

学校读者我要写书评

暂无评论

A systematic review on deep learning implementation in brain tumor segmentation, classification and prediction

引用

Multimedia Tools and Applications 2025年 1-40页

作者： Abid, Muhammad Adeel Munir, Kashif Institute of Computer Science Khwaja Fareed University of Engineering and Information Technology Rahim Yar Khan Pakistan Institute of Information Technology Khwaja Fareed University of Engineering and Information Technology Rahim Yar Khan Pakistan

The brain is the central part of the body that controls the overall functionality of the human body. The formulation of abnormal cells in the brain may lead to a brain tumor. Manual examination of a brain tumor is challenging and time-consuming. Deep learning is purely based on neural networks, and it's beneficial in identifying and diagnosing brain tumors. Different groups of researchers put efforts into implementing deep learning for the classification, segmentation, and prediction of brain tumor. They proposed different models of deep learning and the accuracy of their models in terms of dice score or evaluation parameters that are quite reasonable and acceptable. The primary purpose of this study is to review the different articles implementing deep learning in the segmentation, classification, and prediction of brain tumors. Appropriate keywords are used to extract articles from January 2018 to September 2023. Evaluation is done based on the implementation of deep learning and the accuracy of the proposed model, and a complete data sheet is maintained against each article. A total of 154 articles were collected, and 80 research articles were selected for this study after complete analysis. Each selected article proposed models based on deep learning with reasonable and acceptable accuracy in identifying the brain tumor. 2-D CNN and 3-D CNN are used on Magnetic Resonance Imaging (MRI) images of the brain for brain tumor detection. Thus, much more attention is required in the future on this topic to improve the accuracy of brain tumor identification. © The Author(s), under exclusive licence to Springer science+Business Media, LLC, part of Springer Nature 2025.

关键词： Magnetic resonance imaging

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：