检索结果-内蒙古大学图书馆

arXiv 2024年

作者： Xu, Zhixuan Gao, Chongkai Liu, Zixuan Yang, Gang Tie, Chenrui Zheng, Haozhuo Zhou, Haoyu Peng, Weikun Wang, Debang Hu, Tianrun Chen, Tianyi Yu, Zhouliang Shao, Lin The Department of Computer Science National University of Singapore Singapore Tsinghua Shenzhen International Graduate School Tsinghua University China School of Electronics Engineering and Computer Science Peking University China Department of Mathematics National University of Singapore Singapore Department of Mechanical Engineering National University of Singapore Singapore The Department of Computer Science and Engineering Shanghai Jiao Tong University China The Division of Emerging Interdisciplinary Areas Hongkong University of Science and Technology Hong Kong

To substantially enhance robot intelligence, there is a pressing need to develop a large model that enables general-purpose robots to proficiently undertake a broad spectrum of manipulation tasks, akin to the versatile task-planning ability exhibited by LLMs. The vast diversity in objects, robots, and manipulation tasks presents huge challenges. Our work introduces a comprehensive framework to develop a foundation model for general robotic manipulation that formalizes a manipulation task as contact synthesis. Specifically, our model takes as input object and robot manipulator point clouds, object physical attributes, target motions, and manipulation region masks. It outputs contact points on the object and associated contact forces or post-contact motions for robots to achieve the desired manipulation task. We perform extensive experiments both in the simulation and real-world settings, manipulating articulated rigid objects, rigid objects, and deformable objects that vary in dimensionality, ranging from one-dimensional objects like ropes to two-dimensional objects like cloth and extending to three-dimensional objects such as plasticine. Our model achieves average success rates of around 90%. Supplementary materials and videos are available on our project website at https://***/. Copyright © 2024, The Authors. All rights reserved.

关键词： Robot programming

来源：评论

学校读者我要写书评

暂无评论

Optimizing Fuzzy System of Fuzzy Time Series for Hyper Spectral Image Classification

Optimizing Fuzzy System of Fuzzy Time Series for Hyper Spect...

引用

Optimization Computing and Wireless Communication (ICOCWC), International Conference on

作者： M.S. Nidhya Preeti Naval Ravindra Kumar Department of Computer Science and Information Technology Jain (Deemed to be University) Bangalore India Maharishi School of Engineering & Technology Maharishi University of Information Technology Uttar Pradesh India Department of Mathematics Vivekananda Global University Jaipur India

This research paper examines the capability of fuzzy time collection for hyperspectral photograph classification. Fuzzy time series (FTS) is a time series in which fuzzy standards are used to model the styles within the facts. FTS can be used to explain complex temporal styles in the records, and as a consequence making it possible to categorize photographs more extraordinarily accurately., this look proposes an optimization method primarily based on genetic seek techniques. The optimization algorithm is designed to discover the high-quality FTS parameters that yield first-rate type accuracy. The efficacy of the proposed technique is evaluated on hyperspectral facts set with extraordinary experimental scenarios. The results of the test display that the proposed method can enhance the accuracy of photo classification and the use of FTS considerably. Hence, the proposed method gives a promising technique that can be used to classify hyperspectral snapshots efficiently. The paper affords an optimized fuzzy machine of fuzzy time collection for the hyperspectral photograph category. The proposed device consists of 3 levels: pre-processing, version creation, and optimization. Throughout the pre-processing level, statistical and spectral analyses are executed to acquire the applicable attributes for developing the fuzzy time collection. The model construction degree then uses the bushy time series to extract between-class separability for the photo type. It is followed utilizing the optimization stage, related to the software of differential evolution, to minimize the complexity of the proposed machine while still enhancing the type accuracy. The proposed machine has been correctly carried out to a real-international hyperspectral dataset and demonstrates widespread upgrades in class accuracy over existing methods.

关键词：

来源：评论

学校读者我要写书评

暂无评论

An Analysis of YOLOvX Deep Learning Models for Colon Cancer Detection

引用

Procedia computer science 2025年 260卷 81-90页

作者： Raseena T.P. Aditya Kumar Soni Jitendra Kumar T. Akilan Ashutosh Kumar Singh Department of Computer Applications National Institute of Technology Tiruchirappalli 620015 Tamilnadu India Kotak Mahindra Bank Thane West Mumbai 400604 Maharastra India Department of Mathematics Bioinformatics and Computer Applications Maulana Azad National Institute of Technology Bhopal 462003 M.P. India Department of Software Engineering Lakehead University Thunder Bay Canada Department of Computer Science and Engineering Indian Institute of Information Technology Bhopal India

Around the world, Colorectal cancer is still a prevalent form of cancer, and effective treatment and its impact depends on early diagnosis. Recently, polyp detection methods employ a convolutional neural networks to identify precancerous or malignant polyps in colonoscopy images accurately with high speed and precision. This study evaluates four deep learning models namely YOLOv3n, YOLOv5s, YOLOv7, and YOLOv7x, to determine their effectiveness in detecting colorectal cancer polyp frames efficiently. Among these, YOLOv7x model exhibited outstanding performance on the Hyper Kavasir dataset compared to other models, achieving an F1 score of 88.0%, a recall of 86.4%, a precision of 89.5%, and a mAP of 92.0% at a confidence threshold of 0.353. The achieved results highlight the unique capability of models in detecting CRC polyps for diagnosis. Incorporating such detection models can help physicians significantly in clinical practice to improve their ability to identify malignant polyps, ultimately leading to better treatment for patients. Gastroenterologists can benefit considerably from these models as diagnostic tools because of their speed and accuracy.

关键词： Colorectal cancer Colonoscopy Deep learning YOLO

来源：评论

学校读者我要写书评

暂无评论

Machine Learning based rainfall forecasting in different season in Bangladesh

Machine Learning based rainfall forecasting in different sea...

引用

Recent Progresses in science, Engineering and technology (ICRPSET), International Conference on

作者： Shuva Mai Paul MD Habibur Rahman MD Miskat Hossain Siddique Department of Computer Science and Engineering University of Science and Technology Chittagong Chittagong Bangladesh Department of Mathematics University of Chittagong Chittagong Bangladesh

The main purpose of this study was to assess recent changes in rainfall intensity and seasonal rainfall variability in Bangladesh. By exploring the data collaboratively, daily rainfall records from 34 meteorological stations spread throughout 7 areas of Bangladesh between 1989 and 2018 were used to analyze changes in rain intensity and seasonal rainfall variability. The variability of spring and summer rainfall increased during the last 30 years, but winter fall showed less variability in seasonal totals in the last 30 years[12]. During the last 30 years (1989–2018), the accuracy of rainfall estimation has been above 65%. The changing pattern of the climate means neither scarcity nor heavy rainfall affects rural or urban life to a great extent. The classification analysis models have been developed for the rainfall prediction of 34 metrological stations in Bangladesh. Rainfall data analysis revealed an escalating tendency throughout the monsoon and post-monsoon seasons, while the winter season revealed a descending trend and the pre-monsoon season showed no discernible change. In the same time frame, it was discovered that the annual average rainfall trend was declining at a rate of 0.023 mm per year. The effects of rainfall patterns on agriculture are significant in Bangladesh.

关键词：

来源：评论

学校读者我要写书评

暂无评论

Mining High Utility Sequences with a Novel Utility Function 13

Mining High Utility Sequences with a Novel Utility Function

引用

13th International Conference Knowledge and Systems Engineering, KSE 2021

作者： Huynh, Hanh-Thong Duong, Hai Truong, Tin Le, Bac Fournier-Viger, Philippe University of Science Faculty of Information Technology Vnu Hcmc Ho Chi Minh Viet Nam Dalat University Department of Mathematics and Computer Science Viet Nam University of Science Vnu Hcmc Department of Computer Science Ho Chi Minh Viet Nam Shenzhen University Shenzhen China

ISBN: (纸本)9781665499750

Mining high utility sequential patterns (HUSP) is a popular data mining task. The goal is to find all subsequences that yield a high utility (e.g. high profit) in a quantitative sequence database (QSDB). Traditional algorithms for this task have many uses but a major limitation is that they rely on the maximum or minimum utility measures for calculating the utility of a pattern, thus assuming either a best or worst case scenario. These measures are unsuitable for many real-life applications such as business decision-making. To address this issue, this paper introduces a novel utility function (NUF) to calculate the utility of a sequence in each input sequence, which provides a trade-off between the above two extreme cases. A novel upper bound on NUF is designed as well as search space pruning strategies to eliminate unpromising candidate patterns early. These contributions are integrated into a novel efficient algorithm named FHNewUSM to discover frequent HUSPs with NUF. An experimental study with both real-life and synthetic databases shows that the proposed algorithm is efficient for mining HUSPs with NUF in terms of execution time, memory consumption and scalability. © 2021 IEEE.

关键词： Decision making

来源：评论

学校读者我要写书评

暂无评论

STRL: Writer-Independent Offline Signature Verification with Transformers and Self-Supervised Representation Learning

STRL: Writer-Independent Offline Signature Verification with...

引用

International Conference on computer and Communications (ICCC)

作者： Qianxi Shi Yulong Wang Tongcun Liu Lei Zhang Jianxin Liao State Key Laboratory of Networking and Switching Technology Beijing University of Posts and Telecommunications Beijing China School of Mathematics and Computer Science Zhejiang A&F University Hangzhou China

ISBN: (数字)9798331507077

ISBN: (纸本)9798331507084

Handwritten signature verification is a commonly used method of identity authentication, but given its relatively lower fabrication difficulty compared to other biometric characteristics like facial recognition, designing algorithms for this purpose poses greater challenges. Existing handwritten signature verification algorithms mainly suffer from two drawbacks: firstly, due to the private and confidentiality of signatures, there is often a lack of substantial high-quality signature images for model training. Secondly, the majority of existing algorithms are implemented using convolutional architectures, where the localized neighborhood operations of CNNs limit the model's ability to capture the global interrelationships of signature strokes. To address these challenges, we introduce a self-supervised learning algorithm grounded in the Transformer framework. This approach incorporates convolutional cropping and grayscale inversion for the preprocessing of input images. Additionally, we integrate a convolutional image block encoding module into the network to supplement the network with local contextual information. The superiority of our proposed algorithm over current state-of-the-art self-supervised algorithms is validated through t-SNE visualization modeling and comparative ablation studies, demonstrating its feasibility.

关键词： Training Representation learning Fabrication Handwriting recognition Visualization Image coding Face recognition Self-supervised learning Gray-scale Transformers

来源：评论

学校读者我要写书评

暂无评论

Modeling of Measurement Error in Financial Returns Data

arXiv

引用

arXiv 2024年

作者： Jasra, Ajay Maama, Mohamed Mijatović, Aleksandar School of Data Science The Chinese University of Hong Kong Shenzhen China Applied Mathematics and Computational Science Program Computer Electrical and Mathematical Sciences and Engineering Division King Abdullah University of Science and Technology Thuwal23955-6900 Saudi Arabia Department of Statistics University of Warwick United Kingdom

In this paper we consider the modeling of measurement error for fund returns data. In particular, given access to a time-series of discretely observed log-returns and the associated maximum over the observation period, we develop a stochastic model which models the true log-returns and maximum via a Lévy process and the data as a measurement error there-of. The main technical difficulty of trying to infer this model, for instance Bayesian parameter estimation, is that the joint transition density of the return and maximum is seldom known, nor can it be simulated exactly. Based upon the novel stick breaking representation of [12] we provide an approximation of the model. We develop a Markov chain Monte Carlo (MCMC) algorithm to sample from the Bayesian posterior of the approximated posterior and then extend this to a multilevel MCMC method which can reduce the computational cost to approximate posterior expectations, relative to ordinary MCMC. We implement our methodology on several applications including for real data. Copyright © 2024, The Authors. All rights reserved.

关键词： Markov chains

来源：评论

学校读者我要写书评

暂无评论

A Multi-View Learning-Based Bayesian Ruleset Extraction Algorithm For Accurate Hepatotoxicity Prediction

A Multi-View Learning-Based Bayesian Ruleset Extraction Algo...

引用

2023 IEEE International Conference on Bioinformatics and Biomedicine, BIBM 2023

作者： Chen, Siyang Wu, Lianlian Xu, Yong Liu, Kunhong He, Song Bo, Xiaochen Xiamen University School of Informatics Xiamen China Tianjin University Academy of Medical Engineering and Translational Medicine Tianjin China Fujian University of Technology School of Computer Science and Mathematics Fuzhou China Xiamen University School of Film Department of Digital Media Technology Xiamen China Institute of Health Service and Transfusion Medicine Department of Bioinformatics Beijing China

ISBN: (纸本)9798350337488

The interpretable machine learning method is important in drug discovery. Unlike traditional ensemble learning methods, this paper proposes an interpretable algorithm based on Bayesian rule extraction to obtain reliable and explainable results for hepatotoxicity prediction. To extract information from different types of omics data, our algorithm employs a multi-view learning strategy to enhance performance. Specifically, a random forest is trained in each view, and then the Bayesian rule extraction algorithm is designed to select an optimal rule subset, controlling the size and accuracy of the ruleset through probabilities. These rule sets are integrated through multi-view voting to get the final decisions. The performance of our algorithm is tested on the hepatotoxicity dataset, demonstrating that compared to traditional machine learning algorithms and rule-based algorithms, our approach maintains excellent performance while achieving high interpretability in most cases. Our python source code and the related Supplementary Materials are available at: ***/MLDMXM2017/MV-BRS. © 2023 IEEE.

关键词： Bayesian Ruleset Hepatotoxicity Prediction Interpretable Machine Learning Multi-view Learning

来源：评论

学校读者我要写书评

暂无评论

Analyzing the Performance of Online Adaptive Clustering for Time Series with Head-Based Aggregation

Analyzing the Performance of Online Adaptive Clustering for ...

引用

Optimization Computing and Wireless Communication (ICOCWC), International Conference on

作者： Ankita Agarwal Murli Manohar Gour Karthikeyan M P Maharishi School of Engineering & Technology Maharishi University of Information Technology Uttar Pradesh India Department of Mathematics Vivekananda Global University Jaipur India Department of Computer Science and Information Technology Jain (Deemed to be University) Bangalore India

This paper presents an analytical examine the present day, an internet adaptive clustering technique for time collection statistics with Head-primarily based Aggregation (HBA). Using a Markov chain technique, the authors analyze the asymptotic overall performance of cutting-edge global clustering blunders, which consist of the sum of present-day misclassification and the clustering structure errors. They offer a closed-shape expression new to the predicted global error rate and show its conduct quantitatively beneath distinct eventualities. Thru massive experiments on both synthetic and actual datasets, they show the effectiveness of the proposed method compared to the conventional okay-approach clustering and a few online clustering algorithms. The effects show that the proposed method can successfully seize time collection clustering shape.

关键词：

来源：评论

学校读者我要写书评

暂无评论

DCHT: Deep Complex Hybrid Transformer for Speech Enhancement

DCHT: Deep Complex Hybrid Transformer for Speech Enhancement

引用

International Conference on Digital Data Processing (DDP)

作者： Jialu Li Junhui Li Pu Wang Youshan Zhang School of public policy Cornell University Ithaca NY USA Department of Mathematics School of Science University of Science and Technology Liaoning Anshan China Department of Artificial Intelligence and Computer Science Yeshiva University New York NY USA

Most of the current deep learning-based approaches for speech enhancement only operate in the spectrogram or wave-form domain. Although a cross-domain transformer combining waveform- and spectrogram-domain inputs has been proposed, its performance can be further improved. In this paper, we present a novel deep complex hybrid transformer that integrates both spectrogram and waveform domains approaches to improve the performance of speech enhancement. The proposed model consists of two parts: a complex Swin-Unet in the spectrogram domain and a dual-path transformer network (DPTnet) in the waveform domain. We first construct a complex Swin- $V$ net network in the spectrogram domain and perform speech enhancement in the complex audio spectrum. We then introduce improved DPT by adding memory-compressed attention. Our model is capable of learning multi-domain features to reduce existing noise on different domains in a complementary way. The experimental results on the BirdSoundsDenoising dataset and the VCTK+DEMAND dataset indicate that our method can achieve better performance compared to state-of-the-art methods.

关键词：

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：