检索结果-内蒙古大学图书馆

arXiv 2023年

作者： Tang, Wei Zhang, Weijia Zhang, Min-Ling School of Computer Science and Engineering Southeast University Nanjing210096 China Key Laboratory of Computer Network and Information Integration Southeast University Ministry of Education China School of Information and Physical Sciences The University of Newcastle CallaghanNSW2308 Australia

In many real-world tasks, the concerned objects can be represented as a multi-instance bag associated with a candidate label set, which consists of one ground-truth label and several false positive labels. Multi-instance partial-label learning (MIPL) is a learning paradigm to deal with such tasks and has achieved favorable performances. Existing MIPL approach follows the instance-space paradigm by assigning augmented candidate label sets of bags to each instance and aggregating bag-level labels from instance-level labels. However, this scheme may be suboptimal as global bag-level information is ignored and the predicted labels of bags are sensitive to predictions of negative instances. In this paper, we study an alternative scheme where a multi-instance bag is embedded into a single vector representation. Accordingly, an intuitive algorithm named DEMIPL, i.e., Disambiguated attention Embedding for Multi-Instance Partial-Label learning, is proposed. DEMIPL employs a disambiguation attention mechanism to aggregate a multi-instance bag into a single vector representation, followed by a momentum-based disambiguation strategy to identify the ground-truth label from the candidate label set. Furthermore, we introduce a real-world MIPL dataset for colorectal cancer classification. Experimental results on benchmark and real-world datasets validate the superiority of DEMIPL against the compared MIPL and partial-label learning approaches. Copyright © 2023, The Authors. All rights reserved.

关键词： Diseases

来源：评论

学校读者我要写书评

暂无评论

Multi-dimensional classification via stacked dependency exploitation

引用

Science China(information Sciences) 2020年第12期63卷 104-117页

作者： Bin-Bin JIA Min-Ling ZHANG School of Computer Science and Engineering Southeast University College of Electrical and Information Engineering Lanzhou University of Technology Key Laboratory of Computer Network and Information Integration (Southeast University) Ministry of EducationChina Collaborative Innovation Center of Wireless Communications Technology

Multi-dimensional classification(MDC) aims to build classification models for multiple heterogenous class spaces simultaneously, where each class space characterizes the semantics of an object w.r.t. one specific dimension. Modeling dependencies among class spaces plays a key role in solving MDC tasks, where most approaches work by assuming directed acyclic graph(DAG) structure or random chaining structure over class spaces. Different from existing probabilistic strategies, a deterministic strategy named Seem for dependency modeling is proposed in this paper via stacked dependency exploitation. In the first-level, pairwise dependencies are considered which can be modeled more reliably than modeling full dependencies among all class spaces by DAG or chaining structure. In the second-level, the class label of unseen instance *** class space is determined by adaptively stacking predictive outputs from first-level pairwise *** results show that stacked dependency exploitation leads to superior performance against stateof-the-art MDC approaches.

关键词： machine learning multi-dimensional classification class dependencies deterministic strategy stacked dependency exploitation

来源：评论

学校读者我要写书评

暂无评论

Long-tail learning with foundation model: heavy fine-tuning hurts 24

Long-tail learning with foundation model: heavy fine-tuning ...

引用

Proceedings of the 41st International Conference on Machine Learning

作者： Jiang-Xin Shi Tong Wei Zhi Zhou Jie-Jing Shao Xin-Yan Han Yu-Feng Li National Key Laboratory for Novel Software Technology and School of Artificial Intelligence Nanjing University China School of Computer Science and Engineering Southeast University China and Key Laboratory of Computer Network and Information Integration Southeast University Ministry of Education China National Key Laboratory for Novel Software Technology Nanjing University China

The fine-tuning paradigm in addressing longtail learning tasks has sparked significant interest since the emergence of foundation models. Nonetheless, how fine-tuning impacts performance in long-tail learning was not explicitly quantified. In this paper, we disclose that heavy fine-tuning may even lead to nonnegligible performance deterioration on tail classes, and lightweight fine-tuning is more effective. The reason is attributed to inconsistent class conditions caused by heavy fine-tuning. With the observation above, we develop a low-complexity and accurate long-tail learning algorithms LIFT with the goal of facilitating fast prediction and compact models by adaptive lightweight fine-tuning. Experiments clearly verify that both the training time and the learned parameters are significantly reduced with more accurate predictive performance compared with state-of-the-art approaches. The implementation code is available at https://***/shijxcs/LIFT.

关键词：

来源：评论

学校读者我要写书评

暂无评论

Disease Diagnosis with Cost-Sensitive Grouped Features Based on Deep Reinforcement Learning

Disease Diagnosis with Cost-Sensitive Grouped Features Based...

引用

International Joint Conference on Neural networks (IJCNN)

作者： Yikemaiti Sataer Zhiqiang Gao Xuelian Li School of Computer Science and Engineering Southeast University Nanjing China Key Laboratory of Computer Network and Information Integration (Southeast University) Ministry of Education China School of Foreign Studies Nanjing University of Posts and Telecommunications

Traditional machine learning algorithms rely on the assumption that all features of a given dataset are available for free. However, in disease diagnosis, acquiring features requires a certain cost because there are many concerns, such as monetary data collection costs, patient discomfort in medical procedures, and privacy impacts of data collection that require careful consideration. Moreover, some medical examination items are grouped in the real world, such as lab tests that return multiple measurements, resulting in naturally grouped features. In this work, to consider features that come in groups, we propose a novel feature selection method for disease diagnosis with cost-sensitive grouped features based on deep reinforcement learning (RL). We optimize the objective which is a trade-off between the cost of grouped features and the diagnostic accuracy, by training deep RL models. In addition, the approach is flexible, as it can be improved with new deep reinforcement learning algorithms. We evaluate the effectiveness of the proposed method on four publicly available datasets. Compared with the existing competitive feature selection methods, the proposed method can achieve higher learning accuracy with lower feature costs. This paper takes disease diagnosis as an example, and the proposed method is universal for problems with grouped features of different domains. (The source code is available at https://***/r/DRL-CSGF- Y821).

关键词：

来源：评论

学校读者我要写书评

暂无评论

PE: A Poincare Explanation Method for Fast Text Hierarchy Generation

arXiv

引用

arXiv 2024年

作者： Chen, Qian Li, Dongyang He, Xiaofeng Li, Hongzhao Yi, Hongyu School of Computer Science and Technology East China Normal University Shanghai China NPPA Key Laboratory of Publishing Integration Development ECNUP Shanghai China Sichuan Caizi Software Information Network Co. Ltd China

The black-box nature of deep learning models in NLP hinders their widespread application. The research focus has shifted to Hierarchical Attribution (HA) for its ability to model feature interactions. Recent works model non-contiguous combinations with a time-costly greedy search in Eculidean spaces, neglecting underlying linguistic information in feature representations. In this work, we introduce a novel method, namely Poincare Explanation (PE), for modeling feature interactions with hyperbolic spaces in a time efficient manner. Specifically, we take building text hierarchies as finding spanning trees in hyperbolic spaces. First we project the embeddings into hyperbolic spaces to elicit inherit semantic and syntax hierarchical structures. Then we propose a simple yet effective strategy to calculate Shapley score. Finally we build the the hierarchy with proving the constructing process in the projected space could be viewed as building a minimum spanning tree and introduce a time efficient building algorithm. Experimental results demonstrate the effectiveness of our approach. Copyright © 2024, The Authors. All rights reserved.

关键词： Semantics

来源：评论

学校读者我要写书评

暂无评论

Improving User QoE via Joint Trajectory and Resource Optimization in Multi-UAV Assisted MEC

引用

IEEE Transactions on Services Computing 2025年第3期18卷 1472-1486页

作者： Gao, Yang Tao, Jun Xu, Yifan Wang, Zuyan Gao, Yu Wang, Meiling Hainan University School of Cyberspace Security Haikou570228 China Ministry of Education Key Laboratory of Computer Network and Information Integration Southeast University Nanjing211189 China Southeast University School of Cyber Science and Engineering Nanjing211189 China ZTE Communications Co. Ltd Nanjing210012 China

As a promising network architecture, Mobile Edge Computing (MEC), has been proven that can effectively reduce the end-to-end latency and the energy consumption. The Unmanned Aerial Vehicle (UAV) assisted MEC network, where the UAV can provide the computation offloading services for the mobile users, can further alleviate the huge deployment cost of static edge servers. However, it remains unsolved how multiple cooperative flying UAVs serve the ground users, especially considering that these UAVs may share the same wireless channel and can communicate with the users while flying. In this paper, we first propose the Age of Task (AoT) metric to measure the quality of experience, and then formulate the joint optimization problem to minimize the worst AoT among all the users. Based on the block coordinate descent (BCD) method, this problem is transformed into three non-convex programming sub-problems (i.e., the UAV-user association sub-problem, the UAV trajectory planning sub-problem and the transmit power optimization sub-problem). Specifically, the successive convex approximation (SCA) technique is exploited iteratively to deal with the non-convexity in the UAV trajectory and transmit power optimization. Numerical results show that the proposed scheme outperforms the benchmark offloading schemes in terms of AoT. © 2008-2012 IEEE.

关键词： Drones

来源：评论

学校读者我要写书评

暂无评论

Data-consistent Unsupervised Diffusion Model for Metal Artifact Reduction

Data-consistent Unsupervised Diffusion Model for Metal Artif...

引用

2023 IEEE International Conference on Bioinformatics and Biomedicine, BIBM 2023

作者： Tong, Zhan Wu, Zhan Yang, Yang Mao, Weilong Wang, Shijie Li, Yinsheng Chen, Yang Southeast University Laboratory of Image Science and Technology Nanjing210096 China Southeast University Ministry of Education Key Laboratory of Computer Network and Information Integration Nanjing210096 China Chinese Academy of Sciences Research Center for Medical Artificial Intelligence Shenzhen Institutes of Advanced Technology Shenzhen518055 China School of Computer Science and Engineering Key Lab. of New Generation Artificial Intelligence Technology and Its Interdisciplinary Applications Jiangsu Provincial Joint International Research Laboratory of Medical Information Processing The Laboratory of Image Science and Technology Nanjing210096 China

ISBN: (纸本)9798350337488

Computed Tomography (CT) is an imaging technique widely used in clinical diagnosis. However, high-attenuation metallic implants result in the obstruction of low-energy Xrays and further lead to metal artifacts in the reconstructed CT images. Deep supervised model-based metal artifact reduction(MAR) approaches are limited in clinical applications due to the difficulty in obtaining paired artifact-affected and artifactfree data. Furthermore, these model-based methods lack the consideration of data consistency in the sinogram-domain to perform exact metal trace inpainting. To address these challenges, we propose a Data-consistent unsupErVised diffusiOn model for meTal artifact rEDuction, called DEVOTED-Net. First, DEVOTED-Net leverages prior knowledge to guide the conditional diffusion model for fine-grained metal trace inpainting. Second, an unsupervised MAR framework is designed in the reverse process for the unknown metal traces restoration in the sinogram domain. Third, to further enhance the sinogram-domain data consistency, physics-based consistency constraint loss including conjugateray consistency loss and accumulation-ray consistency loss is designed. Extensive experiments are carried out to verify the performance of our algorithm on the publicly available dataset and clinical experimental dataset. This efficient, accurate, and reliable MAR approach holds great potential in clinics. © 2023 IEEE.

关键词： computed tomography deep unsupervised learning denoising diffusion probabilistic model Metal artifact reduction physics-based consistency constraint

来源：评论

学校读者我要写书评

暂无评论

Task Offloading and Resource Allocation with Privacy Constraints in End-Edge-Cloud Environment 17th

Task Offloading and Resource Allocation with Privacy Constra...

引用

17th CCF Conference on computer Supported Cooperative Work and Social Computing, ChineseCSCW 2022

作者： Zhu, Xia Sun, Wei Li, Xiaoping School of Computer Science and Engineering Southeast University Nanjing China Key Laboratory of Computer Network and Information Integration Ministry of Education Southeast University Nanjing China School of Cyber Science and Engineering Southeast University Nanjing China

ISBN: (纸本)9789819923847

The end-edge-cloud architecture brings an efficient solution to the big data processing problem caused by massive IoT devices. The characteristics of this architecture, such as massive devices, heterogeneous resources, and complex layers, bring new challenges to privacy protection issues. This paper considers the cooperation among terminal devices, edge nodes, and remote clouds to solve the task scheduling problem with privacy constraints by optimizing the offloading decision and resource allocation. A heuristic privacy-aware task offloading and resource allocation algorithm is proposed to maximize the number of successful tasks, which offloads low-privacy and non-privacy tasks to find sub-optimal offloading decisions by offloading sequence generation rules and decision adjustment. Task scheduling algorithms are presented by communication and computing resource allocation strategies for low-privacy and non-privacy tasks. ANOVA technique is used to verify the performance of the proposed algorithm. The experimental results show that the proposed algorithm is superior to others in terms of the number of devices, the amount of different task data, and the proportion of privacy tasks at different levels. © 2023, The Author(s), under exclusive license to Springer Nature Singapore Pte Ltd.

关键词： Cloud computing architecture

来源：评论

学校读者我要写书评

暂无评论

Robust restoration of low-dose cerebral perfusion CT images using NCS-Unet

引用

Nuclear Science and Techniques 2022年第3期33卷 62-76页

作者： Kai Chen Li-Bo Zhang Jia-Shun Liu Yuan Gao Zhan Wu Hai-Chen Zhu Chang-Ping Du Xiao-Li Mai Chun-Feng Yang Yang Chen Southeast University Nanjing 210096China School of Cyber Science and Engineering Southeast UniversityNanjing 210096China Key Laboratory of Computer Network and Information Integration(Southeast University) Ministry of EducationNanjing 210096China Department of Radiology General Hospital of the Northern Theater of the Chinese People’s Liberation ArmyShenyang 110016China Laboratory of Image Science and Technology The School of Computer Science and EngineeringSoutheast UniversityNanjing 210096China Department of Radiology Nanjing Drum Tower HospitalThe Affiliated Hospital of Nanjing University Medical SchoolNanjing 210008China Jiangsu Key Laboratory of Molecular and Functional Imaging Department of RadiologyZhongda HospitalSoutheast UniversityNanjing 210009China Jiangsu Provincial Joint International Research Laboratory of Medical Information Processing School of Computer Science and EngineeringSoutheast UniversityNanjing 210096China

Cerebral perfusion computed tomography(PCT)is an important imaging modality for evaluating cerebrovascular diseases and stroke *** widespread public concern about the potential cancer risks and health hazards associated with cumulative radiation exposure in PCT imaging,considerable research has been conducted to reduce the radiation dose in X-ray-based brain perfusion *** the dose of X-rays causes severe noise and artifacts in PCT *** solve this problem,we propose a deep learning method called *** exceptional characteristics of non-subsampled contourlet transform(NSCT)and the Sobel filter are introduced into *** decomposes the convolved features into high-and low-frequency *** decomposed high-frequency component retains image edges,contrast imaging traces,and noise,whereas the low-frequency component retains the main image *** Sobel filter extracts the contours of the original image and the imaging traces caused by the contrast agent *** extracted information is added to NCS-Unet to improve its performance in noise reduction and artifact *** and quantitative analyses demonstrated that the proposed NCS-Unet can improve the quality of low-dose cone-beam CT perfusion reconstruction images and the accuracy of perfusion parameter calculations.

关键词： Cerebral perfusion CT Low-dose Image denoising Perfusion parameters

来源：评论

学校读者我要写书评

暂无评论

Explainable Topic-Enhanced Argument Mining from Heterogeneous Sources

arXiv

引用

arXiv 2023年

作者： Si, Jiasheng Zhu, Yingjie Shi, Xingyu Zhou, Deyu He, Yulan School of Computer Science and Engineering Key Laboratory of Computer Network and Information Integration Ministry of Education Southeast University China Department of informatics King’s College London United Kingdom Alan Turing Insitute United Kingdom

Given a controversial target such as "nuclear energy", argument mining aims to identify the argumentative text from heterogeneous sources. Current approaches focus on exploring better ways of integrating the target-associated semantic information with the argumentative text. Despite their empirical successes, two issues remain unsolved: (i) a target is represented by a word or a phrase, which is insufficient to cover a diverse set of target-related subtopics;(ii) the sentence-level topic information within an argument, which we believe is crucial for argument mining, is ignored. To tackle the above issues, we propose a novel explainable topic-enhanced argument mining approach. Specifically, with the use of the neural topic model and the language model, the target information is augmented by explainable topic representations. Moreover, the sentence-level topic information within the argument is captured by minimizing the distance between its latent topic distribution and its semantic representation through mutual learning. Experiments have been conducted on the benchmark dataset in both the in-target setting and the cross-target setting. Results demonstrate the superiority of the proposed model against the state-of-the-art baselines. Copyright © 2023, The Authors. All rights reserved.

关键词： Semantics

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：