检索结果-内蒙古大学图书馆

Large language model for table processing: a survey

Frontiers of Computer Science 2025年第2期19卷 71-87页

作者： Weizheng LU Jing ZHANG Ju FAN Zihao FU Yueguo CHEN Xiaoyong DU School of Information Renmin University of ChinaBeijing 100872China Key Laboratory of Data Engineering and Knowledge Engineering Beijing 100872China WPS Office Kingsoft Co.Zhuhai 519080China

Tables,typically two-dimensional and structured to store large amounts of data,are essential in daily activities like database queries,spreadsheet manipulations,Web table question answering,and image table information *** these table-centric tasks with Large Language Models(LLMs)or Visual Language Models(VLMs)offers significant public benefits,garnering interest from academia and *** survey provides a comprehensive overview of table-related tasks,examining both user scenarios and technical *** covers traditional tasks like table question answering as well as emerging fields such as spreadsheet manipulation and table data *** summarize the training techniques for LLMs and VLMs tailored for table ***,we discuss prompt engineering,particularly the use of LLM-powered agents,for various tablerelated ***,we highlight several challenges,including diverse user input when serving and slow thinking using chainof-thought.

关键词： data mining and knowledge discovery table processing large language model

来源：评论

学校读者我要写书评

暂无评论

Bayesian Network Reconstruction and Iterative Divergence Problem Solving Method Based on Norm Minimization

引用

Computer Modeling in engineering & Sciences 2025年第4期143卷 617-637页

作者： Kuo Li Aimin Wang Limin Wang Yuetan Zhao Xinyu Zhu Key Laboratory of Symbolic Computation and Knowledge Engineering of Ministry of Education Jilin UniversityChangchun130012China College of Software Jilin UniversityChangchun130012China

A Bayesian network reconstruction method based on norm minimization is proposed to address the sparsity and iterative divergence issues in network reconstruction caused by noise and missing *** method achieves precise adjustment of the network structure by constructing a preliminary random network model and introducing small-world network characteristics and combines L1 norm minimization regularization techniques to control model complexity and optimize the inference process of variable *** the experiment of game network reconstruction,when the success rate of the L1 norm minimization model’s existence connection reconstruction reaches 100%,the minimum data required is about 40%,while the minimum data required for a sparse Bayesian learning network is about 45%.In terms of operational efficiency,the running time for minimizing the L1 normis basically maintained at 1.0 s,while the success rate of connection reconstruction increases significantly with an increase in data volume,reaching a maximum of 13.2 ***,in the case of a signal-to-noise ratio of 10 dB,the L1 model achieves a 100% success rate in the reconstruction of existing connections,while the sparse Bayesian network had the highest success rate of 90% in the reconstruction of non-existent *** the analysis of actual cases,the maximum lift and drop track of the research method is 0.08 *** mean square error is 5.74 cm^(2).The results indicate that this norm minimization-based method has good performance in data efficiency and model stability,effectively reducing the impact of outliers on the reconstruction results to more accurately reflect the actual situation.

关键词： Bayesian norm minimization network reconstruction iterative divergence sparsity

来源：评论

学校读者我要写书评

暂无评论

Face Reconstruction-Based Generalized Deepfake Detection Model with Residual Outlook Attention

引用

ACM Transactions on Multimedia Computing, Communications and Applications 2025年第4期21卷 1-19页

作者： Shi, Zenan Liu, Wenyu Chen, Haipeng College of Computer Science and Technology Jilin University Changchun China Key Laboratory of Symbolic Computation and Knowledge Engineering of Ministry of Education Jilin University Changchun China

With the continuous development of deep counterfeiting technology, the information security in our daily life is under serious threat. While existing face forgery detection methods exhibit impressive accuracy when applied to datasets such as FaceForensics++ and Celeb-DF, they falter significantly when confronted with out-of-domain scenarios. This causes specialization of learned representations to known forgery patterns presented in the training set, rendering it difficult to detect forgeries with unknown patterns. To address this challenge, we propose a novel end-to-end Face Reconstruction-Based Generalized Deepfake Detection (FRG2D) model with Residual Outlook Attention (ROA), which emphasizes the robust visual representations of genuine faces and discerns the subtle differences between authentic and manipulated facial images. Our methodology entails reconstructing authentic face images using an encoder-decoder architecture based on U-net, facilitating a deeper understanding of disparities between genuine and manipulated facial images. Furthermore, we integrate the convolutional block attention module (CBAM) and channel attention block (CAB) to selectively focus the network's attention on salient features within real face images. Furthermore, we employ ROA to guide the network's focus towards precise features within manipulated facial images. Simultaneously, the computed reconstruction differences obtained through ROA serves as the ultimate representation fed into the classifier for face forgery detection. Both the reconstruction and classification learning processes are optimized end-to-end. Through extensive experimentation, our model demonstrated a substantial improvement in deepfake detection across unknown domains, while maintaining a high accuracy within the known domain. © 2025 Copyright held by the owner/author(s). Publication rights licensed to ACM.

关键词： Deep learning

来源：评论

学校读者我要写书评

暂无评论

A level set model with shape prior constraint for intervertebral disc MRI image segmentation

引用

Multimedia Tools and Applications 2025年第9期84卷 6755-6783页

作者： Tian, Zhuangzhou Wang, Shengsheng College of Computer Science and Technology Jilin University Changchun130012 China Key Laboratory of Symbolic Computation and Knowledge Engineering of Ministry of Education Jilin University Changchun130012 China

Accurate intervertebral disc image segmentation is necessary for further treatment. However, existing methods are difficult to segment due to the intensity inhomogeneity of intervertebral disc MRI images and the similarity in intensity between the intervertebral disc and the surrounding regions. This paper proposes a novel level set model by introducing a shape prior constraint for automatic and accurate segmentation. We utilize a detection model and intensity change to obtain rectangle prior information, which is used to define a shape constraint term. Then, combining data and regularizing terms construct an energy function and employ the steepest descent method minimization to get segmentation results. Finally, we apply the proposed new method to segment intervertebral disc images and get promising results. Extensive experiments and comparisons show that our model is more efficient and accurate. © The Author(s), under exclusive licence to Springer Science+Business Media, LLC, part of Springer Nature 2024.

关键词： Image segmentation

来源：评论

学校读者我要写书评

暂无评论

LIOF: Make the Learned Index Learn Faster With Higher Accuracy

引用

IEEE Transactions on knowledge and Data engineering 2025年第6期37卷 3499-3513页

作者： Ji, Tao Zhong, Kai Sun, Luming Li, Yiyan Li, Cuiping Chen, Hong Ministry of Education Key Laboratory of Data Engineering and Knowledge Engineering Beijing100872 China Renmin University of China School of Information Beijing100872 China Yunxi Technology Company Ltd. Shanghai200120 China

Learned indexes, emerging as a promising alternative to traditional indexes like B+Tree, utilize machine learning models to enhance query performance and reduce memory usage. However, the widespread adoption of learned indexes is limited by their expensive training cost and the need for high accuracy of internal models. Although some studies attempt to optimize the building process of these learned indexes, existing methods are restrictive in scope and applicability. They are usually tailored to specific index types and heavily rely on pre-trained model knowledge, making deployment a challenging task. In this work, we introduce the Learned Index Optimization Framework (LIOF), a general and easily integrated solution aimed at expediting the training process and improving the accuracy of index model for one-dimensional and multi-dimensional learned indexes. The optimization of LIOF for the learned indexes is intuitive, directly providing optimized parameters for index models based on the distribution of node data. By leveraging the correlation between key distribution and node model parameters, LIOF significantly reduces the training epochs required for each node model. Initially, we introduce an optimization strategy inspired by optimization-based meta-learning to train the LIOF to generate optimized initial parameters for index node models. Subsequently, we present a data-driven encoder and a parameter-centric decoder network, which adaptively translate key distribution into a latent variable representation and decode it into optimized node model initialization. Additionally, to further utilize characteristics of key distribution, we propose a monotonic regularizer and focal loss, guiding LIOF training towards efficiency and precision. Through extensive experimentation on real-world and synthetic datasets, we demonstrate that LIOF provides substantial enhancements in both training efficiency and the predictive accuracy for learned indexes. © 1989-2012 IEEE.

关键词： Decoding

来源：评论

学校读者我要写书评

暂无评论

Distributed Leader-Following Formation Control of Networked Mobile Robots via Global Orientation Estimation

引用

IEEE Transactions on Network Science and engineering 2025年第3期12卷 2135-2150页

作者： Wang, Siqi Wang, Heng Che, Weiwei Li, Qing University of Science and Technology Beijing Key Laboratory of Knowledge Automation for Industrial Processes of Ministry of Education School of Automation and Electrical Engineering Beijing100083 China Northeastern University College of Information Science and Engineering Shenyang110819 China

This article proposes a novel distributed leader-following formation control strategy for multiple networked mobile robots, utilizing the relative position measurements among robots. In particular, the leader is assumed to have the same kinematics as the followers and all the robots do not rely on orientation measurements. Firstly, a global orientation estimation law is proposed in the sense that the followers' orientation information is estimated in the leader's reference frame, only based on the leader's orientation estimation and relative bearing information. Secondly, since the leader is not directly connected to all the followers, a new distributed state observer is designed for each follower to estimate the leader's states. Especially, the designed observer not only removes the algebraic loops issue but also eliminates the requirement for the leader's acceleration information. Furthermore, a distributed formation control law is proposed by incorporating the previous estimations, and it is proved that the closed-loop system consisting of the observer and controller is asymptotically stable. Finally, simulation results validate the effectiveness and superiority of the proposed method. © 2013 IEEE.

关键词： Observers Formation control Estimation Robot sensing systems Position measurement Robot kinematics Mobile robots Topology Decentralized control Accuracy

来源：评论

学校读者我要写书评

暂无评论

A multi-strategy enhanced circle search algorithm based on adaptive evolutionary framework and its application in image segmentation

引用

Cluster Computing 2025年第5期28卷 1-46页

作者： Wang, Wei Lu, Yinan Wang, Shengsheng College of Computer Science and Technology Jilin University Jilin Changchun130012 China Key Laboratory of Symbolic Computation and Knowledge Engineering of Ministry of Education Jilin University Jilin Changchun130012 China

Circle search algorithm (CSA) is a meta-heuristic optimization algorithm based on the characteristics of circle geometry. In order to overcome the shortcomings of the original CSA in the optimization process, such as low convergence accuracy and easily falling into local optimal, this paper proposed a multi-strategy enhanced CSA named APRDCSA. Firstly, a mutation strategy guided by multiple superior solutions is proposed, which increases the diversity of the population and the global exploration ability of the algorithm. Secondly, random replacement strategy is introduced in the late iteration to avoid falling into local optima. Finally, a population evolution state assessment framework with adaptive parameter (APSE) based on improvement rate is proposed to improve the probability of finding a better solution. A series of comparative experiments on IEEE CEC 2014 demonstrate the superior convergence speed and solution quality of APRDCSA, as well as its enhanced reliability in avoiding local optima. In addition, to further illustrate the improved performance of APRDCSA and its ability to solve practical applications, we apply it to the multi-thresholding image segmentation (MIS) method based on non-local mean two-dimensional histogram and Kapur entropy. Experimental results show that the proposed method achieves better segmentation results at both low and high threshold levels. © 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

PBCounter:weighted model counting on pseudo-boolean formulas

引用

Frontiers of Computer Science 2025年第3期19卷 55-63页

作者： Yong LAI Zhenghang XU Minghao YIN College of Computer Science and Technology Jilin UniversityChangchun 130012China Key Laboratory of Symbolic Computation and Knowledge Engineering of Ministry of Education Jilin UniversityChangchun 130012China School of Computer Science and Information Technology Northeast Normal UniversityChangchun 130017China

In Weighted Model Counting(WMC),we assign weights to literals and compute the sum of the weights of the models of a given propositional formula where the weight of an assignment is the product of the weights of its *** current WMC solvers work on Conjunctive Normal Form(CNF)***,CNF is not a natural representation for human-being in many *** by the stronger expressive power of Pseudo-Boolean(PB)formulas than CNF,we propose to perform WMC on PB *** on a recent dynamic programming algorithm framework called ADDMC for WMC,we implement a weighted PB counting tool *** compare PBCounter with the state-of-the-art weighted model counters SharpSAT-TD,ExactMC,D4,and ADDMC,where the latter tools work on CNF with encoding methods that convert PB constraints into a CNF *** experiments on three domains of benchmarks show that PBCounter is superior to the model counters on CNF formulas.

关键词： weighted model counting pseudo-boolean constraint algebraic decision diagrams preprocessing techniques

来源：评论

学校读者我要写书评

暂无评论

An Evolutionary Multitasking Algorithm for Efficient Multiobjective Recommendations

IEEE Transactions on Artificial Intelligence

引用

IEEE Transactions on Artificial Intelligence 2025年第3期6卷 518-532页

作者： Tian, Ye Ji, Luke Hu, Yiwei Ma, Haiping Wu, Le Zhang, Xingyi Anhui University Key Laboratory of Intelligent Computing and Signal Processing of Ministry of Education School of Computer Science and Technology Hefei230601 China Anhui University Institutes of Physical Science and Information Technology Hefei230601 China Hefei University of Technology Key Laboratory of Knowledge Engineering with Big Data Hefei230029 China

Represented by evolutionary algorithms and swarm intelligence algorithms, nature-inspired metaheuristics have been successfully applied to recommender systems and amply demonstrated effectiveness, in particular, for multiobjective recommendation. Owing to the population-based search paradigm, these algorithms can produce a number of recommendation lists, making diverse tradeoffs between multiple metrics and meeting the requirements of accuracy, novelty, diversity, and other user preferences. However, these algorithms are criticized for the low efficiency of the optimization process, especially when the number of users is large. To address this issue, this article proposes an evolutionary multitasking-based recommendation method, where each task corresponds to a user and all the tasks are optimized simultaneously, thus highly improving the efficiency of recommendation. To enhance the convergence speed, all the users are divided into multiple populations according to the similarity between their preferences, where each population evolves with internal knowledge transfer between users, and all the populations evolve with external knowledge transfer between populations. Experimental results on various datasets verify that the proposed method can better balance between multiple metrics than classical and deep neural network-based recommendation methods and exhibits significantly higher efficiency than evolutionary multiobjective optimization-based recommendation methods. © 2024 IEEE. All rights reserved.

关键词： Multiobjective optimization

来源：评论

学校读者我要写书评

暂无评论

Orthogonal Projection-Based Fault Detection for Linear Discrete-Time Varying Systems

引用

IEEE Transactions on Automatic Control 2025年第5期70卷 3478-3485页

作者： Li, Linlin Ding, Steven X. Zhong, Maiying Peng, Kaixiang University of Science and Technology Beijing Key Laboratory of Knowledge Automation for Industrial Processes of Ministry of Education School of Automation and Electrical Engineering Beijing100083 China Duisburg47057 Germany Shandong University of Science and Technology College of Electrical Engineering and Automation Qingdao266590 China

This article is devoted to one-class fault detection in linear discrete-time varying (LDTV) systems with uncertainties. Specifically, following the Hilbert Projection theorem, the residual generation problem is solved by means of an orthogonal projection of process data onto system subspaces in Hilbert space. The resulted Pythagorean equation enables establishing an adaptive threshold driven by the residual signal and using gap metric as the similarity measure. In this projection-based fault detection framework, the design and implementation of the residual generators and threshold setting can be uniformly addressed, which provides a systematic solution to the detection issues in linear discrete-time varying systems. Moreover, a comparison with the observer- and differential-algebraic equations-based fault detection schemes is studied, which highlights the improved fault detection capability of the projection-based method. © 1963-2012 IEEE.

关键词： Linear time-invariant system

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：