检索结果-内蒙古大学图书馆

2025 IEEE International Conference on Acoustics, Speech, and Signal Processing, ICASSP 2025

作者： Huang, Diwei Lin, Kunyang Chen, Peihao Du, Qing School of Software Engineering South China University of Technology China Tencent AI Lab China Tencent Robotics X China

ISBN: (纸本)9798350368741

Few-shot audio-visual acoustics modeling seeks to synthesize the room impulse response in arbitrary locations with few-shot observations. To sufficiently exploit the provided few-shot data for accurate acoustic modeling, we present a map-guided framework by constructing acoustic-related visual semantic feature maps of the scenes. Visual features preserve semantic details related to sound and maps provide explicit structural regularities of sound propagation, which are valuable for modeling environment acoustics. We thus extract pixel-wise semantic features derived from observations and project them into a top-down map, namely the observation semantic map. This map contains the relative positional information among points and the semantic feature information associated with each point. Yet, limited information extracted by few-shot observations on the map is not sufficient for understanding and modeling the whole scene. We address the challenge by generating a scene semantic map via diffusing features and anticipating the observation semantic map. The scene semantic map then interacts with echo encoding by a transformer-based encoder-decoder to predict RIR for arbitrary speaker-listener query pairs. Extensive experiments on Matterport3D and Replica dataset verify the efficacy of our framework. © 2025 IEEE.

关键词： Acoustics Modeling Audio-Visual Learning Few-Shot Learning

来源：评论

学校读者我要写书评

暂无评论

Enhancing Intent Detection Through ChatGPT-Driven Data Augmentation 5th

Enhancing Intent Detection Through ChatGPT-Driven Data Augme...

引用

5th Doctoral Symposium on Computational Intelligence, DoSCI 2024

作者： Tesfagergish, Senait Gebremichael Damasevicius, Robertas Kapociute-Dzikiene, Jurgita Department of Software Engineering Kaunas University of Technology Kaunas Lithuania Department of Applied Informatics Vytautas Magnus University Kaunas Lithuania

ISBN: (纸本)9789819771776

This study investigates the improvement of intent detection tasks, focusing on the Facebook Multilingual Task Oriented Dataset, which encompasses 12 classes, including specific intents like weather and alarms. To address the scarcity of datasets in the Amharic language for intent detection, we first translated the English dataset into Amharic using Google Translator. Additionally, we expanded the dataset to include Lithuanian, German, French, and Czech languages. Next, we balance the dataset through data augmentation using ChatGPT, particularly targeting the four least-covered classes by adding 50 sentences to each. For text vectorization, the language-agnostic BERT sentence embedding model was used for its multilingual capabilities and Amharic language support. The classification was based on Cosine similarity, incorporating a majority voting mechanism from the most similar training instances. We conducted multilingual experiments to evaluate the performance of intent detection across various languages, both before and after data augmentation. Our findings highlight the effectiveness of using ChatGPT for data augmentation in intent detection tasks across all investigated languages, particularly in enhancing underrepresented languages and specific intents. We achieved improvement in intent detection performance for minority classes. This shows the benefit of using ChatGPT with imbalanced datasets for multilingual tasks and applications. © The Author(s), under exclusive license to Springer Nature Singapore Pte Ltd. 2025.

关键词： Translation (languages)

来源：评论

学校读者我要写书评

暂无评论

Understanding the effect of cathodic reaction on corrosion-enhanced erosion in sand-entraining electrolyte

引用

材料科学技术（英文版） 2025年第6期209卷 103-116页

作者： Qiliang Zhang Long Hao Wanbin Chen Yi Huang Yunze Xu State Key Laboratory of Structural Analysis Optimization and CAE Software for Industrial EquipmentDalian University of TechnologyDalian 116024China School of Naval Architecture and Ocean Engineering Dalian University of TechnologyDalian 116024China State Key Laboratory for Corrosion and Protection Institute of Metal ResearchChinese Academy of SciencesShenyang 110016China

In this study,the pure erosion behaviour of pure iron and its erosion-corrosion behaviour under different anodic polarization currents were investigated in various cathodic reactions(oxygen reduction,hydrogen ion reduction,and water reduction)using a cylindrical stirring *** corrosion-enhanced erosion(C-E)rates were determined for each *** results revealed that pure iron displayed similar pure erosion behaviour across all three cathodic *** the cathodic reactions involve hydrogen ion reduction or water reduction,the erosion-corrosion of pure iron manifested as uniform damage,with the reduction in hardness being the main cause of the C-E in this ***,in the case of oxy-gen reduction reaction as the cathodic reaction,the erosion-corrosion presented as pitting damage,with the reduction in hardness resulting from localized concentration of anodic current and the formation of easily worn protruding flaky iron structures at the edges of the pits as the main mechanism of the ***,linear and exponential relationships were found between the C-E rate and the anodic current density for uniform damage and pitting damage,***,the concept of surface equivalent hardness was proposed,along with the establishment of a mathematical model for surface equivalent hardness based on the relationships between the C-E rate and the anodic current *** the surface equivalent hardness enables the evaluation of the erosion rate on material surfaces considering the coupled effect.

关键词： Corrosion-enhanced erosion Cathodic reaction Pitting damage Uniform damage Hardness

来源：评论

学校读者我要写书评

暂无评论

ACbot:an IIoT platform for industrial robots

引用

Frontiers of Computer Science 2025年第4期19卷 13-27页

作者： Rui WANG Xudong MOU Tianyu WO Mingyang ZHANG Yuxin LIU Tiejun WANG Pin LIU Jihong YAN Xudong LIU School of Computer Science and Engineering Beihang UniversityBeijing 100191China Zhongguancun Laboratory Beijing 100194China School of Software Beihang UniversityBeijing 100191China School of Mechatronics Engineering Harbin Institute of TechnologyHarbin 150001China Hangzhou Innovation Institute Beihang UniversityHangzhou 310051China School of Information Engineering China University of GeosciencesBeijing 100083China

As the application of Industrial Robots(IRs)scales and related participants increase,the demands for intelligent Operation and Maintenance(O&M)and multi-tenant collaboration *** methods could no longer cover the requirements,while the Industrial Internet of Things(IIoT)has been considered a promising ***,there’s a lack of IIoT platforms dedicated to IR O&M,including IR maintenance,process optimization,and knowledge *** this context,this paper puts forward the multi-tenant-oriented ACbot platform,which attempts to provide the first holistic IIoT-based solution for O&M of *** on an information model designed for the IR field,ACbot has implemented an application architecture with resource and microservice management across the cloud and multiple *** this basis,we develop four vital applications including real-time monitoring,health management,process optimization,and knowledge *** have deployed the ACbot platform in real-world scenarios that contain various participants,types of IRs,and *** date,ACbot has been accessed by 10 organizations and managed 60 industrial robots,demonstrating that the platform fulfills our ***,the application results also showcase its robustness,versatility,and adaptability for developing and hosting intelligent robot applications.

关键词： IIoT platform industrial robots cloud-edge collaboration intelligent applications

来源：评论

学校读者我要写书评

暂无评论

Predictive Caching of File System for Database Workload Management 1

Predictive Caching of File System for Database Workload Mana...

引用

1st International Conference on AIML-Applications for engineering and technology, ICAET 2025

作者： Jadhav, Rohini Jadhav, Veena Pawar, Prajakta Kannaujia, Abhishek Kumar, Abhishek Gohil, Jayesh College of Engineering Department of Information Technology Pune India College of Engineering Department of Computer Engineering Pune India Bharati Vidyapeeth College of Engineering Department of Computer Engineering Lavale Pune India Veritas Principal Software Engineering India

ISBN: (纸本)9798350355611

The proactive caching technique known as 'predictive caching' attempts to improve file system performance by anticipating and pre-fetching data that is likely to be requested in the future. Conventional caching strategies, such LRU (Least Recently Used) and LFU (Least Frequently Used), base their decision on which material to keep in the cache on historical access patterns. Although these methods work well in many situations, they could not completely take advantage of the temporal and geographical proximity of file system access patterns, which could result in less-than-ideal cache use. Utilizing machine learning algorithms, statistical analysis, or heuristics, predictive caching makes predictions about future access patterns based on past data, system status, file properties, or user activity. So we propose a predictive caching technique which reduce access latency and boost system performance by pre-fetching and storing pertinent data ahead of time in anticipation of incoming data accesses. © 2025 IEEE.

关键词： Dynamic Data File Access Pattern Predictive Caching Random Forest Classifier

来源：评论

学校读者我要写书评

暂无评论

EPipe: Pipeline Inference Framework with High-quality Offline Parallelism Planning for Heterogeneous Edge Devices 24

EPipe: Pipeline Inference Framework with High-quality Offlin...

引用

43rd International Conference on Computer-Aided Design, ICCAD 2024

作者： Xiong, Yi Liu, Weihong Zhang, Rui Zu, Yulong Zhu, Zongwei Zhou, Xuehai School of Computer Science and Technology University of Science and Technology of China China Suzhou Institute for Advanced Research University of Science and Technology of China China School of Software Engineering University of Science and Technology of China China

ISBN: (纸本)9798400710773

Pipeline parallelism is essential for edge computing as it effectively consolidates the limited resources of edge devices, enabling the deployment of large Deep Neural Network (DNN) models and accelerating inference processes without compromising the performance of models. Accurate computation and communication latency estimation on heterogeneous edge devices is essential for searching for a superior parallelism plan. However, existing heterogeneous pipeline inference approaches either incur substantial resource wastage during online parallelism planning, as they utilize profiling strategies that occupy physical devices;or rely on cost models with inadequate representational capabilities, leading to inaccurate predictions, thereby harming the result of pipeline planning. This paper proposes EPipe, a novel pipeline inference framework that supports high-quality offline planning in heterogeneous edge environments. EPipe integrates two core components: the Task-Device Co-analyzer (TDC) and the Multi-pipeline Parallelism Planner (MPP). TDC utilizes an undirected connected graph to depict the compatibility of DNNs across device groups and precisely estimates inference and communication latencies through fine-grained modeling. Based on TDC, MPP utilizes a dynamic programming-based genetic algorithm to explore multi-pipeline solutions, extending beyond traditional single-pipeline methods. A comprehensive experimental evaluation on an edge testbed confirms the effectiveness of EPipe, demonstrating significant speedups in inference tasks for both task streams and single tasks. © 2024 Copyright is held by the owner/author(s).

关键词： Edge detection

来源：评论

学校读者我要写书评

暂无评论

U-net of joint spatial domains with multi-scale atrous convolution for rectal image segmentation

引用

Multimedia Tools and Applications 2025年 1-17页

作者： Rao, Yunbo Gao, Li Zeng, Shaoning Shao, Tingting Sun, Jihong School of Information and Software Engineering University of Electronic Science and Technology of China Chengdu611731 China Yangtze Delta Region Institute University of Electronic Science and Technology of China Huzhou313000 China Zhejiang University School of Medicine Zhejiang Hangzhou310058 China

Medical image segmentation is very important for the diagnosis of related diseases. To reduce the labeling work of related medical images, numerous models based on U-Net have been proposed to achieve automatic segmentation of target regions. However, most of these models are only trained in one coordinate system, ignoring the joint effects of different spatial coordinate systems. In addition, most of the encoding modules in these models do not pay attention to multi-scale spatial information. Our proposed solution for the aforementioned challenges involves using U-Net model with joint spatial domains and multi-scale encoding module, which enables us to segment rectal image better. The model includes a self-designed multi-layer dilated convolution encoding module named AIR (Atrous Inception Residual Block), to achieve a multi-scale content fusion. Besides this, it utilizes the center point and polar coordinates to realize attention mechanism and rotation invariance. Furthermore, retraining the output of polar coordinate network with Cartesian coordinate system realizes the translation invariance of segmentation. Compared with the commonly used medical segmentation models, the dice coefficient of our model is improved by about 2% on our in-house rectal dataset. © The Author(s), under exclusive licence to Springer Science+Business Media, LLC, part of Springer Nature 2025.

关键词： Image segmentation

来源：评论

学校读者我要写书评

暂无评论

An Uncertainty Quantization-Based Method for Anti-UAV Detection in Infrared Images

引用

Computers, Materials & Continua 2025年第4期83卷 1439-1458页

作者： Can Wu Wenyi Tang Yunbo Rao Yinjie Chen Hui Ding Shuzhen Zhu Yuanyuan Wang School of Information and Software Engineering University of Electronic Science and Technology of ChinaChengdu610054China State Key Laboratory of Air Traffic Management System Nanjing210000China National Airspace Management Center Beijing100094China

Infrared unmanned aerial vehicle(UAV)target detection presents significant challenges due to the inter-play between small targets and complex *** methods,while effective in controlled environments,often fail in scenarios involving long-range targets,high noise levels,or intricate backgrounds,highlighting the need for more robust *** address these challenges,we propose a novel three-stage UAV segmentation framework that leverages uncertainty quantification to enhance target *** framework incorporates a Bayesian convolutional neural network capable of generating both segmentation maps and probabilistic uncertainty *** utilizing uncer-tainty predictions,our method refines segmentation outcomes,achieving superior detection ***,this marks the first application of uncertainty modeling within the context of infrared UAV target *** evaluations on three publicly available infrared UAV datasets demonstrate the effectiveness of the proposed *** results reveal significant improvements in both detection precision and robustness when compared to state-of-the-art deep learning *** approach also extends the capabilities of encoder-decoder convolutional neural networks by introducing uncertainty modeling,enabling the network to better handle the challenges posed by small targets and complex environmental *** bridging the gap between theoretical uncertainty modeling and practical detection tasks,our work offers a new perspective on enhancing model interpretability and *** codes of this work are available openly at https://***/general-learner/UQ_Anti_UAV(acceessed on 11 November 2024).

关键词： Object segmentation uncertainty quantification bayesian convolutional neural network

来源：评论

学校读者我要写书评

暂无评论

Rethinking-based Code Summarization with Chain of Comments 31

Rethinking-based Code Summarization with Chain of Comments

引用

31st International Conference on Computational Linguistics, COLING 2025

作者： Cao, Liuwen He, Hongkui Huang, Hailin Wang, Jiexin Cai, Yi School of Software Engineering South China University of Technology China Key Laboratory of Big Data and Intelligent Robot South China University of Technology Ministry of Education China

ISBN: (纸本)9798891761964

Automatic code summarization aims to generate concise natural language descriptions (summary) for source code, which can free software developers from the heavy burden of manual commenting and software maintenance. Existing methods focus on learning a direct mapping from pure code to summaries, overlooking the significant heterogeneity gap between code and summary. Moreover, existing methods lack a human-like re-check process to evaluate whether the generated summaries match well with the code. To address these two limitations, we introduce RBCoSum, a novel framework that incorporates the generated Chain Of Comments (COC) as auxiliary intermediate information for the model to bridge the gap between code and summaries. Also, we propose a rethinking process where a learned ranker trained on our constructed ranking dataset scores the extent of matching between the generated summary and the code, selecting the highest-scoring summary to achieve a re-check process. We conduct extensive experiments to evaluate our approach and compare it with other automatic code summarization models as well as multiple code Large Language Models (LLMs). The experimental results show that RBCoSum is effective and outperforms baselines by a large margin. The human evaluation also proves the summaries generated with RBCoSum are more natural, informative, useful, and truthful. © 2025 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

Anything-DT Modeling: A New Framework for Digital Twin Modeling of Wild Objects from a Single RGB Image

Anything-DT Modeling: A New Framework for Digital Twin Model...

引用

2025 International Conference on Electrical Automation and Artificial Intelligence, ICEAAI 2025

作者： Geng, Jun Wang, Siqi Yi, Xiaoyun Xiong, Wanqing Yin, Hongchao Li, Xiaoyu The School of Software Xinjiang University Urumqi China The School of Information and Software Engineering University of Electronic Science and Technology of China Chengdu China

ISBN: (纸本)9798331506797

Given the significant limitations of current digital twin modeling methods in practical applications, their lack of sufficient generalizability, and the inherent diversity and complexity of real-world objects and environments, digital twin modeling from a single RGB image in unconstrained scenarios faces several maj or challenges. In this paper, we first improve the Segment Anything Model (SAM) and propose the CSAM model, which enables more accurate segmentation of any object. Next, we introduce the Anything-DT Modeling framework, which cleverly integrates the CSAM object segmentation model with a vision-language model to accomplish digital twin modeling tasks using only a single image. The CSAM model can maintain high accuracy and efficiency in complex scenarios, providing high-quality input data for subsequent digital twin modeling. The framework fully utilizes the latest advances in computer vision and natural language processing, providing strong technical support for digital twin modeling, thereby ensuring the feasibility of digital twin modeling from a single RGB image. Compared with existing state-of-the-art methods, the Anything-DT Modeling framework demonstrates its ability to produce detailed and accurate digital twin models for a wide range of objects, showcasing promising potential in overcoming the limitations of current approaches. © 2025 IEEE.

关键词： Digital Twin Modeling Image Segmentation Single-View 3D Modeling Thin Object Structures

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：