检索结果-内蒙古大学图书馆

arXiv 2024年

作者： Song, Rui Li, Yingji Shi, Lida Giunchiglia, Fausto Xu, Hao College of Computer Science and Technology Jilin University Changchun China Key Laboratory of Symbol Computation and Knowledge Engineering of Ministry of Education Jilin University China Department of Information Engineering and Computer Science University of Trento Italy

Shortcut learning refers to the phenomenon where models employ simple, non-rob.st decision rules in practical tasks, which hinders their generalization and rob.stness. With the rapid development of Large Language Models (LLMs) in recent years, an increasing numb.r of studies have shown the impact of shortcut learning on LLMs. This paper provides a novel perspective to review relevant research on shortcut learning in In-Context Learning (ICL). It conducts a detailed exploration of the types of shortcuts in ICL tasks, their causes, availab.e b.nchmarks, and strategies for mitigating shortcuts. b.sed on the corresponding ob.ervations, we summarize the unresolved issues in existing research and attempts to outline the future research landscape of shortcut learning. © 2024, CC b..

关键词： Contrastive Learning

来源：评论

学校读者我要写书评

暂无评论

Deep Feature-b.sed Text Clustering and Its Explanation

Deep Feature-Based Text Clustering and Its Explanation

引用

International Conference on Data engineering

作者： Renchu Guan Hao Zhang Yanchun Liang Fausto Giunchiglia Lan Huang Xiaoyue Feng Key Laboratory for Symbol Computation and Knowledge Engineering of the Ministry of Education College of Computer Science and Technology Jilin University Changchun China DISI University of Trento Trento Italy

Text clustering is a critical step in text data analysis and has b.en extensively studied b. the text mining community. Most existing text clustering algorithms are b.sed on the b.g-of-words model, which faces the high-dimensional and sparsity prob.ems and ignores text structural and sequence information. Deep learning-b.sed models such as convolutional neural networks and recurrent neural networks regard texts as sequences b.t lack supervised signals and explainab.e results. In this paper, we propose a deep feature-b.sed text clustering (DFTC) framework that incorporates pretrained text encoders into text clustering tasks. This model, which is b.sed on sequence representations, b.eaks the dependency on supervision. The experimental results show that our model outperforms classic text clustering algorithms on almost all the considered datasets. In addition, the explanation of the clustering results is significant for understanding the principles of the deep learning approach. Our proposed clustering framework includes an explanation module that can help users understand the meaning and quality of the clustering results. Our code is availab.e at https://***/KEAML-JLU/DeepTextClustering.

关键词：

来源：评论

学校读者我要写书评

暂无评论

Community discovery algorithm and its technical improvement b.sed on link structure–taking web.community algorithm as an example

Community discovery algorithm and its technical improvement ...

引用

International Conference on Applications and Techniques in Cyb.r Intelligence, ATCI 2020

作者： Gao, Rui Yang, Wenzhe Shi, Xiaohu Key Laboratory of Symbol Computation and Knowledge Engineering of the Ministry of Education College of Computer Science and Technology Jilin University 2699 Qianjin Street Changchun130012 China

ISBN: (纸本)9783030539795

Link analysis is an important way to discover potential web.communities. This paper analyzes the characteristics of the Web.link structure, and studies the traditional maximum flow algorithm and the maximum flow algorithm b.sed on HITS algorithm for edge capacity allocation. The traditional maximum flow prob.em, b.t there are still shortcomings. b.sed on the analysis of existing link similarity definitions, this paper proposes a new definition of link similarity and topic dissimilarity to b.tter describ. the relationship b.tween linked pages, and to measure b.sed on link similarity and topic dissimilarity. The similarity of the pages gives a more reasonab.e and efficient maximum flow-side capacity allocation scheme. The community discovery of 6 topics shows that the maximum flow algorithm proposed in this paper using the link similarity and topic dissimilarity model to allocate edge capacity can b.tter solve the prob.ems of existing algorithms and significantly improve the quality of the Web.community. This paper proposes new methods and ideas for page similarity measurement, and provides new strategies for improving the Web.community discovery algorithm b.sed on link analysis. © The Editor(s) (if applicab.e) and The Author(s), under exclusive license to Springer Nature Switzerland AG 2021.

关键词： Computer programming

来源：评论

学校读者我要写书评

暂无评论

Toward Moiré-Free and Detail-Preserving Demosaicking

arXiv

引用

arXiv 2023年

作者： Li, Xuanchen Niu, Yan Zhao, b. Shi, Haoyuan An, Zitong State Key Laboratory of Symbol Computation and Knowledge Engineering College of Computer Science and Technology Ministry of Education Jilin University Changchun China The College of Software Jilin University Changchun130012 China

3D convolutions are commonly employed b. demosaicking neural models, in the same way as solving other image restoration prob.ems. Counter-intuitively, we show that 3D convolutions implicitly impede the RGb.color spectra from exchanging complementary information, resulting in spectral-inconsistent inference of the local spatial high frequency components. As a consequence, shallow 3D convolution networks suffer the Moiré artifacts, b.t deep 3D convolutions cause over-smoothness. We analyze the fundamental difference b.tween demosaicking and other prob.ems that predict lost pixels b.tween availab.e ones (e.g., super-resolution reconstruction), and present the underlying reasons for the confliction b.tween Moiré-free and detail-preserving. From the new perspective, our work decouples the common standard convolution procedure to spectral and spatial feature aggregations, which allow strengthening glob.l communication in the spectral dimension while respecting local contrast in the spatial dimension. We apply our demosaicking model to two tasks: Joint Demosaicking-Denoising and Independently Demosaicking. In b.th applications, our model sub.tantially alleviates artifacts such as Moiré and over-smoothness at similar or lower computational cost to currently top-performing models, as validated b. diverse evaluations. Source code will b. released along with paper pub.ication. Copyright © 2023, The Authors. All rights reserved.

关键词： Convolution

来源：评论

学校读者我要写书评

暂无评论

Aspect b.sed Fine-Grained Sentiment Analysis for Pub.ic Policy Opinion Mining 21st

Aspect Based Fine-Grained Sentiment Analysis for Public Poli...

引用

21st International Symposium on knowledge and Systems Sciences, KSS 2022

作者： Zhao, Yueming Li, Ying Liu, Yijun Li, Qianqian College of Computer Science and Technology Key Laboratory of Symbol Computation and Knowledge Engineering of Ministry of Education Jilin University Changchun130012 China Institutes of Science and Development Chinese Academy of Sciences Beijing100190 China School of Public Policy and Management University of Chinese Academy of Sciences Beijing100049 China

ISBN: (纸本)9789811936098

Fine-grained mining has sub.tantial practical significance for understanding pub.ic opinion toward pub.ic policies and optimizing relevant decision-making processes. In this study, an aspect-b.sed sentiment analysis model integrated with b.th a sentiment lexicon for the policy domain and mutual information (MI) is estab.ished from the perspective of attitudinal orientations toward several elements of pub.ic policies, including sub.ects, ob.ects, and tools to analyze comments. First, terms related to the aspects of comments on different policy elements are extracted using the association rules in conjunction with the word2vec algorithm. In addition, the contextual neighb.r principle is employed to extract terms that express evaluative opinions and to extricate aspect-b.sed two-tuples from comments. Sub.equently, the sentiment lexicon for the policy domain is expanded using the sentiment orientation pointwise mutual information (SO-PMI) algorithm. Four machine learning models (i.e., the support vector machine, logistic regression, naive b.yes, and k-nearest neighb.rs models) and two deep learning models (i.e., the convolutional neural network (CNN) and long short-term memory models) are compared in terms of their classification performance b.sed on 87,304 comments on pub.ic policies crawled from the Weib. platform. The results show that the CNN model integrated with the domain-specific sentiment lexicon and mutual information is effective at facilitating policy element-oriented fine-grained sentiment analysis of pub.ic opinion toward policies and therefore has practical decision-making value for the government to b.tter understand the policy demands of the pub.ic. © 2022, The Author(s), under exclusive license to Springer Nature Singapore Pte Ltd.

关键词： Pub.ic policy

来源：评论

学校读者我要写书评

暂无评论

A novel multi-task learning network for skin lesion classification b.sed on multi-modal clues and lab.l-level fusion

引用

Computers in b.ology and Medicine 2024年 175卷 108549-108549页

作者： Lin, Qifeng Guo, Xiaoxin Feng, b. Guo, Juntong Ni, Shuang Dong, Hongliang Key Laboratory of Symbol Computation and Knowledge Engineering of Ministry of Education Jilin University 2699 Qianjin Street Changchun130012 China College of Computer Science and Technology Jilin University 2699 Qianjin Street Changchun130012 China College of Software Jilin University 2699 Qianjin Street Changchun130012 China

In this paper, we propose a multi-task learning (MTL) network b.sed on the lab.l-level fusion of metadata and hand-crafted features b. unsupervised clustering to generate new clustering lab.ls as an optimization goal. We propose a MTL module (MTLM) that incorporates an attention mechanism to enab.e the model to learn more integrated, variab.e information. We propose a dynamic strategy to adjust the loss weights of different tasks, and trade off the contrib.tions of multiple b.anches. Instead of feature-level fusion, we propose lab.l-level fusion and comb.ne the results of our proposed MTLM with the results of the image classification network to achieve b.tter lesion prediction on multiple dermatological datasets. We verify the effectiveness of the proposed model b. quantitative and qualitative measures. The MTL network using multi-modal clues and lab.l-level fusion can yield the significant performance improvement for skin lesion classification. © 2024 Elsevier Ltd

关键词： Metadata

来源：评论

学校读者我要写书评

暂无评论

Consistency lab.l-activated region generating network for weakly supervised medical image segmentation

引用

Computers in b.ology and Medicine 2024年 173卷 108380-108380页

作者： Du, Wei Huo, Yongkang Zhou, Rixin Sun, Yu Tang, Shiyi Zhao, Xuan Li, Ying Li, Gaoyang Key Laboratory of Symbol Computation and Knowledge Engineering of Ministry of Education College of Computer Science and Technology Jilin University Changchun130012 China Translational Medical Center for Stem Cell Therapy and Institute for Regenerative Medicine Shanghai East Hospital Shanghai200092 China Bioinformatics Department School of Life Sciences and Technology Tongji University Shanghai200092 China

The current methods of auto-segmenting medical images are limited due to insufficient and amb.guous pathonmorphological lab.ling. In clinical practice, rough classification lab.ls (such as disease or normal) are more commonly used than precise segmentation masks. However, there is still much to b. explored regarding utilizing these weak clinical lab.ls to accurately determine the lesion mask and guide medical image segmentation. In this paper, we proposed a weakly supervised medical image segmentation model to directly generate the lesion mask through a class activation map (CAM) guided cycle-consistency lab.l-activated region transferring network. Cycle-consistency enforces that the mappings b.tween the two domains should b. reversib.e, which ensures that the original image can b. reconstructed from the translated image. We developed a complementary b.anches fusion module to address the issue of b.urry b.undaries in CAM-guided segmentation. The complementary b.anch preserves the original semantic information of the non-lesion region and perfectly fuses the transferred feature of the lesion region with a complementary mask-constrained fake image generation process to clear the b.undary of the lesion and non-lesion regions. This module allows the class transformation to focus solely on the lab.l-activated region, resulting in more explicit segmentation. This model can accurately identify different region of medical images at the pixel-level while preserving the overall semantic structure semantion. It organizes disease lab.ls and corresponding regions during image synthesis. Our method utilizes a joint discrimination strategy that significantly enhances the precision of the produced lesion mask. Extensive experiments of the proposed method on b.aTs, ISIC and COVID-19 datasets demonstrate superior performance over existing state-of-the-art methods. The code and datasets are availab.e at: https://***/mlcb.jlu/MedImgSeg. © 2024 Elsevier Ltd

关键词： Image segmentation

来源：评论

学校读者我要写书评

暂无评论

Automatic Detection of Lumb.r Spondylolisthesis and Instab.lity Using Supine Magnetic Resonance Images

SSRN

引用

SSRN 2023年

作者： Liu, Zewen Zhang, Han Zhao, Hang Liu, Yuwu Zhang, Chi Shi, Xiaohu College of Software Jilin University Changchun130021 China Guangyuan Central Hospital China College of Computer Science and Technology Jilin University Changchun130021 China Key Laboratory of Symbol Computation and Knowledge Engineering Ministry of Education Changchun130021 China

b.ckground and Ob.ective: Lumb.r spondylolisthesis is a condition in which one verteb.a is displaced relative to the verteb.a b.low it. Lumb.r instab.lity can b. divided into static and dynamic spondylolisthesis depending on whether the degree of slippage changes significantly in images taken in different postures. Usually, the degree of slippage of lumb.r instab.lity is not ob.ious on magnetic resonance (MR) images taken in the supine position;therefore, it is easy to miss the detection of spondylolisthesis with lumb.r instab.lity b.sed on slippage. This feature of lumb.r instab.lity also poses a challenge for the automatic detection of dynamic spondylolisthesis, particularly in supine MR images. This study aimed to develop a two-stage model for the automatic detection of spondylolisthesis and lumb.r instab.lity using supine MR ***: Specifically, we used You Only Look Once (YOLO) to detect and crop consecutive images near the interverteb.al discs (IVDs) on b.th sagittal and transverse axis magnetic resonance images, and then classified the cropped images using weight-sharing convolutional neural networks (CNNs). Consistency regularization and a new contrast loss function ware also developed to improve the rob.stness. The MR data we used includes 108 mixed samples of lumb.r instab.lity and spondylolisthesis (39 cases of lumb.r instab.lity and 69 cases of spondylolisthesis) and 182 normal ***: We found that YOLO could detect the complete image sequence b.tween L1 and S1 with 97.24% accuracy (ACC). Compared with the ground truth lab.ls, the mean of intersection over union (mIOU) is 84.58%. In addition, the ACC and mIoU for transverse axis image are 98.62% and 85.71% respectively. In the classification stage, CNN achieved a maximum ACC of 97.9% with transverse axis image as inputs and using consistency ***: While MR images offer a wider range of applications in disease detection compared to X-rays, the identification of con

关键词： Stab.lity

来源：评论

学校读者我要写书评

暂无评论

A Review of dataset and lab.ling methods for causality extraction 28

A Review of dataset and labeling methods for causality extra...

引用

28th International Conference on computational Linguistics, COLING 2020

作者： Xu, Jinghang Zuo, Wanli Liang, Shining Zuo, Xianglin College of Computer Science and Technology Jilin University China Key Laboratory of Symbol Computation and Knowledge Engineering of Ministry of Education China

ISBN: (纸本)9781952148279

Causality represents the most important kind of correlation b.tween events. Extracting causality from text has b.come a promising hot topic in NLP. However, there is no mature research systems, evaluation rules and datasets for pub.ic evaluation. Moreover, there is a lack of unified causal sequence lab.ling methods, which constitute the key factors that hinder the progress of causality extraction research. We survey the limitations and shortcomings of existing causality research field comprehensively from the aspects of b.sic concepts, extraction methods, experimental data, and lab.ling methods, so as to provide reference for future research on causality extraction. We summarize the existing causality datasets, explore their practicab.lity and extensib.lity from multiple perspectives. Aiming at the prob.em of causal sequence lab.ling, we analyze the existing methods of causal sequence lab.ling, with a summarizations of its regulation. Multiple candidate causal lab.ling sequences are put forward according to lab.ling controversy to explore the optimal lab.ling method through experiments, and suggestions are provided for selecting lab.ling method. © 2020 COLING 2020 - 28th International Conference on computational Linguistics, Proceedings of the Conference. All rights reserved.

关键词： Extraction

来源：评论

学校读者我要写书评

暂无评论

An AutoEncoder-b.sed Matrix Factorization Approach to Estimating Cell Proportion from b.lk Tumor RNA-seq Data

An AutoEncoder-Based Matrix Factorization Approach to Estima...

引用

IEEE International Conference on b.oinformatics and b.omedicine (b.b.)

作者： Yingze Xu Yan Wang Xuping Xie Feilong Wang Qiong Chen Huiyan Sun College of Computer Science and Technology Key Laboratory of Symbol Computation and Knowledge Engineering of Ministry of Education Jilin University Changchun China College of Computer Science and Technology School of Artificial Intelligence Key Laboratory of Symbol Computation and Knowledge Engineering of Ministry of Education Jilin University Changchun China School of Artificial Intelligence Key Laboratory of Symbol Computation and Knowledge Engineering of Ministry of Education Jilin University Changchun China

ISBN: (纸本)9781665429825

The deconvolution of infiltrating immune cells and stromal cells from complex tumor tissues is significant for studying the impact of these cells on tumor development, as well as assisting cancer therapies. Integrating cell-specific marker genes from several references, we put forward an AutoEncoder-b.sed matrix factorization method to estimate the cell proportion in heterogeneous samples. The proportion predicted b. our method achieved over 0.95 Pearson correlation coefficient (PCC) with the ground truth. Moreover, through analyzing association b.tween cell proportion of tumor tissues and clinical information of the tumor patients, we found that the proportion of cancer-associated fib.ob.ast (CAF) gradually increased with the progress of tumor, and the proportion of b.cell in tissues was significantly related to the five-year survival rate of tumor patients.

关键词： Head Deconvolution Medical treatment Melanoma Fib.ob.asts Neck Reliab.lity

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：