检索结果-内蒙古大学图书馆

2024 Conference on Empirical Methods in Natural Language Processing, EMNLP 2024

作者： Yuan, Xinfeng Yuan, Siyu Cui, Yuhan Lin, Tianhe Wang, Xintao Xu, Rui Chen, Jiangjie Yang, Deqing School of Data Science Fudan University China Shanghai Key Laboratory of Data Science School of Computer Science Fudan University China

ISBN: (纸本)9798891761643

Large language models (LLMs) have demonstrated impressive performance and spurred numerous AI applications, in which role-playing agents (RPAs) are particularly popular, especially for fictional characters. The prerequisite for these RPAs lies in the capability of LLMs to understand characters from fictional works. Previous efforts have evaluated this capability via basic classification tasks or characteristic imitation, failing to capture the nuanced character understanding with LLMs. In this paper, we propose evaluating LLMs' character understanding capability via the character profiling task, i.e., summarizing character profiles from corresponding materials, a widely adopted yet understudied practice for RPA development. Specifically, we construct the CROSS dataset from literature experts and assess the generated profiles by comparing them with ground truth references and evaluating their applicability in downstream tasks. Our experiments, which cover various summarization methods and LLMs, have yielded promising results. These results strongly validate the character understanding capability of LLMs. Resources are available at https://***/Joanna0123/character_profiling. © 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

User Distribution Mapping Modelling with Collaborative Filtering for Cross Domain Recommendation 24

User Distribution Mapping Modelling with Collaborative Filte...

引用

33rd ACM Web Conference, WWW 2024

作者： Liu, Weiming Chen, Chaochao Liao, Xinting Hu, Mengling Su, Jiajie Tan, Yanchao Wang, Fan College of Computer Science Zhejiang University China College of Computer and Data Science Fuzhou University China

ISBN: (纸本)9798400701719

User cold-start recommendation aims to provide accurate items for the newly joint users and is a hot and challenging problem. Nowadays as people participant in different domains, how to recommend items in the new domain for users in an old domain has become more urgent. In this paper, we focus on the Dual Cold-Start Cross Domain Recommendation (Dual-CSCDR) problem. That is, providing the most relevant items for new users on the source and target domains. The prime task in Dual-CSCDR is to properly model user-item rating interactions and map user expressive embeddings across domains. However, previous approaches cannot solve Dual-CSCDR well, since they separate the collaborative filtering and distribution mapping process, leading to the error superimposition issue. Moreover, most of these methods fail to fully exploit the cross-domain relationship among large number of non-overlapped users, which strongly limits their performance. To fill this gap, we propose User Distribution Mapping model with Collaborative Filtering (UDMCF), a novel end-to-end cold-start cross-domain recommendation framework for the Dual-CSCDR problem. UDMCF includes two main modules, i.e., rating prediction module and distribution alignment module. The former module adopts one-hot ID vectors and multi-hot historical ratings for collaborative filtering via a contrastive loss. The latter module contains overlapped user embedding alignment and general user subgroup distribution alignment. Specifically, we innovatively propose unbalance distribution optimal transport with typical subgroup discovering algorithm to map the whole user distributions. Our empirical study on several datasets demonstrates that UDMCF significantly outperforms the state-of-the-art models under the Dual-CSCDR setting. © 2024 ACM.

关键词： Collaborative filtering

来源：评论

学校读者我要写书评

暂无评论

ENHANCED LABEL PROPAGATION THROUGH AFFINITY MATRIX FUSION FOR SOURCE-FREE DOMAIN ADAPTATION 3

ENHANCED LABEL PROPAGATION THROUGH AFFINITY MATRIX FUSION FO...

引用

3rd Conference on Lifelong Learning Agents, CoLLAs 2024

作者： Guo, Li Xia, Yuxuan Wang, Shengjie Department of Computer Science and Data Science New York University Shanghai China

Source-free domain adaptation (SFDA) has gained significant attention as a method to transfer knowledge from a pre-trained model on source domains toward target domains without accessing the source data. Recent research in SFDA has predominately adopted a self-training paradigm, focusing on utilizing local consistency constraints to refine pseudo-labels during self-training. These methods aim to encourage similar predictions among samples residing in local neighborhoods. However, despite their effectiveness, the significance of global consistency is often overlooked in such approaches. Moreover, such self-training based adaptation processes suffer from the "confirmation bias": models use self-generated sub-optimal pseudo-labels to guide their subsequent training, resulting in a loop of self-reinforcing errors. In this study, we address the aforementioned limitations through two key contributions. Firstly, we introduce a label propagation method that seamlessly enforces both local and global consistency, leading to more coherent label predictions within the target domain. Secondly, to mitigate the "confirmation bias", we aggregate the affinity matrix derived from current and historical models during the label propagation process. This approach takes advantage of different snapshots of the model to obtain a more accurate representation of the underlying graph structure, significantly enhancing the efficacy of label propagation and resulting in more refined pseudo-labels. Extensive experimental evaluations demonstrate the superiority of our approach over existing methods by a large margin. Our findings not only highlight the significance of incorporating global consistency within the SFDA framework but also offer a novel approach to mitigate the confirmation bias that arises from the use of noisy pseudo-labels in the self-training paradigm. © 2024 CoLLAs. All Rights Reserved.

关键词： Domain Knowledge

来源：评论

学校读者我要写书评

暂无评论

An efficient Quasi-Affine Transformation Evolutionary algorithm with fixed dimension updating and its application in UAV 3D path planning

引用

Journal of Intelligent and Fuzzy Systems 2024年第4期46卷 9755-9781页

作者： Sung, Tien-Wen Zhao, Baohua Zhang, Xin Lee, Chao-Yang Fang, Qingjun Fujian Provincial Key Laboratory of Big Data Mining and Applications College of Computer Science and Mathematics Fujian University of Technology Fuzhou China Department of Computer Science and Information Engineering National Yunlin University of Science and Technology Yunlin Taiwan

Quasi-Affine Transformation Evolutionary (QUATRE) algorithm is a kind of swarm-based collaborative optimization algorithm that solves the problem of a position deviation in a DE search by using the co-evolution matrix M instead of the cross-control parameter CR in the differential evolution algorithm (DE). However, QUATRE shares some of the same weaknesses as DE, such as premature convergence and search stagnation. Inspired by the artificial bee colony algorithm (ABC), we propose a new QUATRE algorithm to improve these problems that ranks all the individuals and evolves only the poorer half of the population. In an evolving population, individuals of different levels intersect with dimensions of different sizes to improve search efficiency and accuracy. In addition, we establish a better selection framework for the parent generation individuals and select more excellent parent individuals to complete the evolution for the individuals trapped in search stagnation. To verify the performance of the new QUATRE algorithm, we divide the comparison algorithm into three groups, including ABC variant group, DE variant group, and QUATRE variant group, and the CEC2014 test suite is used for the comparison. The experimental results show the new QUATRE algorithm performance is competitive. We also successfully apply the new QUATRE algorithm on the 3D path planning of UAV, and compared with the other famous algorithm performance it is still outstanding, which verifies the algorithm's practicability. © 2024 - IOS Press. All rights reserved.

关键词： Optimization

来源：评论

学校读者我要写书评

暂无评论

Age-of-Information-Aware Federated Learning

引用

Journal of computer science & Technology 2024年第3期39卷 637-653页

作者：徐殷肖明军吴晨吴杰周津锐孙贺 Student Member CCF 1.School of Computer Science and TechnologyUniversity of Science and Technology of ChinaHefei 230026China Suzhou Institute for Advanced Research University of Science and Technology of ChinaSuzhou 215123China School of Data Science University of Science and Technology of ChinaHefei 230026China Department of Computer and Information Sciences Temple UniversityPhiladelphiaPA 19122U.S.A. CCF IEEE

Federated learning(FL)is an emerging privacy-preserving distributed computing paradigm,enabling numerous clients to collaboratively train machine learning models without the necessity of transmitting clients’private datasets to the central *** most existing research where the local datasets of clients are assumed to be unchanged over time throughout the whole FL process,our study addresses such scenarios in this paper where clients’datasets need to be updated periodically,and the server can incentivize clients to employ as fresh as possible datasets for local model *** primary objective is to design a client selection strategy to minimize the loss of the global model for FL loss within a constrained *** this end,we introduce the concept of“Age of Information”(AoI)to quantitatively assess the freshness of local datasets and conduct a theoretical analysis of the convergence bound in our AoI-aware FL *** on the convergence bound,we further formulate our problem as a restless multi-armed bandit(RMAB)***,we relax the RMAB problem and apply the Lagrangian Dual approach to decouple it into multiple ***,we propose a Whittle’s Index Based Client Selection(WICS)algorithm to determine the set of selected *** addition,comprehensive simulations substantiate that the proposed algorithm can effectively reduce training loss and enhance the learning accuracy compared with some state-of-the-art methods.

关键词： federated learning Age of Information restless multi-armed bandit Whittle’s index

来源：评论

学校读者我要写书评

暂无评论

Enabling high fault-tolerant embedding capability of alternating group graphs

引用

Future Generation computer Systems 2024年 158卷 110-121页

作者： Zhuang, Hongbin Li, Xiao-Yan Wang, Dajin Lin, Cheng-Kuan Zhao, Kun College of Computer and Data Science Fuzhou University Fuzhou350108 China Department of Computer Science Montclair State University Upper MontclairNJ07043 United States Department of Computer Science National Yang Ming Chiao Tung University Hsinchu30010 Taiwan

The Hamiltonian path/cycle serves as a robust tool for transmitting messages within parallel and distributed systems. However, the prevalent device-intensive nature of these systems often leads to the occurrence of faults. Tackling the critical challenge of tolerating numerous faults when constructing Hamiltonian paths and cycles in these systems is of utmost significance. The alternating group graph AGn is a suitable interconnection network for building parallel and distributed systems due to its outstanding topological properties. Many studies have been dedicated to the fault-tolerant embedding of Hamiltonian paths and cycles in AGn, but they all fail to reach the desired level of fault-tolerance. In this paper, we aim to enhance the edge fault-tolerant embedding capability of AGn by leveraging a newly emerged fault model, called the Partitioned Edge Fault (PEF) model. We prove the existence of Hamiltonian paths/cycles tolerating large-scale edge faults in AGn under the PEF model. Additionally, we propose a fault-tolerant Hamiltonian path embedding algorithm for AGn under the PEF model and validate its effectiveness through experiments. To gauge the significance of each missed edge, we conduct experimental calculations for the average path length of every edge along the Hamiltonian path produced by the proposed embedding algorithm. Furthermore, the fault-tolerance comparison and experimental analysis reveal that our results improve the fault-tolerant embedding capability of AGn from a linear level to an exponential one. To our knowledge, this is the first time that a Hamiltonian path/cycle embedding approach is proposed and actually implemented for AGn with large-scale edge faults. © 2024 Elsevier B.V.

关键词： Fault tolerance

来源：评论

学校读者我要写书评

暂无评论

Influence of different form of multimedia on motivation and topic perception in augmented reality learning experiences

引用

Multimedia Tools and Applications 2025年 1-28页

作者： Yap, Shook Chin Rahmat, Rahmita Wirza Ali, Siti Khadijah School of Computing and Data Science Xiamen University Malaysia Selangor Sepang Malaysia Faculty of Computer Science and Information Technology Universiti Putra Malaysia Selangor Serdang Malaysia

It has been widely proven that Augmented Reality (AR) brings numerous benefits in learning experiences, including enhancing learning outcomes and motivation. However, not many studies investigate how different forms of multimedia help in motivating students in AR learning experiences. Instead of examining whether AR can be used in education to motivate students, our research focused on investigating how and when AR learning experiences influence motivation when integrated with multimedia learning. A between-subjects 2 × 3 factorial design was employed in this study. The first factor was the type of AR (static-AR or animated-AR), and the second factor was the modality of the description (concurrent narration-text, text only, or narration only). Six experimental conditions were designed, each equipped with different versions of AR learning applications with various combinations of multimedia formats. One hundred and forty-four computer science and information technology undergraduate students (70 males and 74 females) were randomly assigned to one of the six experimental conditions. data collection was conducted with questionnaires, and data were analyzed using univariate analysis of variance (ANOVA). Results show that attention and satisfaction subscales of the ARCS motivational model can be enhanced with 3D animation implemented in AR. Descriptions delivered with concurrent narration and written text in AR are effective in enhancing motivation as well. Besides, participants’ perceived understanding and interest in learning the topic increased after AR learning experiences. © The Author(s), under exclusive licence to Springer science+Business Media, LLC, part of Springer Nature 2025.

关键词： Contrastive Learning

来源：评论

学校读者我要写书评

暂无评论

Inductive Lottery Ticket Learning for Graph Neural Networks

引用

Journal of computer science & Technology 2024年第6期39卷 1223-1237页

作者： Yong-Duo Sui Xiang Wang Tianlong Chen Meng Wang Xiang-Nan He Tat-Seng Chua School of Data Science University of Science and Technology of ChinaHefei 230027China Department of Electrical and Computer Engineering The University of Texas at AustinAustin 78712U.S.A School of Computer Science and Information Engineering Hefei University of TechnologyHefei 230009China School of Computing National University of SingaporeSingapore

Graph neural networks (GNNs) have gained increasing popularity, while usually suffering from unaffordable computations for real-world large-scale applications. Hence, pruning GNNs is of great need but largely unexplored. The recent work Unified GNN Sparsification (UGS) studies lottery ticket learning for GNNs, aiming to find a subset of model parameters and graph structures that can best maintain the GNN performance. However, it is tailed for the transductive setting, failing to generalize to unseen graphs, which are common in inductive tasks like graph classification. In this work, we propose a simple and effective learning paradigm, Inductive Co-Pruning of GNNs (ICPG), to endow graph lottery tickets with inductive pruning capacity. To prune the input graphs, we design a predictive model to generate importance scores for each edge based on the input. To prune the model parameters, it views the weight’s magnitude as their importance scores. Then we design an iterative co-pruning strategy to trim the graph edges and GNN weights based on their importance scores. Although it might be strikingly simple, ICPG surpasses the existing pruning method and can be universally applicable in both inductive and transductive learning settings. On 10 graph-classification and two node-classification benchmarks, ICPG achieves the same performance level with 14.26%–43.12% sparsity for graphs and 48.80%–91.41% sparsity for the GNN model.

关键词： lottery ticket hypothesis graph neural networks neural network pruning

来源：评论

学校读者我要写书评

暂无评论

Deployment and Offloading Optimization for Space-Air-Ground Integrated Network with Improved Differential Evolution Algorithm 10

Deployment and Offloading Optimization for Space-Air-Ground ...

引用

10th IEEE Smart World Congress, SWC 2024

作者： Zheng, Jingjing Zheng, Hongqiang Chen, Xing Zhang, Jianshan Fuzhou University College of Computer and Data Science FuZhou350116 China MinJiang University College of Computer and Data Science FuZhou350116 China

ISBN: (纸本)9798331520861

The development of the Space-Air-Ground Integrated Network (SAGIN) represents a significant advancement in satellite Internet technology, primarily because it facilitates the provision of pervasive networking solutions for the Internet of Things (IoT). Central to minimizing energy consumption in this context is the strategic deployment of Unmanned Aerial Vehicles (UAVs) and the efficient offloading of computational tasks. This is due to the fact that the quality of communication between UAVs and ground-based mobile devices is a crucial determinant of both communication latency and energy usage. This study presents an in-depth exploration of a Mobile Edge Computing (MEC) framework within a SAGIN context, in which multiple UAVs serve as edge computing servers, thereby accelerating computational services to terrestrial devices. In parallel, Low Earth Orbit (LEO) satellites are leveraged to provide cloud computing services. We introduce a Dual-layer nested joint optimization technique fortified by an improved differential evolution. The outer layer of the algorithm employs a phased differential evolution technique to optimize the positioning of UAVs. Simultaneously, the inner layer adopts a greedy approach to optimize the offloading ratio for computational tasks. The effectiveness and superior performance of the proposed method are substantiated through extensive numerical simulations. Comparative analysis with existing benchmark techniques demonstrates a significant reduction in the energy consumption of the system, underscoring the potential of the proposed framework. © 2024 IEEE.

关键词： Computation offloading

来源：评论

学校读者我要写书评

暂无评论

LesionAid: vision transformers-based skin lesion generation and classification – A practical review

引用

Multimedia Tools and Applications 2025年 1-22页

作者： K, Mallikharjuna Rao Krishna, Ghanta Sai Supriya, Kundrapu Sorgile, Meetiksha Data Science and Artificial Intelligence International Institute of Information Technology Naya Raipur India Computer Science and Engineering International Institute of Information Technology Naya Raipur India

Skin cancer is one of the most prevalent forms of human cancer. It is recognized mainly visually, beginning with clinical screening and continuing with the dermoscopic examination, histological assessment, and specimen collection. Deep convolutional neural networks (CNNs) perform highly segregated and potentially universal tasks against a classified fine-grained object. This study suggests a novel multi-class prediction framework that uses ViT and ViTGAN to categories skin lesions. To address the class disparity, GANs (Generative Adversarial Networks) based on vision transformers are used. The framework comprises four main phases: ViTGANs, Image processing, and explainable AI. Phase 1 consists of generating synthetic images using ViTGAN to balance all the classes in the dataset. To enhance the amount of the data, the second phase involves using various morphological processes and data augmentation techniques. In phases three and four, a ViT model for edge computing systems that can recognize patterns and classify skin lesions from the user's skin that is visible in the picture is developed. In phase 3, after classifying the lesions into the desired class with ViT, we will use explainable AI (XAI) that leads to more explainable results (using activation maps, etc.) while ensuring high predictive accuracy. The results demonstrate that the model used for generating synthetic images has achieved an FID score of 13.32 and the ViT model has achieved 99.2% as its training accuracy and 97.4% as its validation accuracy. The whole framework is compared with the existing frameworks for skin lesion detection. And explainable AI (XAI) is used in the proposed framework in order to increase model transparency and boost trust among users by illustrating the main factors affecting predictions. This interpretability helps physicians in making more informed decisions by providing clear insights into the model's reasoning. © The Author(s), under exclusive licence to Springer science+Bu

关键词： Generative adversarial networks

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：