检索结果-内蒙古大学图书馆

31st International Conference on Computational Linguistics, COLING 2025

作者： Hou, Haowen Zeng, Peigen Ma, Fei Yu, Fei Richard Shenzhen China College of Computer Science and Software Engineering Shenzhen University Shenzhen China Shool of Information Technology Carleton University Canada

ISBN: (纸本)9798891761964

Visual Language Models (VLMs) have rapidly progressed with the recent success of large language models. However, there have been few attempts to incorporate efficient linear Recurrent Neural Networks (RNNs) architectures into VLMs. In this study, we introduce VisualRWKV, the first application of a linear RNN model to multimodal learning tasks, leveraging the pre-trained RWKV language model. We propose a data-dependent recurrence and sandwich prompts to enhance our modeling capabilities, along with a 2D image scanning mechanism to enrich the processing of visual sequences. Extensive experiments demonstrate that VisualRWKV achieves competitive performance compared to Transformer-based models like LLaVA-1.5 on various benchmarks. Compared to LLaVA-1.5, VisualRWKV has a speed advantage of 3.98 times and can save 54% of GPU memory when reaching an inference length of 24K tokens. To facilitate further research and analysis, we have made the checkpoints and the associated code publicly accessible at the following GitHub repository: https://***/howard-hou/VisualRWKV. © 2025 Association for Computational Linguistics.

关键词： Recurrent neural networks

来源：评论

学校读者我要写书评

暂无评论

VisualRWKV: Exploring Recurrent Neural Networks for Visual Language Models

arXiv

引用

arXiv 2024年

关键词： Recurrent neural networks

来源：评论

学校读者我要写书评

暂无评论

Decoupling of Direction and Length for Cumulative Step Size Adaptation 17

Decoupling of Direction and Length for Cumulative Step Size ...

引用

17th International Conference on Computational Intelligence and Security, CIS 2021

作者： Zhang, Shuo Li, Zhenhua Yang, Detian Wang, Shuo Cai, Xinye School of Computer Science and Technology Nanjing University of Aeronautics and Astronautics Nanjing China Shool of Information Technology and Engineering Tianjin University of Technology and Education Tianjin China

ISBN: (纸本)9781665494892

Covariance matrix adaptation evolution strategy (CMA-ES) is a successful evolutionary algorithm for continuous optimization. The cumulative step size adaptation (CSA) is the default method for adapting the step size in CMA-ES and achieves remarkable performance in practice. It follows the idea of exploiting the correlations between search directions in consecutive generations. In this paper, we investigate the effects of the update directions and the length in CSA. We find that sampling from uniform sphere distribution can yield comparative performance to the default setting, and normalizing the update direction in CSA does not hurt the performance. Further, the CSA with actual search direction cannot perform well, and the correlation between the default direction of CSA and the actual search direction is rather small. It indicates that CSA does not accumulate the actual search directions. We argue that the step size and the covariance matrix should be updated by decoupled search information, and accumulating the actual search direction results in inferior performance due to the same update direction in two evolution paths for the step size and covariance matrix updates. © 2021 IEEE.

关键词： Covariance matrix

来源：评论

学校读者我要写书评

暂无评论

computer-aided Visual Modeling of Rice Leaf Growth Based on Machine Learning 23

Computer-aided Visual Modeling of Rice Leaf Growth Based on ...

引用

23rd International Conference on Soft Computing and Measurements, SCM 2020

作者： Yi, Wenlong Dai, Shiming Jiang, Yingzhao Yuan, Chao Yang, Le Jiangxi Agricultural University School of Software Engineering Nanchang China Jiangxi Agricultural University Shool of Computer and Information Engineering Nanchang China

ISBN: (纸本)9781728196923

Machine learning can abstract the complex plant growth and development into a high-dimensional feature space, and transform a complex biological process into a mathematical problem. In this paper, with the computer-aided modeling of rice leaf growth as an example, improving the accuracy of prediction models for external environmental factors of rice growth and parameters of leaf shape evolution is investigated. Two machine learning tools, SVR and CNN, are selected to compare and analyze the training and prediction errors of 709 collected sample data. The experimental results show that the prediction accuracy of CNN is about 2 times higher than that of SVR. However, the learning speed of SVR in solving small sample regression is 50 times higher than that of CNN. Finally, the obtained parameters for rice leaf growth shape prediction are subjected to geometric description using the B-spline function, and visual simulation is carried out by visual C++ programming language and OpenGL 3.2. Three-dimensional visual models of plant growth and development with an external growth environment are established using machine learning. The ideal plant morphology is obtained by adjusting external environmental factors reasonably through quantitative analysis. It provides an information tool for the transformation of traditional empirical agricultural production to precise mode. © 2020 IEEE.

关键词： Machine learning

来源：评论

学校读者我要写书评

暂无评论

Rethinking Quality of Experience for Metaverse Services: A Consumer-based Economics Perspective

arXiv

引用

arXiv 2022年

作者： Du, Hongyang Ma, Bohao Niyato, Dusit Kang, Jiawen Xiong, Zehui Yang, Zhaohui The School of Computer Science and Engineering Nanyang Technological University Singapore The Shool of Civil and Environmental Engineering Nanyang Technological University Singapore The School of Automation Guangdong University of Technology China The Pillar of Information Systems Technology and Design Singapore University of Technology and Design Singapore The College of Information Science and Electronic Engineering Zhejiang University China

The Metaverse is considered to be one prototype of the next-generation Internet, which contains people’s expectations for the future world. However, the academic discussion of the Metaverse still mainly focused on the technical system design, and few research studied Metaverse challenges from the perspective of consumers, i.e., Metaverse users. One difficulty is that the analysis from the consumer’s perspective requires interdisciplinary theoretical framework and quantifiable Quality of Experience (QoE) measurements. In this article, pioneering from consumers’ point of view, we explore an interaction between Metaverse system design and consumer behaviors. Specifically, we rethink QoE and propose an framework that encompasses both the Metaverse service providers (MSPs) and consumer considerations. From the macro perspective, we introduce a joint optimization scheme that simultaneously considers the Metaverse system design, consumers’ utility, and profitability of the MSPs. From the micro perspective, we advocate the Willingness-to-Pay (WTP) as an easy-to-implement measurement for future Metaverse system studies. To illustrate the usability of the proposed integrated framework, a use case, i.e., virtual traveling, is presented. We show that our framework can benefit the MSPs in offering competitive and economical service design to consumers while maximizing the profit. © 2022, CC BY.

关键词： Consumer behavior

来源：评论

学校读者我要写书评

暂无评论

Decoupling of Direction and Length for Cumulative Step Size Adaptation

Decoupling of Direction and Length for Cumulative Step Size ...

引用

International Conference on Computational Intelligence and Security

作者： Shuo Zhang Zhenhua Li Detian Yang Shuo Wang Xinye Cai School of Computer Science and Technology Nanjing University of Aeronautics and Astronautics Nanjing China Shool of Information Technology and Engineering Tianjin University of Technology and Education Tianjin China

ISBN: (纸本)9781665494908

关键词： Correlation Evolutionary computation Security Covariance matrices Optimization

来源：评论

学校读者我要写书评

暂无评论

computer-aided Visual Modeling of Rice Leaf Growth Based on Machine Learning

Computer-aided Visual Modeling of Rice Leaf Growth Based on ...

引用

International Conference on Soft Computing and Measurements (SCM)

作者： Wenlong Yi Shiming Dai Yingzhao Jiang Chao Yuan Le Yang School of Software Engineering Jiangxi Agricultural University Nanchang China Shool of Computer and Information Engineering Jiangxi Agricultural University Nanchang China

ISBN: (数字)9781728196923

ISBN: (纸本)9781728196930

关键词： Biological system modeling Predictive models Computational modeling Machine learning Image color analysis Data models Training

来源：评论

学校读者我要写书评

暂无评论

Two Algorithms Analysing Discharge Parameters Based on Neural-network and Wavelet Transformer

Two Algorithms Analysing Discharge Parameters Based on Neura...

引用

EMC/SI/PI and EMC Europe Symposium, IEEE International Joint

作者： Fangming Ruan Sheng Guan Yang Meng Lan Yin Yanli Chen Kui Zhou School of big data and computer science Guizhou Normal University Guiyang China Shool of big data and information engineering Guizhou University Guiyang China Shandong Computer Science Center (National Supercomputing Center in Jinan) Qilu University of Technology (Shandong Academy of Sciences) Jinan China School of big data and computer Guizhou Normal University Guiyang China

Special relationship exists between environmental conditions and discharge characteristic parameters in electrostatic discharge (ESD) events. The neural network can explore the potential law between input and output if taken discharge condition parameters as a neural network input. Characteristics of discharge results are affected by environmental conditions, and hence discharge parameters can be described with the output of a neural network. Two algorithms of artificial intelligence were used to analyzing discharge results in electrostatic discharge. The nonlinear relationship between discharge conditions and discharge effect may be a new potential discharge feature. Noise in discharge current can be suppressed with wavelet and Kalman filter method. The characteristics measured in the real experiment were compared with the prediction parameters from the neuro-network calculation result. According to the prediction data, the discussion was conducted on correctness accuracy and the discharge process trend analysis.

关键词： Electrodes Electric potential Current measurement Neural networks Electrostatic measurements Electrostatic discharges Prediction algorithms

来源：评论

学校读者我要写书评

暂无评论

arXiv

引用

arXiv 2019年

作者： Chen, Shiming Zhang, Peng Xie, Guo-Sen Peng, Qinmu Cao, Zehong Yuan, Wei You, Xinge The Department of Electronic Information and Communication Huazhong University of Science and Technology Wuhan430074 China The Shool of Computer Science and Engineering Nanjing University of Science and Technology Nanjing China STEM The University of South Australia Adelaide Australia

Dynamic texture (DT) exhibits statistical stationarity in the spatial domain and stochastic repetitiveness in the temporal dimension, indicating that different frames of DT possess a high similarity correlation that is critical prior knowledge. However, existing methods cannot effectively learn a synthesis model for high-dimensional DT from a small number of training samples. In this paper, we propose a novel DT synthesis method, which makes full use of similarity as prior knowledge to address this issue. Our method is based on the proposed kernel similarity embedding, which can not only mitigate the high-dimensionality and small sample issues, but also has the advantage of modeling nonlinear feature relationships. Specifically, we first put forward two hypotheses that are essential for the DT model to generate new frames using similarity correlations. Then, we integrate kernel learning and the extreme learning machine into a unified synthesis model to learn kernel similarity embeddings for representing DTs. Extensive experiments on DT videos collected from the internet and two benchmark datasets, i.e., Gatech Graphcut Textures and Dyntex, demonstrate that the learned kernel similarity embeddings can provide discriminative representations for DTs. Further, our method can preserve the long-term temporal continuity of the synthesized DT sequences with excellent sustainability and generalization. Meanwhile, it effectively generates realistic DT videos with higher speed and lower computation than the current state-of-the-art methods. The code and more synthesis videos are available at our project page https://***/Similarity-page/***. Copyright © 2019, The Authors. All rights reserved.

关键词： Textures

来源：评论

学校读者我要写书评

暂无评论

SHREC 2019 - Monocular image based 3D model retrieval 12

SHREC 2019 - Monocular image based 3D model retrieval

引用

12th Eurographics Workshop on 3D Object Retrieval, 3DOR 2019, in conjunction with the 40th Annual Conference of the European Association for computer Graphics, EG 2019

作者： Li, Wenhui Liu, Anan Nie, Weizhi Song, Dan Li, Yuqian Wang, Weijie Xiang, Shu Zhou, Heyu Bui, Ngoc-Minh Cen, Yunchi Chen, Zenian Chung-Nguyen, Huy-Hoang Diep, Gia-Han Do, Trong-Le Doubrovski, Eugeni L. Duong, Anh-Duc Geraedts, Jo M.P. Guo, Haobin Hoang, Trung-Hieu Li, Yichen Liu, Xing Liu, Zishun Luu, Duc-Tuan Ma, Yunsheng Nguyen, Vinh-Tiep Nie, Jie Ren, Tongwei Tran, Mai-Khiem Tran-Nguyen, Son-Thanh Tran, Minh-Triet Vu-Le, The-Anh Wang, Charlie C.L. Wang, Shijie Wu, Gangshan Yang, Caifei Yuan, Meng Zhai, Hao Zhang, Ao Zhang, Fan Zhao, Sicheng Shool of Electrical and Information Engineering Tianjin University China University of Science VNU-HCM Viet Nam State Key Laboratory of Virtual Reality Technology and System Beihang University China Delft University of Technology Netherlands University of Information Technology VNU-HCM Viet Nam Nanjing University China SuoAo Technology Center SAEE University of Science and Technology Beijing China Chinese University of Hong Kong Hong Kong School of Software Dalian University of Technology China Department of Electrical Engineering and Computer Sciences University of California Berkeley United States Ocean University of China China

ISBN: (纸本)9783038680772

Monocular image based 3D object retrieval is a novel and challenging research topic in the field of 3D object retrieval. Given a RGB image captured in real world, it aims to search for relevant 3D objects from a dataset. To advance this promising research, we organize this SHREC track and build the first monocular image based 3D object retrieval benchmark by collecting 2D images from ImageNet and 3D objects from popular 3D datasets such as NTU, PSB, ModelNet40 and ShapeNet. The benchmark contains classified 21,000 2D images and 7,690 3D objects of 21 categories. This track attracted 9 groups from 4 countries and the submission of 20 runs. To have a comprehensive comparison, 7 commonly-used retrieval performance metrics have been used to evaluate their retrieval performance. The evaluation results show that the supervised cross domain learning get the superior retrieval performance (Best NN is 97.4 %) by bridging the domain gap with label information. However, there is still a big challenge for unsupervised cross domain learning (Best NN is 61.2%), which is more practical for the real application. Although we provided both view images and OBJ file for each 3D model, all the participants use the view images to represent the 3D model. One of the interesting work in the future is directly using the 3D information and 2D RGB information to solve the task of monocular Image based 3D model retrieval. © 2019 The Author(s)

关键词： 3D modeling

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：