检索结果-内蒙古大学图书馆

Smart Photonic and Optoelectronic Integrated Circuits 2025

作者： Tan, Bo Xue Wang, Jiantao Poon, Andrew W. Photonic Device Laboratory Department of Electronic and Computer Engineering The Hong Kong University of Science and Technology Clear Water Bay Hong Kong

ISBN: (纸本)9781510684881

Aluminum gallium arsenide (AlGaAs) is a promising photonic platform for photon-pair generation through spontaneous parametric down-conversion (SPDC) or spontaneous four-wave mixing due to its high optical nonlinearities. In this paper, we report our progress in the design and fabrication of high-quality-factor waveguide-coupled microring resonators in an AlGaAs-on-insulator (AlGaAsOI) platform using i-line (365 nm) photolithography. We design the waveguide geometries for the modal-phase matching for SPDC. Our preliminary results suggest the potential for efficient photon-pair generation in an AlGaAsOI platform for quantum photonic applications. © 2025 SPIE.

关键词： Four wave mixing

来源：评论

学校读者我要写书评

暂无评论

State Definition in Deep Reinforcement Learning: A Smoothness Index Approach 11

State Definition in Deep Reinforcement Learning: A Smoothnes...

引用

11th RSI International Conference on Robotics and Mechatronics, ICRoM 2023

作者： Dehdasthian, Soroosh Kalhor, Ahmad Masouleh, Mehdi Tale University of Tehran Human and Robot Intraction Laboratory School of Electrical and Computer Engineering Department Iran

ISBN: (纸本)9798350308105

In the pursuit of developing autonomous agents capable of learning optimal behaviors, reinforcement learning has emerged as a foundational frame-work. This paper introduces a novel approach for state definition in deep reinforcement learning systems, with a focus on solving pixel-based control scenarios. The proposed frame-work harnesses the capabilities of Convolutional Variational Auto-Encoders to extract compact and meaningful latent representations from high-dimensional pixel observations. These representations serve as the foundation for constructing an efficient state space for reinforcement learning agents. To further enhance the state construction process, the concept of the Smoothness Index is introduced. Smoothness index acts as a criterion for feature selection within the latent space, facilitating the identification of the most relevant and informative states. Demonstrating the effectiveness of iterative smoothness index-based forward selection, this process notably compresses the observation space while retaining essential information. This compression enables the creation of more efficient reinforcement learning agents. The procedure involves collecting and preprocessing observation data given by the OpenAI Gym car racing environment, encoding observations using a convolutional variational auto-encoder, and iteratively selecting features based on smoothness index. Subsequently, the Proximal Policy Optimization algorithm is employed as a policy-based controller to evaluate the effectiveness of the proposed frame-work. Results indicate that this approach yields a remarkably compact state space that either maintains or enhances reinforcement learning agent performance. This innovative frame-work paves the way for the development of more efficient and adaptive reinforcement learning agents in real-world image-based applications. © 2023 IEEE.

关键词： Reinforcement learning

来源：评论

学校读者我要写书评

暂无评论

Adjusting Logit in Gaussian Form for Long-Tailed Visual Recognition

IEEE Transactions on Artificial Intelligence

引用

IEEE Transactions on Artificial Intelligence 2024年第10期5卷 5026-5039页

作者： Li, Mengke Cheung, Yiu-Ming Lu, Yang Hu, Zhikai Lan, Weichao Huang, Hui Shenzhen518132 China Hong Kong Baptist University Department of Computer Science 999077 Hong Kong Xiamen University Fujian Key Laboratory of Sensing and Computing for Smart City School of Informatics Xiamen361005 China Shenzhen University College of Computer Science and Software Engineering Shenzhen518060 China

It is not uncommon that real-world data are distributed with a long tail. For such data, the learning of deep neural networks becomes challenging because it is hard to classify tail classes correctly. In the literature, several existing methods have addressed this problem by reducing classifier bias, provided that the features obtained with long-tailed data are representative enough. However, we find that training directly on long-tailed data leads to uneven embedding space. That is, the embedding space of head classes severely compresses that of tail classes, which is not conducive to subsequent classifier learning. This article therefore studies the problem of long-tailed visual recognition from the perspective of feature level. We introduce feature augmentation to balance the embedding distribution. The features of different classes are perturbed with varying amplitudes in Gaussian form. Based on these perturbed features, two novel logit adjustment methods are proposed to improve model performance at a modest computational overhead. Subsequently, the distorted embedding spaces of all classes can be calibrated. In such balanced-distributed embedding spaces, the biased classifier can be eliminated by simply retraining the classifier with class-balanced sampling data. Extensive experiments conducted on benchmark datasets demonstrate the superior performance of the proposed method over the state-of-the-art ones. © 2020 IEEE.

关键词： Benchmarking

来源：评论

学校读者我要写书评

暂无评论

On the Design of Novel Attention Mechanism for Enhanced Efficiency of Transformers 24

On the Design of Novel Attention Mechanism for Enhanced Effi...

引用

61st ACM/IEEE Design Automation Conference, DAC 2024

作者： Jha, Sumit Kumar Jha, Susmit Ewetz, Rickard Velasquez, Alvaro Computer Science Department Florida International University MiamiFL United States Computer Science Laboratory Sri International Menlo ParkCA United States Electrical and Computer Engineering University of Central Florida OrlandoFL United States Department of Computer Science University of Colorado Boulder BoulderCO United States

ISBN: (纸本)9798400706011

We present a new xor-based attention function for efficient hardware implementation of transformers. While the standard attention mechanism relies on matrix multiplication between the key and the transpose of the query, we propose replacing the computation of this attention function with bitwise xor operations. We mathematically analyze the information-theoretic properties of the standard multiplication-based attention, demonstrating that it preserves input entropy, and then computationally show that the xor-based attention approximately preserves the entropy of its input despite small variations in correlations between the inputs. Across various admittedly simple tasks, including arithmetic, sorting, and text generation, we show comparable performance to baseline methods using scaled GPT models. The xor-based computation of the attention function shows substantial improvement in power consumption, latency, and circuit area compared to the corresponding multiplication-based attention function. This hardware efficiency makes xor-based attention more compelling for the deployment of transformers under tight resource constraints, opening new application domains in sustainable energy-efficient computing. Additional optimizations to the xor-based attention function can further improve efficiency of transformers. © 2024 Copyright is held by the owner/author(s). Publication rights licensed to ACM.

关键词： Distribution transformers

来源：评论

学校读者我要写书评

暂无评论

Broadband spin and angle co‑multiplexed waveguide‑based metasurface for six‑channel crosstalk‑free holographic projection

引用

eLight 2024年第1期4卷 223-237页

作者： Zeyang Liu Hao Gao Taigao Ma Vishva Ray Niu Liu Xinliang Zhang L.Jay Guo Cheng Zhang School of Optical and Electronic Information Huazhong University of Science and TechnologyWuhan 430074China Wuhan National Laboratory for Optoelectronics Huazhong University of Science and TechnologyWuhan 430074China Department of Physics University of MichiganAnn ArborMI 48109USA Lurie Nanofabrication Facility Department of Electrical Engineering and Computer ScienceUniversity of MichiganAnn ArborMI 48109USA Department of Electrical Engineering and Computer Science University of MichiganAnn ArborMI 48109USA

Metasurface-based holograms,or metaholograms,offer unique advantages including enhanced imaging quality,expanded field of view,compact system size,and broad operational ***-channel metaholograms,capable of switching between multiple projected images based on the properties of illuminating light such as state of polarization and angle of incidence,have emerged as a promising solution for realizing switchable and dynamic holographic ***,existing designs typically grapple with challenges such as limited multiplexing channels and unwanted crosstalk,which severely constrain their practical ***,we present a new type of waveguidebased multi-channel metaholograms,which support six independent and fully crosstalk-free holographic display channels,simultaneously multiplexed by the spin and angle of guided incident light within the glass *** employ a k-space translation strategy that allows each of the six distinct target images to be selectively translated from evanescent-wave region to the center of propagation-wave region and projected into free space without crosstalk,when the metahologram is under illumination of a guided light with specific spin and azimuthal *** addition,by tailoring the encoded target images,we implement a three-channel polarization-independent metahologram and a two-channel full-color(RGB)***,the number of multiplexing channels can be further increased by expanding the k-space’s central-period region or combing the k-space translation strategy with other multiplexing techniques such as orbital angular momentum *** work provides a novel approach towards realization of high-performance and compact holographic optical elements with substantial information capacity,opening avenues for applications in AR/VR displays,image encryption,and information storage.

关键词： Metasurface Multi-channel metahologram Color metahologram Waveguide display

来源：评论

学校读者我要写书评

暂无评论

Generalized Few-shot 3D Point Cloud Segmentation with vision-Language Model

arXiv

引用

arXiv 2025年

作者： An, Zhaochong Sun, Guolei Liu, Yun Li, Runjia Han, Junlin Konukoglu, Ender Belongie, Serge Department of Computer Science University of Copenhagen Denmark Computer Vision Laboratory ETH Zurich Switzerland College of Computer Science Nankai University China Department of Engineering Science University of Oxford United Kingdom

Generalized few-shot 3D point cloud segmentation (GFS-PCS) adapts models to new classes with few support samples while retaining base class segmentation. Existing GFS-PCS methods enhance prototypes via interacting with support or query features but remain limited by sparse knowledge from few-shot samples. Meanwhile, 3D vision-language models (3D VLMs), generalizing across open-world novel classes, contain rich but noisy novel class knowledge. In this work, we introduce a GFS-PCS framework that synergizes dense but noisy pseudo-labels from 3D VLMs with precise yet sparse few-shot samples to maximize the strengths of both, named GFS-VL. Specifically, we present a prototype-guided pseudo-label selection to filter low-quality regions, followed by an adaptive infilling strategy that combines knowledge from pseudo-label contexts and few-shot samples to adaptively label the filtered, unlabeled areas. Additionally, we design a novel-base mix strategy to embed few-shot samples into training scenes, preserving essential context for improved novel class learning. Moreover, recognizing the limited diversity in current GFS-PCS benchmarks, we introduce two challenging benchmarks with diverse novel classes for comprehensive generalization evaluation. Experiments validate the effectiveness of our framework across models and datasets. Our approach and benchmarks provide a solid foundation for advancing GFS-PCS in the real world. The code is at here. © 2025, CC BY-NC-SA.

关键词： Visual languages

来源：评论

学校读者我要写书评

暂无评论

Transient Angle Stability of Inverters Equipped with Robust Droop Control

引用

CSEE Journal of Power and Energy Systems 2023年第2期9卷 659-670页

作者： Chen Qi Keyou Wang Qing-Chang Zhong Jin Xu Guojie Li Key Laboratory of Control of Power Transmission and Conversion Ministry of EducationShanghai Jiao Tong UniversityShanghai 200240China Department of Electrical and Computer Engineering Illinois Institute of TechnologyChicagoIllinois60616USA

Transient angle stability of inverters equipped with the robust droop controller is investigated in this *** first,the conditions on the control references to guarantee the existence of a feasible post-disturbance operating point are ***,the post-disturbance equilibrium points are found and their stability properties are ***,the attraction regions of the stable equilibrium points are accurately depicted by calculating the stable and unstable manifolds of the surrounding unstable equilibrium points,which presents an explanation to system transient ***,the transient control considerations are provided to help the inverter ridethrough the disturbance and maintain its stability *** is shown that the transient angle stability is not a serious problem for droop controlled inverters with proper control settings.

关键词： Droop controlled inverter equilibrium point transient angle stability (un)stable manifold

来源：评论

学校读者我要写书评

暂无评论

On the Attractive and Repulsive Forces of Generalized Stochastic Neighbor Embedding With Alpha-Divergence

引用

IEEE Access 2024年 12卷 90380-90394页

作者： Lin, Hsin-Yi Tseng, Huan-Hsin Chien, Jen-Tzung Seton Hall University Department Of Mathematics And Computer Science South OrangeNJ07079 United States Brookhaven National Laboratory Computational Science Initiative UptonNY11973 United States Institute Of Electrical And Computer Engineering National Yang Ming Chiao Tung University Hsinchu30010 Taiwan

Stochastic neighbor embedding (SNE) performs nonlinear transformation from high-dimensional observation space to low-dimensional latent space which preserves neighbor affinities. Data pairs in latent space tend to be crowded due to the dimensionality reduction. To mitigate the crowding problem, certain characteristics are favorable in the design of the SNE setting. This study presents a fundamental analysis of SNE that not only generalizes the previous SNEs but also provides a systematic way to understand the intrinsic properties. From the perspective of theoretical connection, we are able to conceive a new generalized SNE (g-SNE) by introducing a regularized power-law distribution with the α-divergence for manifold learning. The proposed method generalizes and incorporates various favorable features for the clustering process. In addition, the proposed method provides high flexibility, admitting tailored realizations to properly reflect the similarity between original and dimension-reduced samples. Experiments are performed to analyze the proposed method, and its effectiveness is demonstrated with several learning tasks. © 2024 The Authors.

关键词： Data visualization

来源：评论

学校读者我要写书评

暂无评论

Advancing Intrusion Detection Systems Innovative Approaches for Prevention and Response

Advancing Intrusion Detection Systems Innovative Approaches ...

引用

2024 International Conference on Communication, computer Sciences and engineering, IC3SE 2024

作者： Reddy, M. Siva Swetha Krishna, B.S.S. Murali Asha, V. Nijhawan, Ginni Rana, Ajay Jabbar, Ahmed Sabah Abed Al-Zahra Institute Of Aeronautical Engineering Department Of Computer Science And Engineering Telangana Hyderabad India Mlr Institute Of Technology Department Of Computer Science And Engineering Telangana Hyderabad India Computer Application New Horizon College Of Engineering Bangalore India Lovely Professional University Phagwara India Amity School Of Engineering And Technology Amity University Greater Noida India The Islamic University Medical Laboratory Technology Department Najaf Iraq

ISBN: (纸本)9798350366846

A group intruder detection system built on K-Nearest Neighbors, Decision Trees, Neural Networks, Support Vector Machines, and Random Forests is shown in this study. A thorough study on ablation shows how the algorithms work together to make breach detection better. We compare how well the suggested method works in terms of memory, F1 score, false negative rate, accuracy, and detection time. Bar charts, line graphs, pie charts, stacked bar graphs, area plots, and scatter plots all show that the new way is better than the old ones. Tables are also used to organize data. An ensemble takes the best features of Random Forests, Support Vector Machines, Neural Networks, Decision Trees, and K-Nearest Neighbors and puts them all together in a single package. Broad powers make the suggested method better at solving cybersecurity problems across a range of review factors. The ablation study shows why each method is important and how they work together to make an adaptive intruder detection system. In a world where safety is always changing, the proposed answer is scalable, flexible, easy to set up, and cheap. © 2024 IEEE.

关键词： Nearest neighbor search

来源：评论

学校读者我要写书评

暂无评论

Deep learning:Applications,architectures,models,tools,and frameworks:A comprehensive survey

引用

CAAI Transactions on Intelligence Technology 2023年第3期8卷 581-606页

作者： Mehdi Gheisari Fereshteh Ebrahimzadeh Mohamadtaghi Rahimi Mahdieh Moazzamigodarzi Yang Liu Pijush Kanti Dutta Pramanik Mohammad Ali Heravi Abolfazl Mehbodniya Mustafa Ghaderzadeh Mohammad Reza Feylizadeh Saeed Kosari School of Computer Science and Technology Harbin Institute of Technology(Shenzhen)ShenzhenChina Department of Cognitive Computing Institute of Computer Science and EngineeringSaveetha School of EngineeringSaveetha Institute of Medical and Technical SciencesChennaiIndia Department of Computer Science Islamic Azad UniversityTehranIran Computer Engineering Department Bu‐Ali Sina UniversityHamedanIran Department of Mathematics and Statistics Iran University of Science and TechnologyTehranIran Department of Mathematics and Statistics University of Northern British ColumbiaBCCanada Peng Cheng Laboratory ShenzhenChina Deprtment of Technology Jodhpur Institute of Engineering&TechnologyJodhpurIndia Department of Computer Engineering Sadjad University of TechnologyMashhadIran Department of Electronics and Communications Engineering Kuwait College of Science and TechnologyDoha DistrictKuwait Department of Artificial Intelligence Smart University of Medical SciencesTehranIran Department of Industrial Engineering Shiraz BranchIslamic Azad UniversityShirazIran Institute of Computing Science and Technology Guangzhou UniversityGuangzhouChina

Deep Learning(DL)is a subfield of machine learning that significantly impacts extracting new *** using DL,the extraction of advanced data representations and knowledge can be made *** effective DL techniques help to find more hidden *** learning has a promising future due to its great performance and *** need to understand the fundamentals and the state‐of‐the‐art of DL to leverage it effectively.A survey on DL ways,advantages,drawbacks,architectures,and methods to have a straightforward and clear understanding of it from different views is explained in the ***,the existing related methods are compared with each other,and the application of DL is described in some applications,such as medical image analysis,handwriting recognition,and so on.

关键词： data mining data privacy deep learning

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：