检索结果-内蒙古大学图书馆

FDAM: Filter-Dedicated Approximate Multiplier Design for Real-Time CNN Acceleration

IEEE Transactions on computer-Aided Design of Integrated Circuits and Systems 2025年

作者： Nguyen, Kihwan Kim Xuan Truong Jeon, Dongsuk Lee, Hyuk-Jae Korea Republic of program Department of Electrical Engineering Seoul08826 Korea Republic of Inter-University Semiconductor Research Center Seoul National University Graduate School of Convergence Science and Technology Research Institute for Convergence Science Seoul08826 Korea Republic of Department of Electrical and Computer Engineering Seoul08826 Korea Republic of

Video super-resolution (VSR) is widely used in various high-definition applications, such as HDTVs and smartphones, requiring a dedicated upscaling technique for realtime full-HD generation. To reduce on-chip buffers for largesize output feature maps, a streaming VSR accelerator may employ an output stationary dataflow, leading to a large energy consumption caused by frequent filter switching. To mitigate this, we introduce a new filter-dedicated multiplier design for realtime VSR acceleration. We replace costly multipliers with adders, shifters, and multiplexers (MUXes), referred to as unified multiple constant multiplications (UMCM). The conventional UMCM, however, may incur considerable area/power overhead due to the unified topology constraint among different filter sets. To address this problem, we propose a new approximated MCM (AMCM) problem to relax the constraint and an approximate compatible graph synthesis (A-CGS) framework to efficiently solve AMCM by jointly searching for approximated filters and constructing a unified graph. Additionally, we suggest a lightweight finetuning method by freezing approximated filters and only finetuning biases, which can recover the original model’s accuracy within a few epochs. Experimental results with synthetic data demonstrate that AMCM reduces the area by up to 49.8%, 44.8%, and 40.3% when considering constant sets of 2, 4, and 8, respectively. Our designs with SR application achieve up to a 73.3% reduction in energy consumption. Experiments with the Set5 and Set14 datasets show that our model with bias correction achieves similar restoration performance compared to the eightbit models. © 1982-2012 IEEE.

关键词： Acceleration

来源：评论

学校读者我要写书评

暂无评论

3D Geometric Shape Assembly via Efficient Point Cloud Matching

arXiv

引用

arXiv 2024年

作者： Lee, Nahyuk Min, Juhong Lee, Junha Kim, Seungwook Lee, Kanghee Park, Jaesik Cho, Minsu Department of Computer Science and Engineering POSTECH Pohang Korea Republic of Graduate School of Artificial Intelligence POSTECH Pohang Korea Republic of Department of Computer Science and Engineering Seoul National University Seoul Korea Republic of Interdisciplinary Program in Artificial Intelligence Seoul National University Seoul Korea Republic of

Learning to assemble geometric shapes into a larger target structure is a pivotal task in various practical applications. In this work, we tackle this problem by establishing local correspondences between point clouds of part shapes in both coarse- and fine-levels. To this end, we introduce Proxy Match Transform (PMT), an approximate high-order feature transform layer that enables reliable matching between mating surfaces of parts while incurring low costs in memory and computation. Building upon PMT, we introduce a new framework, dubbed Proxy Match TransformeR (PMTR), for the geometric assembly task. We evaluate the proposed PMTR on the large-scale 3D geometric shape assembly benchmark dataset of Breaking Bad and demonstrate its superior performance and efficiency compared to state-of-the-art methods. Project page: https://***/pmtr. Copyright © 2024, The Authors. All rights reserved.

关键词： Large datasets

来源：评论

学校读者我要写书评

暂无评论

Optimization of Node Reduction Using BRKGA for GNN-Based Traffic Speed Forecasting

Optimization of Node Reduction Using BRKGA for GNN-Based Tra...

引用

Congress on Evolutionary Computation

作者： Yuto Inokuchi Pedro Henrique González Israel Mendonça Masayoshi Aritsugi Graduate School of Science and Technology Kumamoto University Kumamoto Japan System Engineering and Computer Science Program Federal University of Rio de Janeiro Rio de Janeiro Brazil Faculty of Advanced Science and Technology Kumamoto University Kumamoto Japan

ISBN: (数字)9798350308365

ISBN: (纸本)9798350308372

While Spatio-Temporal Graph Convolutional Networks (STGCNs) are an effective method for traffic speed fore-casting, their training and inference tend to be time-consuming. In this paper, we aim to refine these networks by strategically reducing their number of nodes, thereby boosting computational efficiency. The nodes in these graphs represent data observed for road segments, and by analyzing the interconnections and layout of the graph, we can identify nodes with minimal contribution to overall performance. Removing these nodes can potentially decrease computation time while maintaining the prediction accuracy. We employ the Biased Random-Key Genetic Algorithm (BRKGA) to identify a good set of nodes for removal, based on a predefined percentage reduction of the original graph size (e.g., retaining 95 % of the original graph). We evaluate different graph size configurations, ranging from 95 % to 70 % node retention, to determine the least impactful node set performance. Our experiments on three real-world datasets reveal that reducing nodes can decrease computation time by up to 29%, and as a byproduct of removing noise, even improve the prediction accuracy.

关键词： Training Accuracy Graph convolutional networks Roads Noise Layout Evolutionary computation

来源：评论

学校读者我要写书评

暂无评论

Object Detection using Event Camera: A MoE Heat Conduction based Detector and A New Benchmark Dataset

arXiv

引用

arXiv 2024年

作者： Wang, Xiao Jin, Yu Wu, Wentao Zhang, Wei Zhu, Lin Jiang, Bo Tian, Yonghong School of Computer Science and Technology Anhui University Hefei China School of Artificial Intelligence Anhui University Hefei China Peng Cheng Laboratory Shenzhen China Beijing Institute of Technology Beijing China National Key Laboratory for Multimedia Information Processing School of Computer Science Peking University China School of Electronic and Computer Engineering Shenzhen Graduate School Peking University China

Object detection in event streams has emerged as a cutting-edge research area, demonstrating superior performance in low-light conditions, scenarios with motion blur, and rapid movements. Current detectors leverage spiking neural networks, Transformers, or convolutional neural networks as their core architectures, each with its own set of limitations including restricted performance, high computational overhead, or limited local receptive fields. This paper introduces a novel MoE (Mixture of Experts) heat conduction-based object detection algorithm that strikingly balances accuracy and computational efficiency. Initially, we employ a stem network for event data embedding, followed by processing through our innovative MoE-HCO blocks. Each block integrates various expert modules to mimic heat conduction within event streams. Subsequently, an IoU-based query selection module is utilized for efficient token extraction, which is then channeled into a detection head for the final object detection process. Furthermore, we are pleased to introduce EvDET200K, a novel benchmark dataset for event-based object detection. Captured with a high-definition Prophesee EVK4-HD event camera, this dataset encompasses 10 distinct categories, 200,000 bounding boxes, and 10,054 samples, each spanning 2 to 5 seconds. We also provide comprehensive results from over 15 state-of-the-art detectors, offering a solid foundation for future research and comparison. The source code of this paper will be released on: https://***/EventAHU/OpenEvDET Copyright © 2024, The Authors. All rights reserved.

关键词： Convolutional neural networks

来源：评论

学校读者我要写书评

暂无评论

Efficient Node Reduction Heuristic for GNN-Based Traffic Speed Forecasting

Efficient Node Reduction Heuristic for GNN-Based Traffic Spe...

引用

IEEE/ACM International Symposium on Big Data Computing (BDC)

作者： Yuto Inokuchi Pedro Henrique González Masayoshi Aritsugi Israel Mendonça Graduate School of Science and Technology Kumamoto University Kumamoto Japan System Engineering and Computer Science Program Federal University of Rio de Janeiro Rio de Janeiro Brazil Faculty of Advanced Science and Technology Kumamoto University Kumamoto Japan

ISBN: (数字)9798350367300

ISBN: (纸本)9798350367317

In this paper, we aim to reduce the number of nodes from Graph Neural Networks (GNNs), thereby simplifying models and reducing computational costs. GNNs are highly effective for various tasks, such as prediction, classification, and clustering, due to their ability to learn node and edge attributes and relationships, and they have been utilized for intelligent transportation systems recently by converting sensor networks into graph structures. Deep spatio-temporal neural networks, including Spatio-Temporal Graph Convolutional Networks (STGCNs), capture spatial and temporal dependencies, making them suitable for traffic speed forecasting, traffic demand prediction, and travel time estimation. Despite their success, GNNs face challenges in industrial applications due to significant memory usage and time consumption. In this paper, we propose a new approach to node reduction that outperforms existing methods in computational efficiency. Our experiments on two real-world traffic datasets demonstrate that using the heuristic and edge information to reduce nodes can cut computation time of optimization up to 95% and, by eliminating noise, can even enhance prediction accuracy.

关键词： Graph convolutional networks Noise Memory management Estimation Computational efficiency Forecasting Optimization Intelligent transportation systems Faces Genetic algorithms

来源：评论

学校读者我要写书评

暂无评论

Efficient Hyperparameter Optimization Using Deep Q-Network and BRKGA

Efficient Hyperparameter Optimization Using Deep Q-Network a...

引用

Congress on Evolutionary Computation

作者： Kosei Kobayashi Masayoshi Aritsugi Pedro Henrique González Israel Mendonça Graduate School of Science and Technology Kumamoto University Kumamoto Japan Faculty of Advanced Science and Technology Kumamoto University Kumamoto Japan System Engineering and Computer Science Program Federal University of Rio de Janeiro Rio de Janeiro Brazil

ISBN: (数字)9798350308365

ISBN: (纸本)9798350308372

Hyperparameter optimization (HPO) is paragon to maximize performance when designing machine learning models. Among different HPO methods, Genetic Algorithm (GA) based optimization is considered effective because it allows a wide and diverse range of solutions to be explored. However, GA's exploratory nature makes this type of algorithm to evaluate many solutions that do not improve the overall performance. This is specially costly when the objective function to be evaluated is time-consuming, like in the HPO field. In this paper, we propose an efficient hybrid algorithm that is able to reduce computational cost by combining deep reinforcement learning with the Biased Random Key Genetic Algorithm (BRKGA), a variant of genetic algorithms. Our reinforcement learning agent has a decision-making role during the population's fitness calculation, in which it filters out chromosomes that would not improve the overall fitness of the population. The agent uses small amounts of pre-trained data to identify trends in potentially good solutions, and carry out its decision process. We conduct experiments on eight different datasets to assess the effectiveness of the proposed method, and the results show that the proposed method can significantly reduce the computation time of hyperparameter search using BRKGA (up to 44% reduction in computational time) without compromising the quality of the solution (no statistically difference in results).

关键词： Machine learning algorithms Filters Sociology Evolutionary computation Hyperparameter optimization Market research Linear programming

来源：评论

学校读者我要写书评

暂无评论

MIN-DAS: The Digital Asset System Based on MIN Architecture

MIN-DAS: The Digital Asset System Based on MIN Architecture

引用

Communication Technology and information Technology (ICCTIT), International Conference on

作者： Rui Chen Hui Li Wuyang Li Yimin Yu Shao Kaili Kedan Li Shenzhen Graduate School Peking University Shenzhen China Information School University of Yunnan Finance and Economics Kunming China Faculty of Engineering Huanghe Science and Technology University Zhengzhou China Department of Computer Science University of Illinois Urbana-Champaign USA

ISBN: (数字)9798331528973

ISBN: (纸本)9798331528980

In this paper, we propose an innovative digital asset system based on the MIN (MIN-DAS), which integrates PPoV consortium blockchain technology along with NFT identifiers, achieving the rights confirmation, decentralization, and high security for digital asset trading. We first deployed the MIN-DAS within the high-security MIN network and designed a built-in MIN-VPN security tunnel, enabling end-to-end communication encryption, thereby significantly bolstering transaction security. Subsequently, we adopted the PPoV consortium blockchain as the core processing engine, greatly increasing the speed of transaction processing to accommodate large-scale transaction demands. Moreover, MIN-DAS supports stablecoin transactions, effectively mitigating risks stemming from cryptocurrency market fluctuations. MIN-DAS provides a secure and efficient solution for digital asset trading. Our experimental findings reveal that, compared to existing traditional digital asset platforms, MIN-DAS demonstrates significant superiority in both performance and security.

关键词： Fluctuations Smart contracts TCPIP User interfaces Throughput Encryption Security Nonfungible tokens information technology Engines

来源：评论

学校读者我要写书评

暂无评论

Tumor segmentation using CNN for automatic diagnosis of bone tumor in X-ray image

Tumor segmentation using CNN for automatic diagnosis of bone...

引用

International Symposium on Soft Computing and Intelligent Systems (SCIS)

作者： Kaito Furuo Kento Morita Tomohito Hagi Tomoki Nakamura Kunihiro Asanuma Akihiro Sudo Tetsushi Wakabayashi Computer and Information Science Graduate School of Engineering Mie University Japan Depautment of Orthopaedic Surgery Graduate School of Medicine Mie University Japan

Bone tumors are often firstly diagnosed at the nearby hospital, because they cause bone pain and swelling. The number of patients are quite small (0.3” of population), and the identification of malignancy requires a great effort even for specialists. Therefore, there is a need to develop a system that can automatically extract bone tumor regions and estimate benign and malignant conditions. In this study, we propose a method for automatically extracting bone tumor regions using convolutional neural networks. We compared the segmentation performance of U-Net, DeepLab v3, and Mask R-CNN, and inspected the effect of tumor area cropping before semantic segmentation. Experiments using 118 knee X-ray images of 18 patients revealed that U-Net achieved the highest dice coefficient (0.67) in original x-ray image. In the experiment using cropped x-ray images, DeepLab v3 achieved highest 0.92 dice coefficient.

关键词： Pain Hospitals Semantic segmentation Sociology Training data Bones Convolutional neural networks

来源：评论

学校读者我要写书评

暂无评论

Leverage from Blockchain in Commodity Exchange: Asset-Backed Token with Ethereum Blockchain Network and Smart Contract 4th

Leverage from Blockchain in Commodity Exchange: Asset-Backed...

引用

4th International Conference on Smart Trends for Computing and Communications, SmartCom 2020

作者： Richard Heryadi, Yaya Lukas Trisetyarso, Agung Information Systems Department School of Information Systems Bina Nusantara University Jakarta11840 Indonesia Computer Science Department BINUS Graduate Program Doctor of Computer Science Bina Nusantara Univesity Jakarta11840 Indonesia Faculty of Engineering Universitas Katolik Indonesia Atma Jaya Jakarta12930 Indonesia

ISBN: (纸本)9789811552236

Ethereum is one of the most popular blockchain platforms with a high number of adoption in the blockchain world today. Ethereum token (ERC-20) can tokenize any real-world object while it is also possible to exchange the token across the network participants. This utility from the token drives many blockchain enthusiasts to create decentralized applications. ERC-20 token could be used to collect project funds, decision voting, or even can be used to provide investment platforms for people that want to hedge the volatility of the currency with buying an asset that is back to back with the crypto-token. This research shows the main characteristics of asset-backed token and its implementation on a complete end-to-end "digitalization assets system" (assets buy and assets redemption). The result shows that asset-backed token needs support from government and other legalizing parties to verify the legitimacy of the asset that backing the token itself. © 2021, The Editor(s) (if applicable) and The Author(s), under exclusive license to Springer Nature Singapore Pte Ltd.

关键词： Blockchain

来源：评论

学校读者我要写书评

暂无评论

A comparative study of augmented inverse propensity weighted estimators using outcome-oriented covariate selection via penalization with outcome-adaptive lasso

arXiv

引用

arXiv 2024年

作者： Hongo, Wataru Ando, Shuji Tsuchida, Jun Sozu, Takashi Department of Information and Computer Technology Graduate School of Engineering Tokyo University of Science Tokyo Japan Advanced Quantitative Sciences Development Novartis Pharma K.K. Tokyo Japan Department of Information Sciences Faculty of Science and Technology Tokyo University of Science Chiba Japan Department of Data Science Faculty of Data Science Kyoto Women’s University Kyoto Japan Department of Information and Computer Technology Faculty of Engineering Tokyo University of Science Tokyo Japan

Background: When estimating causal effects from observational data with numerous covariates, employing penalized covariate selection can improve the estimation efficiency. Outcome-oriented covariate selection, which involves selecting covariates related to the outcome, can enhance efficiency, even for propensity score (PS) methods. For outcome-oriented covariate selection in PS models, outcome-adaptive lasso can be used for penalization with the oracle property. The performance of inverse propensity weighted (IPW) estimators using outcome-adaptive lasso was shown to be superior to that of the IPW estimators using other covariate selection methods for parametric models. However, the augmented IPW (AIPW) estimator is typically employed as a doubly robust estimator for the average treatment effect, which requires both PS and outcome models. Despite this, which covariate selection method for outcome models should be combined with the outcome-adaptive lasso to form the AIPW estimator remains unclear. We evaluated the performance of the AIPW estimators using the outcome-adaptive lasso for PS models and various outcome-oriented covariate selection via penalization for outcome models. Methods: We conducted numerical experiments to evaluate the performance of AIPW estimators using various covariate selection via penalization. In numerical experiments, we assessed bias, standard error, and root-mean-squared error. We applied the AIPW estimators to a clinical trial dataset for reference, comparing their point estimates and standard errors obtained via bootstrap. Results: In the numerical experiments, the performance of the AIPW estimators using outcome-oriented covariate selection via penalization with the oracle property for both PS and outcome models was superior to that of the other estimators and similar to that of the AIPW estimator, which relies on true confounders and outcome predictors. In contrast, the bias of the AIPW estimators not relying on the oracle property was

关键词： Mean square error

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：