检索结果-内蒙古大学图书馆

15th IEEE International Conference on Knowledge Graph, ICKG 2024

作者： Fang, Runnan Qiao, Shuofei Xi, Zekun Zhejiang University Department of Software Technology Hangzhou China Zhejiang University Department of Computer Science Hangzhou China

ISBN: (纸本)9798331508821

Guideline Knowledge is helpful for LLM (Large Language Model) based Agent in embody task planning. In this work, we introduce Textgrad to optimize the Guideline Knowledge for the agent's embodied tasks. This allows the model to automatically optimize the Guideline Knowledge by calculating the text gradients in the Guideline Knowledge and analyzing the loss in failed trajectories. © 2024 IEEE.

关键词： Chatbots

来源：评论

学校读者我要写书评

暂无评论

How far are we to GPT-4V?Closing the gap to commercial multimodal models with open-source suites

引用

Science China(Information Sciences) 2024年第12期67卷 5-22页

作者： Zhe CHEN Weiyun WANG Hao TIAN Shenglong YE Zhangwei GAO Erfei CUI Wenwen TONG Kongzhi HU Jiapeng LUO Zheng MA Ji MA Jiaqi WANG Xiaoyi DONG Hang YAN Hewei GUO Conghui HE Botian SHI Zhenjiang JIN Chao XU Bin WANG Xingjian WEI Wei LI Wenjian ZHANG Bo ZHANG Pinlong CAI Licheng WEN Xiangchao YAN Min DOU Lewei LU Xizhou ZHU Tong LU Dahua LIN Yu QIAO Jifeng DAI Wenhai WANG State Key Laboratory for Novel Software Technology Nanjing University Shanghai AI Laboratory School of Computer Science Fudan University SenseTime Research Department of Information Engineering The Chinese University of Hong Kong Department of Electronic Engineering Tsinghua University

In this paper, we introduce InternVL 1.5, an open-source multimodal large language model(MLLM) to bridge the capability gap between open-source and proprietary commercial models in multimodal understanding. We introduce three simple improvements.(1) Strong vision encoder: we explored a continuous learning strategy for the large-scale vision foundation model — InternViT-6B, boosting its visual understanding capabilities, and making it can be transferred and reused in different LLMs.(2) Dynamic high-resolution: we divide images into tiles ranging from 1 to 40 of 448×448 pixels according to the aspect ratio and resolution of the input images, which supports up to 4K resolution input.(3) High-quality bilingual dataset: we carefully collected a high-quality bilingual dataset that covers common scenes, document images,and annotated them with English and Chinese question-answer pairs, significantly enhancing performance in optical character recognition(OCR) and Chinese-related tasks. We evaluate InternVL 1.5 through a series of benchmarks and comparative studies. Compared to both open-source and proprietary commercial models, InternVL 1.5 shows competitive performance, achieving state-of-the-art results in 8 of 18 multimodal benchmarks. Code and models are available at https://***/OpenGVLab/InternVL.

关键词： multimodal model open-source vision encoder dynamic resolution bilingual dataset

来源：评论

学校读者我要写书评

暂无评论

MSEs Credit Risk Assessment Model Based on Federated Learning and Feature Selection

引用

computers, Materials & Continua 2023年第6期75卷 5573-5595页

作者： Zhanyang Xu Jianchun Cheng Luofei Cheng Xiaolong Xu Muhammad Bilal School of Computer and Software Nanjing University of Information Science&TechnologyNanjing210044China State Key Laboratory Novel Software Technology Nanjing UniversityNanjing210023China Department of Computer and Electronics Systems Engineering Hankuk University of Foreign StudiesYongin-siGyeonggi-do17035Korea

Federated learning has been used extensively in business inno-vation scenarios in various *** research adopts the federated learning approach for the first time to address the issue of bank-enterprise information asymmetry in the credit assessment ***,this research designs a credit risk assessment model based on federated learning and feature selection for micro and small enterprises(MSEs)using multi-dimensional enterprise data and multi-perspective enterprise *** proposed model includes four main processes:namely encrypted entity alignment,hybrid feature selection,secure multi-party computation,and global model ***,a two-step feature selection algorithm based on wrapper and filter is designed to construct the optimal feature set in multi-source heterogeneous data,which can provide excellent accuracy and *** addition,a local update screening strategy is proposed to select trustworthy model parameters for aggregation each time to ensure the quality of the global *** results of the study show that the model error rate is reduced by 6.22%and the recall rate is improved by 11.03%compared to the algorithms commonly used in credit risk research,significantly improving the ability to identify ***,the business operations of commercial banks are used to confirm the potential of the proposed model for real-world implementation.

关键词： Federated learning feature selection credit risk assessment MSEs

来源：评论

学校读者我要写书评

暂无评论

CK-index: A Distribution-Aware Learned Index for Composite Keys 22

CK-index: A Distribution-Aware Learned Index for Composite K...

引用

22nd IEEE International Symposium on Parallel and Distributed Processing with Applications, ISPA 2024

作者： Wei, Zhengyang Ye, Baoliu Cai, Miao Hohai University College of Computer and Software China Nanjing University Department of Computer Science and Technology China Nanjing University of Aeronautics and Astronautics College of Computer Science and Technology China

ISBN: (纸本)9798331509712

The learned index is a high-performance index structure that uses machine learning methods to predict key positions in a large key space efficiently. Existing learned indexes suffer from underfitting of key-to-position mapping, leading to poor lookup performance. This paper finds that a data distribution property in the widely-used composite key schema addresses this issue effectively. Specifically, the composite key consists of an agglomerate of attributes. Keys with the same attribute value have a regular data distribution, which leads to a higher fitness of key-to-position mapping. Applying the property, we introduce CK-index, a distribution-aware learned index for composite keys. CK-index divides the key space according to attribute values and trains each learned model separately for an attribute to achieve high fitness of key-to-position mapping. Furthermore, it achieves low data storage consumption via storing composite key's attributes instead of the whole keys. We evaluate the CK-index using real-world datasets. Evaluation results demonstrate that CK-index performs much better in lookup performance, bulk loading time and space consumption compared to B+Tree, RMI, PGM-index and ALEX. © 2024 IEEE.

关键词： Composite key Data distribution Learned index

来源：评论

学校读者我要写书评

暂无评论

Optimized digital watermarking: Harnessing the synergies of Schur matrix factorization, DCT, and DWT for superior image ownership proofing

引用

Multimedia Tools and Applications 2025年第18期84卷 19817-19852页

作者： Al-Aiash, Issa Alquran, Rabee AlJamal, Mahmoud Alsarhan, Ayoub Aljaidi, Mohammad Al-Fraihat, Dimah Department of Information Technology Faculty of Prince Al-Hussein Bin Abdallah II for Information Technology The Hashemite University Zarqa13133 Jordan Department of Computer Science Faculty of Information Technology Zarqa University Zarqa13110 Jordan Department of Software Engineering Faculty of Information Technology Isra University Amman11622 Jordan

The contemporary context of abundant digital dissemination inherently gives rise to the need for media protection and the clear identification of ownership rights. This paper responded to this nagging issue of unauthorized media capture by developing a new digital watermarking approach uniting the Discrete Wavelet Transform DWT, Discrete Cosine Transform DCT, and Schur matrix factorization. The proposed watermark, which was a positive vector uniformly distributed, presented a high degree of robustness to several adversarial operating conditions, including histogram equalization, salt and pepper noise, ripple impact, smoothing attack, and cropping. The obtained results reveal the effectiveness of our approach. The correlation coefficient between watermarked and original images remains exceptionally high at 0.999998, ensuring minimal visual distortion. Mean Squared Error (MSE) graphs for both high-frequency (HH) and low-frequency (LL) domains demonstrate negligible changes after attacks, confirming the watermark's robustness. Notably, the watermark extraction process remains successful even after histogram equalization, as indicated by significantly lower MSE compared to random vectors. Moreover, the average processing times for different attacks are promising;Histogram Equalization and Salt and Pepper attacks each took 0.3 s, Ripple Attack required 0.28 s, Smoothing Attack completed in 0.29 s, and the Cropping Attack took 0.6 s. The robustness of our approach appears against the salt and pepper attack, which historically presented a vulnerability when employing Schur matrix factorization in image watermarking, indicating the novelty of and this contribution to improving media ownership verification. By combining disparate transformation methodologies, this technique offers a promising avenue for reinforcing the integrity and authenticity of digital media, thereby strengthening the foundations of intellectual property rights in the digital domain. © The Author(s), und

关键词： Discrete wavelet transforms

来源：评论

学校读者我要写书评

暂无评论

Unveiling the Power of Deep Learning: A Comparative Study of LSTM, BERT, and GRU for Disaster Tweet Classification

引用

IEIE Transactions on Smart Processing and Computing 2023年第6期12卷 526-534页

作者： Ullah, Ihsan Jamil, Anum Ul Hassan, Imtiaz Kim, Byung-Seo Department of Software and Communication Engineering Hongik University Korea Republic of Department of Physics NED University of Engineering & Technology Karachi Pakistan Department of Computer Science and Information Technology NED University of Engineering & Technology Karachi Pakistan

Disasters have serious effects on people's lives and buildings. Therefore, social media platforms, such as Twitter, have become more critical. They are crucial tools for responding to and managing disasters effectively. This study examined the effectiveness of various deep learning models, such as bidirectional encoder representations from transformers (BERT), gated recurrent units (GRU), and long short-term memory (LSTM) for classifying disaster-related tweets. Twitter data related to different disasters were collected using hashtags. The data were then cleaned, preprocessed, and manually annotated by a team. The annotated data were divided into training, validation, and testing sets. The data were used to train three models based on BERT, GRU, and LSTM for the categorical classification of disaster tweets. Finally, the three models were evaluated and compared using the test data. BERT achieved an accuracy of 96.2%, making it the most effective model. In contrast, the LSTM and GRU models achieved an accuracy of 93.2% and 88.4%, respectively. These findings underscore the potential effectiveness of deep learning models in classifying disaster-related tweets, offering insights that could enhance disaster management strategies, refine social media monitoring processes, bolster public safety, and provide directions for future research. © 2023 Institute of Electronics and Information Engineers. All rights reserved.

关键词： Disasters

来源：评论

学校读者我要写书评

暂无评论

Offshore software Maintenance Outsourcing Process Model Validation:A Case Study Approach

引用

computers, Materials & Continua 2023年第3期74卷 5035-5048页

作者： Atif Ikram Masita Abdul Jalil Amir Bin Ngah Adel Sulaiman Muhammad Akram Ahmad Salman Khan Faculty of Ocean Engineering Technology and Informatics University Malaysia TerengganuKuala TerengganuMalaysia Department of Computer Science&Information Technology Faculty of Information TechnologyThe University of LahoreLahore54000Pakistan Department of Computer Science College of Computer Science and Information SystemsNajran UniversityNajranSaudi Arabia Department of Software Engineering Faculty of Information TechnologyThe University of LahoreLahore54000Pakistan

The successful execution and management of Offshore software Maintenance Outsourcing(OSMO)can be very beneficial for OSMO vendors and the OSMO *** a lot of research on software outsourcing is going on,most of the existing literature on offshore outsourcing deals with the outsourcing of software development *** frameworks have been developed focusing on guiding software systemmanagers concerning offshore software ***,none of these studies delivered comprehensive guidelines for managing the whole process of *** is a considerable lack of research working on managing OSMO from a vendor’s ***,to find the best practices for managing an OSMO process,it is necessary to further investigate such complex and multifaceted phenomena from the vendor’s *** study validated the preliminary OSMO process model via a case study research *** results showed that the OSMO process model is applicable in an industrial setting with few *** industrial data collected during the case study enabled this paper to extend the preliminary OSMO process *** refined version of the OSMO processmodel has four major phases including(i)Project Assessment,(ii)SLA(iii)Execution,and(iv)Risk.

关键词： Offshore outsourcing process model model validation vendor challenges case study

来源：评论

学校读者我要写书评

暂无评论

A review of generative models for virtual human motion driving

引用

Science China(Information Sciences) 2025年

作者： Pengcheng WANG Haolun PENG Siyang MAO Shaojiang WANG Jie TANG Department of Computer Science and Technology Tsinghua University School of Artificial Intelligence Nanjing University Nanjing Institute of Software Technology School of Computer Science and Technology Beijing Institute of Technology Institute of Software Chinese Academy of Sciences Institute of Industrial Artificial Intelligence Chinese Academy of Sciences

Virtual human motion driving focuses on generating and controlling realistic human motions, from facial expressions to body movements. These motions are driven by various types of input signals, such as visual and acoustic features,textual prompts, or a combination thereof. This survey delivers an in-depth examination of generative models for virtual human motion driving, with a specific emphasis on recent models. A taxonomy of virtual human motion driving networks designed for talking-face and human-pose generation is provided. The former mainly concentrates on lip synchronization,differentiation of emotions, and personalized expressions, while the latter mainly includes co-speech gesture generation and text-to-motion prediction. Moreover, available datasets and evaluation metrics for virtual human motion driving tasks are discussed, applications and real products related to virtual human motion driving are explored, along with their challenges,limitations, and potential future developments. The objective of this survey is to gain a comprehensive understanding of the present advancements in talking-face and human-pose generation models, with a focus on the future potential of virtual human motion driving. This endeavor aims to lay the groundwork for the development of extensive applications for virtual humans.

关键词：

来源：评论

学校读者我要写书评

暂无评论

Indexing dynamic encrypted database in cloud for efficient secure k-nearest neighbor query

引用

Frontiers of computer Science 2024年第1期18卷 187-198页

作者： Xingxin LI Youwen ZHU Rui XU Jian WANG Yushu ZHANG College of Computer Science and Technology Nanjing University of Aeronautics and AstronauticsNanjing 210016China Department of Mathematical Informatics University of TokyoTokyo 113-8654Japan Guangxi Key Laboratory of Trusted Software Guilin University of Electronic TechnologyGuilin 541004China School of Computer Science China University of GeosciencesWuhan 430074China

Secure k-Nearest Neighbor(k-NN)query aims to find k nearest data of a given query from an encrypted database in a cloud server without revealing privacy to the untrusted cloud and has wide applications in many areas,such as privacy-preservingmachine elearning gand secure biometric *** solutions have been put forward to solve this challenging ***,the existing schemes still suffer from various limitations in terms of efficiency and *** this paper,we propose a new encrypt-then-index strategy for the secure k-NN query,which can simultaneously achieve sub-linear search complexity(efficiency)and support dynamical update over the encrypted database(flexibility).Specifically,we propose a novel algorithm to transform the encrypted database and encrypted query points in the *** indexing the transformed database using spatial data structures such as the R-tree index,our strategy enables sub-linear complexity for secure k-NN queries and allows users to dynamically update the encrypted *** the best of our knowledge,the proposed strategy is the first to simultaneously provide these two *** theoretical analysis and extensive experiments,we formally prove the security and demonstrate the efficiency of our scheme.

关键词： cloud computing secure k-NN query sub-linear complexity dynamicupdate

来源：评论

学校读者我要写书评

暂无评论

Efficient Federated Learning Using Dynamic Update and Adaptive Pruning with Momentum on Shared Server Data

引用

ACM Transactions on Intelligent Systems and technology 2024年第6期15卷 1-28页

作者： Liu, Ji Jia, Juncheng Zhang, Hong Yun, Yuhui Wang, Leye Zhou, Yang Dai, Huaiyu Dou, Dejing Hithink RoyalFlush Information Network Co. Ltd. Hangzhou China School of Computer Science and Technology Soochow University Suzhou China Collaborative Innovation Center of Novel Software Technology and Industrialization Nanjing China Baidu Research Beijing China Key Lab of High Confidence Software Technologies Ministry of Education and Software Institute Peking University Beijing China Department of Computer Science and Software Engineering Auburn University AuburnAL United States Department of Electrical and Computer Engineering North Carolina State University RaleighNC United States Bedi Cloud and School of Computer Science Fudan University Beijing China

Despite achieving remarkable performance, Federated Learning (FL) encounters two important problems, i.e., low training efficiency and limited computational resources. In this article, we propose a new FL framework, i.e., FedDUMAP, with three original contributions, to leverage the shared insensitive data on the server in addition to the distributed data in edge devices so as to efficiently train a global model. First, we propose a simple dynamic server update algorithm, which takes advantage of the shared insensitive data on the server while dynamically adjusting the update steps on the server in order to speed up the convergence and improve the accuracy. Second, we propose an adaptive optimization method with the dynamic server update algorithm to exploit the global momentum on the server and each local device for superior accuracy. Third, we develop a layer-adaptive model pruning method to carry out specific pruning operations, which is adapted to the diverse features of each layer so as to attain an excellent tradeoff between effectiveness and efficiency. Our proposed FL model, FedDUMAP, combines the three original techniques and has a significantly better performance compared with baseline approaches in terms of efficiency (up to 16.9 times faster), accuracy (up to 20.4% higher), and computational cost (up to 62.6% smaller). © 2024 Copyright held by the owner/author(s). Publication rights licensed to ACM.

关键词： Federated learning

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：