检索结果-内蒙古大学图书馆

DPK:Deep Neural Network Approximation of the First Piola-Kirchhoff Stress

Advances in Applied Mathematics and Mechanics 2024年第1期16卷 75-100页

作者： Tianyi Hu Jerry Zhijian Yang Cheng Yuan Institute of Artificial Intelligence School of Computer ScienceWuhan UniversityWuhanHubei 430072China School of Mathematics and Statistics Wuhan UniversityWuhanHubei 430072China Hubei Key Laboratory of Computational Science Wuhan UniversityWuhanHubei 430072China

This paper presents a specific network architecture for approximation of the first Piola-Kirchhoff *** neural network enables us to construct the constitutive relation based on both macroscopic observations and atomistic simulation *** contrast to traditional deep learning models,this architecture is intrinsic symmetric,guarantees the frame-indifference and material-symmetry of ***,we build the approximation network inspired by the Cauchy-Born rule and virial stress *** numerical results and theory analyses are presented to illustrate the learnability and effectiveness of our network.

关键词： Piola-Kirchhoff stress deep neural networks Cauchy-Born rule

来源：评论

学校读者我要写书评

暂无评论

Modeling Super-Lightweight Cellular Traffic Prediction via Multiservice and Multimodal Feature Fusion Network

IEEE Networking Letters

引用

IEEE Networking Letters 2024年第1期6卷 16-20页

作者： Li, Yingqi Hao, Mingxiang Sun, Xiaochuan Zhang, Haijun University of Science and Technology Beijing School of Computer and Communication Engineering Beijing100083 China North China University of Science and Technology College of Artificial Intelligence Tangshan063210 China Hebei Key Laboratory of Industrial Perception Tangshan063210 China

Cellular Traffic Prediction has proven to be a key enabler towards automatic network management. However, to pursue performance improvement, the existing studies mainly focus on developing complex deep neural network models, which suffer from extensive computation cost and large model size inevitably. Such models are quite difficult to be deployed on resource-constrained devices. In this letter, we propose a multiservice and multimodal feature fusion network for super-lightweight cellular network traffic prediction, namely m^2FFNet , to address the issue. Briefly speaking, such a network consists of a duel feature extraction channel based on grouped 3D convolution for capturing multiservice feature and multimodal feature (yielded from wavelet transform decomposition), respectively. Simulation results demonstrate that our proposal can achieve comparable prediction accuracy as the state-of-the-art deep learning methods, meanwhile obtaining much less computation burden with rather few model size. © 2019 IEEE.

关键词： Feature extraction

来源：评论

学校读者我要写书评

暂无评论

Deep Learning Operators Performance Tuning for Changeable Sized Input Data on Tensor Accelerate Hardware

引用

IEEE Transactions on computers 2025年第6期74卷 2101-2113页

作者： Mu, Pengyu Liu, Yi Wang, Rui Liu, Guoxiang An, Hangcheng Zhao, Qianhe Yang, Hailong Xie, Chenhao Luan, Zhongzhi Gong, Chunye Qian, Depei Beihang University School of Computer Science and Engineering Beijing100191 China Beihang University Hangzhou Innovation Institute Hangzhou310051 China Shanghai Artificial Intelligence Laboratory Shanghai200003 China National Supercomputer Center Tianjin300457 China

The operator library is the fundamental infrastructure of deep learning acceleration hardware. Automatically generating the library and tuning its performance is promising because the manual development by well-trained and skillful programmers is costly in terms of both time and money. Tensor hardware has the best computing efficiency for deep learning applications, but the operator library programs are hard to tune because the tensor hardware primitives have many limitations. Otherwise, the performance is difficult to be fully explored. The recent advancement in LLM exacerbates this problem because the size of input data is not fixed. Therefore, mapping the computing tasks of operators to tensor hardware units is a significant challenge when the shape of the input tensor is unknown before the runtime. We propose DSAT, a deep learning operator performance autotuning technique for changeable-sized input data on tensor hardware. To match the input tensor's undetermined shape, we choose a group of abstract computing units as the basic building blocks of operators for changeable-sized input tensor shapes. We design a group of programming tuning rules to construct a large exploration space of the variant implementation of the operator programs. Based on these rules, we construct an intermediate representation of computing and memory access to describe the computing process and use it to map the abstract computing units to tensor primitives. To speed up the tuning process, we narrow down the optimization space by predicting the actual hardware resource requirement and providing an optimized cost model for performance prediction. DSAT achieves performance comparable to the vendor's manually tuned operator libraries. Compared to state-of-the-art deep learning compilers, it improves the performance of inference by 13% on average and decreases the tuning time by an order of magnitude. © 1968-2012 IEEE.

关键词： Tensors

来源：评论

学校读者我要写书评

暂无评论

Designing a skilled soccer team for RoboCup: exploring skill-set-primitives through reinforcement learning

引用

Neural Computing and Applications 2025年 1-36页

作者： Abreu, Miguel Reis, Luís Paulo Lau, Nuno LIACC/LASI/FEUP Artificial Intelligence and Computer Science Laboratory Faculty of Engineering University of Porto Porto Portugal IEETA/LASI/DETI Institute of Electronics and Informatics Engineering of Aveiro Department of Electronics Telecommunications and Informatics University of Aveiro Aveiro Portugal

The RoboCup 3D soccer simulation league serves as a competitive platform for showcasing innovation in autonomous humanoid robot agents through simulated soccer matches. Our team, FC Portugal, developed a new codebase from scratch in Python after RoboCup 2021. The team’s performance relies on a set of skills centered around novel unifying primitives and a custom, symmetry-extended version of the proximal policy optimization algorithm. Our methods have been thoroughly tested in official RoboCup matches, where FC Portugal has won the last two main competitions, in 2022 and 2023. This paper presents our training framework, as well as a timeline of skills developed using our skill-set-primitives, which considerably improve the sample efficiency and stability of skills, and motivate seamless transitions. We start with a significantly fast Sprint-Kick developed in 2021 and progress to the most recent skill set, including a multi-purpose omnidirectional walk, a dribble with unprecedented ball control, a solid kick, and a push skill. The push addresses low-level collision scenarios and high-level strategies to increase ball possession. We address the resource-intensive nature of this task through an innovative multi-agent learning approach. Finally, we release the team’s codebase to the RoboCup community, providing other teams with a robust and modern foundation upon which they can build new features. © The Author(s) 2025.

关键词： Reinforcement learning

来源：评论

学校读者我要写书评

暂无评论

Machine Learning for Brain Imaging Genomics Methods:A Review

引用

Machine intelligence Research 2023年第1期20卷 57-78页

作者： Mei-Ling Wang Wei Shao Xiao-Ke Hao Dao-Qiang Zhang College of Computer Science and Technology Nanjing University of Aeronautics and AstronauticsNanjing 211106China Key Laboratory of Pattern Analysis and Machine Intelligence Ministry of Industry and Information TechnologyNanjing 211106China School of Artificial Intelligence Hebei University of TechnologyTianjin 300401China

In the past decade,multimodal neuroimaging and genomic techniques have been increasingly *** an interdiscip-linary topic,brain imaging genomics is devoted to evaluating and characterizing genetic variants in individuals that influence phenotyp-ic measures derived from structural and functional brain *** technique is capable of revealing the complex mechanisms by macroscopic intermediates from the genetic level to cognition and psychiatric disorders in *** is well known that machine learn-ing is a powerful tool in the data-driven association studies,which can fully utilize priori knowledge(intercorrelated structure informa-tion among imaging and genetic data)for association *** addition,the association study is able to find the association between risk genes and brain structure or function so that a better mechanistic understanding of behaviors or disordered brain functions is *** this paper,the related background and fundamental work in imaging genomics are first ***,we show the univari-ate learning approaches for association analysis,summarize the main idea and modelling in genetic-imaging association studies based on multivariate machine learning,and present methods for joint association analysis and outcome ***,this paper discusses some prospects for future work.

关键词： Brain imaging genomics machine learning multivariate analysis association analysis outcome prediction

来源：评论

学校读者我要写书评

暂无评论

Towards Gradient Equalization and Feature Diversification for Long-Tailed Multi-Label Image Recognition

引用

IEEE Transactions on Multimedia 2025年 27卷 3489-3500页

作者： Chen, Zhao-Min Cui, Quan Zhang, Xiaoqin Deng, Ruoxi Xia, Chaoqun Lu, Shijian Nanjing University National Key Laboratory for Novel Software Technology Nanjing210093 China Waseda University Graduate School of Information Production and Systems Fukuoka169-8050 Japan Wenzhou University College of Computer Science and Artificial Intelligence Wenzhou325060 China Nanyang Technological University School of Computer Science and Engineering 639798 Singapore

Multi-label image recognition with convolutional neural networks has achieved remarkable progress in the past few years. However, most existing multi-label image recognition methods suffer from the long-tailed data distribution problem, i.e., head categories occupy most training samples, while tailed classes have few samples. This work firstly studies the influence of long-tailed data distribution on existing multi-label image recognition methods. Based on this, two crucial issues of the existing methods are identified: 1) severe gradient imbalance between head and tailed categories, even though re-balancing strategies are adopted;2) the lack of diversity of tail category training samples. To tackle the first issue, this paper proposes a group sampling strategy to create group- wise balanced data distribution. Meanwhile, a dynamic gradient balancing loss is proposed to equalize the gradient for all categories. To tackle the second issue, this paper proposes a diversity enhancement module to fuse the information across all categories, preventing the network from overfitting tail classes. Furthermore, it also balances the gradient, promoting the discriminability of learned classifiers. Our method significantly outperforms the baseline method and achieves competitive performance with state-of-the-art methods on VOC-LT and COCO-LT datasets. Extensive ablation studies are conducted to verify the effectiveness of the essential proposals. © 1999-2012 IEEE.

关键词： Labeled data

来源：评论

学校读者我要写书评

暂无评论

Fault Tolerance in The IoT: A Taxonomy Based on Techniques 3

Fault Tolerance in The IoT: A Taxonomy Based on Techniques

引用

3rd International Conference on Theoretical and Applicative Aspects of computer science, ICTAACS 2023

作者： Oussane, Sakina Benkaouha, Haroun Djouama, Amir Usthb University Lsi Laboratory Computer Science Faculty Algiers Algeria National School of Artificial Intelligence Lsi Laboratory Algiers Algeria

ISBN: (纸本)9798350385854

In order to ensure that IoT systems are reliable, efficient and robust, it is important for failures to be easily detectable and effectively managed to minimize their impact on users. This is why fault tolerance approaches are necessary in the Internet of Things (IoT). This article provides an overview of failures in IoT systems and proposes a classification of fault tolerance techniques. It also explores various existing fault tolerance approaches that can be used to develop effective solutions. © 2023 IEEE.

关键词： Fault tolerance

来源：评论

学校读者我要写书评

暂无评论

Heterogeneous Domain Adaptation via Correlative and Discriminative Feature Learning

引用

IEEE Transactions on Multimedia 2025年 27卷 3447-3461页

作者： Lu, Yuwu Lin, Dewei Shen, Linlin Zhou, Yicong Pan, Jiahui South China Normal University School of Artificial Intelligence Foshan528225 China Shenzhen University College of Computer Science and Software Engineering Shenzhen518055 China Shenzhen University Guangdong Key Laboratory of Intelligent Information Processing Shenzhen518060 China University of Nottingham Ningbo China Department of Computer Science Ningbo315100 China University of Macau Department of Computer and Information Science China

Heterogeneous domain adaptation seeks to learn an effective classifier or regression model for unlabeled target samples by using the well-labeled source samples but residing in different feature spaces and lying different distributions. Most recent works have concentrated on learning domain-invariant feature representations to minimize the distribution divergence via target pseudo-labels. However, two critical issues need to be further explored: 1) new feature representations should be not only domain-invariant but also category-correlative and discriminative, and 2) alleviating the negative transfer caused by the incorrect pseudo-labeling target samples could boost the adaptation performance during the iterative learning process. To address these issues, in this paper, we put forward a novel heterogeneous domain adaptation method to learn category-correlative and discriminative representations, referred to as correlative and discriminative feature learning (CDFL). Specifically, CDFL aims to learn a feature space where class-specific feature correlations between the source and target domains are maximized, the divergences of marginal and conditional distribution between the source and target domains are minimized, and the distances of inter-class distribution are forced to be maximized to ensure the discriminative ability. Meanwhile, a selective pseudo-labeling procedure based on the correlation coefficient and classifier prediction is introduced to boost class-specific feature correlation and discriminative distribution alignment in an iteration way. Extensive experiments certify that CDFL outperforms the state-of-the-art algorithms on five standard benchmarks. © 1999-2012 IEEE.

关键词： Contrastive Learning

来源：评论

学校读者我要写书评

暂无评论

On Unsupervised Uncertainty-Driven Speech Pseudo-Label Filtering and Model Calibration 48

On Unsupervised Uncertainty-Driven Speech Pseudo-Label Filte...

引用

48th IEEE International Conference on Acoustics, Speech and Signal Processing, ICASSP 2023

作者： Dawalatabad, Nauman Khurana, Sameer Laurent, Antoine Glass, James MIT Computer Science and Artificial Intelligence Laboratory CambridgeMA United States LIUM - Le Mans University France

ISBN: (纸本)9781728163277

Pseudo-label (PL) filtering forms a crucial part of Self-Training (ST) methods for unsupervised domain adaptation. Dropout-based Uncertainty-driven Self-Training (DUST) proceeds by first training a teacher model on source domain labeled data. Then, the teacher model is used to provide PLs for the unlabeled target domain data. Finally, we train a student on augmented labeled and pseudo-labeled data. The process is iterative, where the student becomes the teacher for the next DUST iteration. A crucial step that precedes the student model training in each DUST iteration is filtering out noisy PLs that could lead the student model astray. In DUST, we proposed a simple, effective, and theoretically sound PL filtering strategy based on the teacher model's uncertainty about its predictions on unlabeled speech utterances. We estimate the model's uncertainty by computing disagreement amongst multiple samples drawn from the teacher model during inference by injecting noise via dropout. In this work, we show that DUST's PL filtering, as initially used, fail under severe source and target domain mismatch. We suggest several approaches to eliminate or alleviate this issue. Further, we bring insights from the research in neural network model calibration to DUST and show that a well-calibrated model correlates strongly with a positive outcome of the DUST PL filtering step. We demonstrate effectiveness of our methods on target domain GigaSpeech YouTube dataset. © 2023 IEEE.

关键词： Students

来源：评论

学校读者我要写书评

暂无评论

LQTTrack:Multi-Object Tracking by Focusing on Low-Quality Targets Association

引用

computers, Materials & Continua 2024年第10期81卷 1449-1470页

作者： Suya Li Ying Cao Hengyi Ren Dongsheng Zhu Xin Xie Henan Key Laboratory of Big Data Analysis and Processing School of Computer and Information EngineeringHenan UniversityKaifeng475001China College of Information Science and Technology&Artificial Intelligence Nanjing Forestry UniversityNanjing210037China School of Computer Nanjing University of Posts and TelecommunicationsNanjing210023China

Multi-object tracking(MOT)has seen rapid improvements in recent ***,frequent occlusion remains a significant challenge in MOT,as it can cause targets to become smaller or disappear entirely,resulting in lowquality targets,leading to trajectory interruptions and reduced tracking *** from some existing methods,which discarded the low-quality targets or ignored low-quality target ***,with a lowquality association strategy(LQA),is proposed to pay more attention to low-quality *** the association scheme of LQTTrack,firstly,multi-scale feature fusion of FPN(MSFF-FPN)is utilized to enrich the feature information and assist in subsequent data ***,the normalized Wasserstein distance(NWD)is integrated to replace the original Inter over Union(IoU),thus overcoming the limitations of the traditional IoUbased methods that are sensitive to low-quality targets with small sizes and enhancing the robustness of low-quality target ***,the third association stage is proposed to improve the matching between the current frame’s low-quality targets and previously interrupted trajectories from earlier frames to reduce the problem of track fragmentation or error tracking,thereby increasing the association success rate and improving overall multi-object tracking *** experimental results demonstrate the competitive performance of LQTTrack on benchmark datasets(MOT17,MOT20,and DanceTrack).

关键词： Low-quality targets association strategy feature fusion multi-object tracking tracking-by-detection

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：