检索结果-内蒙古大学图书馆

Unsupervised AoA Estimation Based on Dual-Path Knowledge-Aware Auto-Encoders

IEEE Transactions on Wireless Communications 2025年第5期24卷 4425-4439页

作者： Guo, Zhiheng Xiao, Yuanzhang Wang, Xijun Chen, Xiang University of Hawaii Hawaii Advanced Wireless Technologies Institute HonoluluHI United States Sun Yat-sen University School of Electronics and Information Technology Guangzhou510275 China The Chinese University of Hong Kong Hong Kong Guangdong Provincial Key Laboratory of Big Data Computing Shenzhen China

In this paper, an unsupervised deep learning-based framework based on dual-path model-driven auto-encoders (AE) is proposed for angle-of-arrivals (AoAs) estimation in massive MIMO systems. Specifically designed for AoA estimation, the proposed framework differs from the conventional AE in two aspects. Firstly, unlike conventional auto-encoders, our framework employs a dual-path neural network for the encoder, decoupling the estimated parameters and enabling independent updates of each paths. Secondly, the decoder has fixed weights that implement the signal propagation model, instead of learnable parameters. This knowledge-aware decoder ensures the output of meaningful physical parameters (i.e., AoAs) which is unattainable by conventional AEs. We also conduct a thorough analysis to characterize the multiple global optima and local optima of the estimation problem. This analysis inspires the design of a low-complexity two-phase training scheme and confirms the convergence of our proposed framework. Consequently, our framework addresses two key challenges in unsupervised learning: the lack of interpretability and the convergence to local optima. Extensive simulations validate our theoretical analysis and demonstrate the performance improvements of our proposed framework. © 2002-2012 IEEE.

关键词： Unsupervised learning

来源：评论

学校读者我要写书评

暂无评论

Detection of Brain Tumor Using Transfer Learning Using Conventional Autoencoder with Long Short Term Memory Method

Detection of Brain Tumor Using Transfer Learning Using Conve...

引用

Mobile computing and Sustainable Informatics (ICMCSI), International Conference on

作者： D Agalya S. Kamalakkannan Department of Computer Science School of Computing Sciences Vels Institute of Science Technology & Advanced Studies Chennai India Department of Information Technology School of Computing Sciences Vels Institute of Science Technology & Advanced Studies Chennai India

ISBN: (数字)9798331522667

ISBN: (纸本)9798331522674

Brain tumor (BT) has generate a significant health challenge by putting pressure on healthy brain parts or spreading into other areas as well as blocking the flow of fluid around the brain. BT diagnosis is an extensive and time consumption process that depends primarily on radiologists experience and interpretive capabilities. For the area of medical imaging, automatic segmentation of images by means of Deep Learning (DL) approaches. The DL approaches assist in transforming the sector, resulted in improving precision and effectiveness of investigations. However, there are various feature extraction mechanisms available. Therefore, the Recurrent Neural Network (RNN) is utilized for feature extraction and image classification. In model execution, the data have trained in transforming the test image as well as data features for minimizing the domain shift is calculated through the Convolutional Autoencoder (CAE) for reconstruction loss. This research has concentrated in building a model with VGG16 as a single test that subjected at inference and existing method is adopted as neural networks for AE as Transfer Learning (TL) that performs an image analysis task such as segmentation and even set as an adopter for pre training the model. The AE used to train from the source dataset and perform as the adaptors in optimizing during testing using a test subject for effective computation. Moreover, the Long Short Term Memory (LSTM) is utilized as RNN model with CAE for providing improved detection of BT in health care industries. Hence, the proposed CAE with LSTM is compared with AE with Convolutional Neural Network (CNN) for evaluating BT detection using MRI dataset with various BT type classifications.

关键词： Training Recurrent neural networks Computational modeling Autoencoders Transfer learning Predictive models Convolutional neural networks Long short term memory Image classification Testing

来源：评论

学校读者我要写书评

暂无评论

Large magnetoresistance in a quantum dot between two reservoirs with spin accumulation

引用

Physical Review B 2025年第10期111卷 104407-104407页

作者： R. Jansen S. Yuasa Research Center for Emerging Computing Technologies National Institute of Advanced Industrial Science and Technology (AIST) Tsukuba Ibaraki 305-8568 Japan

A few-electron quantum dot is shown to exhibit a millionfold magnetoresistance if the source and drain electron reservoirs have spin accumulations ΔμS and ΔμD that are induced and controlled by two ferromagnetic contacts. If ΔμS and ΔμD have the same sign, a single-electron spin is trapped on the quantum dot and the current vanishes due to Pauli spin blockade between the single quantum dot and the spin-polarized reservoirs. Charge can flow freely when ΔμS and ΔμD are antiparallel. Transport calculations for sequential tunneling show that when the thermal energy is small compared to the singlet-triplet splitting energy, the current on/off ratio is above 106 for Si quantum dots. The device can operate as a spin transistor with huge magnetoresistance, but can also readily be switched to operate as a spin qubit by reconfiguring the source-drain bias and the gate voltage.

关键词： Nanocrystals

来源：评论

学校读者我要写书评

暂无评论

Leveling Up Together: Fostering Positive Growth and Safe Online Spaces for Teen Roblox Developers 25

Leveling Up Together: Fostering Positive Growth and Safe Onl...

引用

Proceedings of the 2025 CHI Conference on Human Factors in computing Systems

作者： Yubin Choi Jeanne Choi Joseph Seering Computer Science Korea Advanced Institute of Science & Technology Daejeon Republic of Korea School of Computing Korea Advanced Institute of Science & Technology Daejeon Republic of Korea

来源：评论

学校读者我要写书评

暂无评论

A Prompt Learning Framework with Large Language Model Augmentation for Few-shot Multi-label Intent Detection

A Prompt Learning Framework with Large Language Model Augmen...

引用

2025 IEEE International Conference on Acoustics, Speech, and Signal Processing, ICASSP 2025

作者： Zhuang, Ning Wei, Xiao Li, Junlei Wang, Xiaobao Wang, Chenyang Wang, Longbiao Dang, Jianwu Tianjin Key Laboratory of Cognitive Computing and Application College of Intelligence and Computing Tianjin University Tianjin China Shenzhen China Co. Ltd. Tianjin China Shenzhen Institute of Advanced Technology Chinese Academy of Sciences Shenzhen China

ISBN: (纸本)9798350368741

Intent detection (ID) is essential in spoken language understanding, especially in multi-label settings where intent labels are interdependent and diverse. Existing methods like SE-MLP and QA-FT struggle in few-shot settings, due to limited data availability and efficiency concerns. To address this, we introduce a Prompt Learning framework with large language Model Augmentation (PLMA) for few-shot multi-label ID. In this study, we make three contributions. First, PLMA integrates large language models (LLMs) with small language models (SLMs), using LLMs to enhance SLMs, with prompt learning as the core framework. Second, it leverages LLMs to improve the model's understanding of both queries and labels by intent spans extraction and answer space expansion. Third, PLMA refines the QA-FT. A novel QA format template is designed to allow single-step training and inference for each input, improving efficiency. Experiments on NLU++ show that PLMA significantly outperforms other baselines in both in-domain and cross-domain settings, demonstrating its effectiveness in combining the strengths of both small and large models. © 2025 IEEE.

关键词： large language model multi-label intent detection prompt learning spoken language understanding

来源：评论

学校读者我要写书评

暂无评论

DaCoM: Strategies to Construct Domain-specific Low-resource Language Machine Translation Dataset 31

DaCoM: Strategies to Construct Domain-specific Low-resource ...

引用

31st International Conference on Computational Linguistics, COLING 2025

作者： Kang, Junghoon Tak, Keunjoo Choi, Joung Su Kim, Myunghyun Jang, Junyoung Kang, Youjin AI Center HD Korea Shipbuilding & Offshore Engineering Korea Republic of DT Innovation Deparment HD Hyundai Samho Korea Republic of School of Computing Korea Advanced Institute of Science & Technology Korea Republic of

ISBN: (纸本)9798891761971

Translation of low-resource languages in industrial domains is essential for improving market productivity and ensuring foreign workers have better access to information. However, existing translators struggle with domain-specific terms, and there is a lack of expert annotators for dataset creation. In this work, we propose DaCoM, a methodology for collecting low-resource language pairs from industrial domains to address these challenges. DaCoM is a hybrid translation framework enabling effective data collection. The framework consists of a large language model and neural machine translation. Evaluation verifies existing models perform inadequately on DaCoM-created datasets, with up to 53.7 BLEURT points difference depending on domain inclusion. DaCoM is expected to address the lack of datasets for domain-specific low-resource languages by being easily pluggable into future state-of-the-art models and maintaining an industrial domain-agnostic approach. ©2025 Association for Computational Linguistics.

关键词： Computational linguistics

来源：评论

学校读者我要写书评

暂无评论

Enhancement of Containerization-based Cloud computing Performance using Hybrid Metaheuristic Optimization Algorithm: A Review

Enhancement of Containerization-based Cloud Computing Perfor...

引用

Electronics and Renewable Systems (ICEARS), International Conference on

作者： G. Ramasamy S. Kamalakkannan Department of Computer Science School of Computing Sciences Vels Institute of Science Technology & Advanced Studies Pallavaram Chennai India Department of Information Technology School of Computing Sciences Vels Institute of Science Technology & Advanced Studies Pallavaram Chennai India

ISBN: (数字)9798331509675

ISBN: (纸本)9798331509682

Container-based virtualisation has become a cornerstone of Cloud computing (CC) due to its lightweight and scalable properties when compared to traditional virtual machines. However, it is still difficult to optimising resource allocation in containerised systems, and current approaches frequently fall short in addressing problems like high computational costs, scalability, cold-start scenarios, and Graphics Processing Unit (GPU) management. Container resource allocation is critical for effectively managing computing workloads in cloud settings, enabling scalability, performance optimization, and cost effectiveness. It is extensively used in a various industry, including e-commerce, healthcare, and finance, where dynamic workloads call for efficient resource management. Traditional resource allocation techniques, while their importance, frequently face difficulties such as unequal load distribution and resource underutilization. To address these constraints, this paper suggests a hybrid metaheuristic optimization algorithm that addresses limitations such as network distance, execution time, and migration costs to provide balanced load distribution. This study attempts to analyse the efficiency, time consumption and latency of resource allocation in dynamic cloud systems by addressing existing limitations.

关键词： Cloud computing Costs Scalability Heuristic algorithms Metaheuristics Graphics processing units Containers Dynamic scheduling Resource management Optimization

来源：评论

学校读者我要写书评

暂无评论

MixDiT: Accelerating Image Diffusion Transformer Inference With Mixed-Precision MX Quantization

IEEE Computer Architecture Letters

引用

IEEE Computer Architecture Letters 2025年

作者： Kim, Daeun Hwang, Jinwoo Oh, Changhun Park, Jongse Korea Advanced Institute of Science and Technology School of Computing Graduate School of AI Semiconductor Daejeon34141 Korea Republic of

Diffusion Transformer (DiT) has driven significant progress in image generation tasks. However, DiT inferencing is notoriously compute-intensive and incurs long latency even on datacenter-scale GPUs, primarily due to its iterative nature and heavy reliance on GEMM operations inherent to its encoder-based structure. To address the challenge, prior work has explored quantization, but achieving low-precision quantization for DiT inferencing with both high accuracy and substantial speedup remains an open problem. To this end, this paper proposes MixDiT, an algorithm-hardware co-designed acceleration solution that exploits mixed Microscaling (MX) formats to quantize DiT activation values. MixDiTquantizes the DiT activation tensors by selectively applying higher precision to magnitude-based outliers, which produce mixed-precision GEMM operations. To achieve tangible speedup from the mixed-precision arithmetic, we design a MixDiTaccelerator that enables precision-flexible multiplications and efficient MX precision conversions. Our experimental results show that MixDiTdelivers a speedup of 2.10-5.32× over RTX 3090, with no loss in FID. © 2025 IEEE.

关键词： Signal encoding

来源：评论

学校读者我要写书评

暂无评论

Correcting on-chip distortion of control pulses with silicon spin qubits

引用

Chinese Physics B 2025年第1期34卷 265-271页

作者： Ming Ni Rong-Long Ma Zhen-Zhen Kong Ning Chu Wei-Zhu Liao Sheng-Kai Zhu Chu Wang Gang Luo Di Liu Gang Cao Gui-Lei Wang Hai-Ou Li Guo-Ping Guo CAS Key Laboratory of Quantum Information University of Science and Technology of China(USTC)Hefei 230026China CAS Center for Excellence in Quantum Information and Quantum Physics University of Science and Technology of ChinaHefei 230026China Integrated Circuit Advanced Process Research and Development Center Institute of MicroelectronicsChinese Academy of Sciences(CAS)Beijing 100029China Hefei National Laboratory Hefei 230088China Beijing Superstring Academy of Memory Technology Beijing 100176China Origin Quantum Computing Company Limited Hefei 230026China

In semiconductor quantum dot systems,pulse distortion is a significant source of coherent errors,which impedes qubit characterization and ***,we demonstrate two calibration methods using a two-qubit system as the detector to correct distortion and calibrate the transfer function of the control *** methods are straightforward to implement,robust against noise,and applicable to a wide range of qubit *** two methods differ in correction accuracy and *** first,coarse predistortion(CPD)method,partially mitigates *** second,all predistortion(APD)method,measures the transfer function and significantly enhances exchange oscillation *** methods use exchange oscillation homogeneity as the metric and are suitable for any qubit driven by a diabatic *** believe these methods will enhance qubit characterization accuracy and operation quality in future applications.

关键词： quantum computation quantum dot pulse distortion

来源：评论

学校读者我要写书评

暂无评论

Simulation of Emulsion Evolution in Shear Flow Field Under The Control Of Bidirectional Pulsed Electric Field

SSRN

引用

SSRN 2025年

作者： Wang, Heping Wu, Yi Zhang, Xiaohang Key Laboratory of Engineering Mathematics and Advanced Computing of Nanchang Institute of Technology School of Sciences Nanchang Institute of Technology Nanchang330099 China

This article established a lattice Boltzmann model coupled with electric and shear field. It explored the impact of the combined action of bidirectional pulsed electric field(BPEF) and shear field on the demulsification process. Simulation results reveal that: Under the action of pure shear field, increasing the shear effect will lead oil droplets to move around the flow field center, and forming a stable state eventually. However, after applying bidirectional pulsed electric field on both sides of the flow field, the electric field force breaks this stable state and promotes further aggregation of oil droplets. Analysis of streamline diagrams indicates that: Bidirectional pulsed electric field periodically changes the direction of the electric field force. It leads to induce an unbalanced velocity distribution in the flow field. And this phenomenon can accelerate or decelerate the aggregation of oil droplets. Quantitative analysis results also show the synergistic effect of bidirectional pulsed electric field and shear field on emulsion. It accelerates the aggregation speed of oil droplets and enhances the degree of aggregation. These simulation results are significant important for understand the synergistic demulsification effects of electric and shear fields. © 2025, The Authors. All rights reserved.

关键词： Demulsification

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：