检索结果-内蒙古大学图书馆

arXiv 2022年

作者： Nakwijit, Pakawat Purver, Matthew Cognitive Science Research Group School of Electronic Engineering and Computer Science Queen Mary University of London United Kingdom Department of Knowledge Technologies Jožef Stefan Institute Ljubljana Slovenia

User-generated content is full of misspellings. Rather than being just random noise, we hypothesise that many misspellings contain hidden semantics that can be leveraged for language understanding tasks. This paper presents a fine-grained annotated corpus of misspelling in Thai, together with an analysis of misspelling intention and its possible semantics to get a better understanding of the misspelling patterns observed in the corpus. In addition, we introduce two approaches to incorporate the semantics of misspelling: Misspelling Average Embedding (MAE) and Misspelling Semantic Tokens (MST). Experiments on a sentiment analysis task confirm our overall hypothesis: additional semantics from misspelling can boost the micro F1 score up to 0.4-2%, while blindly normalising misspelling is harmful and suboptimal. © 2022, CC BY.

关键词： Semantics

来源：评论

学校读者我要写书评

暂无评论

Picosecond optical-to-x-ray (Cd,Mg)Te radiation detectors

Picosecond optical-to-x-ray (Cd,Mg)Te radiation detectors

引用

2023 Conference on Lasers and Electro-Optics, CLEO 2023

作者： Cheng, J. Komissarov, I. Chen, G. Chakraborty, D. Kutcher, S. Chen, H. Trivedi, S. Sobolewski, Roman Materials Science Graduate Program University of Rochester RochesterNY14627 United States Laboratory for Laser Energetics University of Rochester RochesterNY14623 United States Department of Electrical and Computer Engineering University of Rochester RochesterNY14627 United States Brimrose Technology Corp SparksMD21152 United States

We developed a dual optical/x-ray ultrafast photodetector based on in-house grown Cdo * Mg0.03Te single crystals. The detector is characterized by ~200 ps full-width-at-half-maximum, readout-electronics limited photor... 详细信息

ISBN: (纸本)9781957171258

关键词： X ray detectors

来源：评论

学校读者我要写书评

暂无评论

Exploring the science and Art of UAV Light Painting: From Equations and Pixels to Long-Exposure Photography

Exploring the Science and Art of UAV Light Painting: From Eq...

引用

International Conference on Unmanned Aircraft Systems (ICUAS)

作者： Leonardo A. Fagundes-Junior Celso O. Barcelos Raquel F. Vassallo Alexandre S. Brandão Department of Electrical Engineering Robotics Specialization Center Graduate Program on Computer Science Federal University of Viçosa Viçosa MG Brazil Department of Electrical Engineering Federal University of Espirito Santo Vitória ES Brazil

ISBN: (数字)9798350357882

ISBN: (纸本)9798350357899

The use of Unmanned Aerial Vehicles (in short, UAVs, aka drones) for cultural and entertainment purposes, such as drone light shows, has grown exponentially. One such innovative and creative application is the visual arts using drones to explore long-exposure photography. Light painting are generally performed indoors and outdoors in a dedicated space with a human using a moving light source to create spectacular images and save the movement perception in a picture. In this work, we propose a robotic perception system designed to choreograph UAV movements based on time-parametric curves or image edges, serving as reference motions. Our framework begins by processing a digital image, extracting its contours through boundary tracing, and subsequently generating a safe, navigable, and precise path for UAV motion planning. This process involves optimizing waypoints within the UAV workspace to determine a feasible trajectory that encompasses all designated points or computes safe trajectories utilizing established mathematical equations. The validation of the motion planning is performed through light painting, where the UAV can either fly through the motion reference to mimic the original image. The generated trajectories on light painting mode by physical robots are compared against the ground-truth demonstrating the accuracy of the applied control scheme.

关键词： Photography Visualization Art Autonomous aerial vehicles Trajectory Planning Robots

来源：评论

学校读者我要写书评

暂无评论

Democratizing engineering education: An online Bachelor of science in engineering degree 53

Democratizing engineering education: An online Bachelor of S...

引用

53rd IEEE ASEE Frontiers in Education International Conference, FIE 2023

作者： Stites, Nick A. Yowell, Janet L. Hoenigman, Rhonda Kuskin, William Rose, Abigail B. University of Colorado Boulder Integrated Teaching & Learning Program BoulderCO United States University of Colorado Boulder Computer Science Department BoulderCO United States University of Colorado Boulder English Department BoulderCO United States University of Colorado Boulder Engineering Digital Education BoulderCO United States

ISBN: (纸本)9798350336429

The College of engineering & Applied science at the University of Colorado Boulder proposes a low-cost, online, accredited Bachelor of science in engineering (BSE) degree. The BSE will provide a pathway for community college (CC) students, or students with similar transfer credits, to complete a degree in engineering at a cost similar to attending CC. The educationally underserved communities across Colorado are often low-income, racially and ethnically diverse, and rural;the BSE's primary goal is to provide access to an engineering education for these students and, thus, open pathways to high-demand, high-wage careers. The BSE will reach an underserved population of learners who can stay in their local community while simultaneously pursuing an engineering education. A fully online, ABET-accredited engineering degree is surprisingly novel-only nine other online engineering degrees in the nation are accredited-but an R1-level university degree at a near-community-college cost is unprecedented. The BSE will invite the institution's best instructors to innovate in online education-virtual labs, team-based projects, flipped asynchronous discussions, and hybrid-learning activities. Student success support will include wrap-around advising to foster the development of the academic, affective, financial, and career support that students need to succeed in engineering. The BSE will utilize five administrative affordances: 1) success coaches to provide wrap-around support and guidance to students;2) course facilitators who provide personalized support;3) stackable credentials;4) eight-week semesters, allowing students to quickly complete portions of their education;and 5) mentorship programming to foster community and to help students develop their (inter)disciplinary identity. The online BSE creates a new pathway for underserved students to matriculate to a flagship public research university for a degree in engineering. This work-in-progress outlines the proposed structur

关键词： community college engineering pathways online education two-year transfer undergraduate education

来源：评论

学校读者我要写书评

暂无评论

Early Detection and Classification of Melanoma Based on Android MobileNet V2 Convolutional Neural Network

Early Detection and Classification of Melanoma Based on Andr...

引用

International Seminar on Application for Technology of Information and Communication (ISemantic)

作者： Jasmine Iswarini Dyah Prabandari Christy Atika Sari Eko Hari Rachmawanto Candra Irawan Bilal R. Altamer Mohamed Ayad Alkhafaji Study Program in Informatics Engineering University of Dian Nuswantoro Semarang Indonesia Study Program in Information Systems University of Dian Nuswantoro Semarang Indonesia Mosul University Presidency University of Mosul Mosul Iraq Department of Computer Science National University of Science and Technology DhiQar Iraq

Melanoma is a malignant form of cancer that affects the skin and has a particularly high mortality rate, so it requires early detection to increase the level of safety for users. Diagnosis and detection of skin cancer are usually done through manual screening and visual inspection. This process requires a long time, has high complexity, is subjective, and is prone to errors. CNN is one of the algorithms with advantages in accurate classification. In this research, early detection and classification of melanoma cancer were carried out based on two classes, namely benign and malignant using the Convolutional Neural Network method. Our proposed method yields an accuracy of 81.11% for the validation data. The accuracy results obtained can be improved by using more datasets and increasing the number of layers used. This study uses the CNN method using MobileNet V2 architecture to detect melanoma skin cancer. The class used is benign and malignant.

关键词：

来源：评论

学校读者我要写书评

暂无评论

A Hybrid Rolling and Flying Robot for Pipeline Fault Detection Using Deep Learning

A Hybrid Rolling and Flying Robot for Pipeline Fault Detecti...

引用

Advances in science and engineering Technology International Conferences (ASET)

作者： Megan Ghaly Rohan Mitra Abdullah Al Rayess Assem Ahmed Dara Varam Mohammad A. Jaradat Michel Pasquier Ahmed Khalil Department of Mechanical Engineering American University of Sharjah Sharjah UAE Department of Computer Science and Engineering American University of Sharjah Sharjah UAE Mechatronics Engineering Gradute Program American University of Sharjah Sharjah UAE

The Middle Eastern and North African region is highly reliant on the oil and gas industry. Subsequently, the need for pipeline inspection and fault diagnosis has become paramount. Current inspection methods rely on manual procedures that introduce sources of error into the results. In this paper, the design for a hybrid rolling and flying unmanned aerial vehicle (UAV) for pipeline inspection is proposed. The UAV was developed with a rolling frame capable of landing on pipelines to externally inspect them for cracks, leaks, or faults by traversing along the pipeline's surface while maintaining its stability. The hybrid UAV uses a camera to classify cracks along the surface of the pipeline using deep learning models. The paper then tests several models and deploys the best model found during testing -MobileNet100 along with the proposed design for the UAV. Finally, the prototype that was presented in this paper has been tested within a controlled environment to verify the compliance of the flying, rolling and fault detection subsystems.

关键词：

来源：评论

学校读者我要写书评

暂无评论

Low-Crosstalk in 2D Waveguide Crossings by Metaheuristic Optimization Methods

Low-Crosstalk in 2D Waveguide Crossings by Metaheuristic Opt...

引用

Optics and Photonics Conference (SBFoton IOPC), SBFoton International

作者： Luana da França Vieira C. E. Rubio-Mercedes Vitaly Felix Rodriguez Esquerre Department of Computer Science Federal University of Western Bahia Barreiras Brazil Physics Engineering Program State University of Mato Grosso do Sul Dourados Brazil Graduate School of Electrical Engineering Federal University of Bahia Salvador Brazil

ISBN: (数字)9798350388176

ISBN: (纸本)9798350388183

This work describes the investigation of planar waveguide crossings using metaheuristic methods associated with the two-dimensional finite element method (2D-FEM). An optimization and measurement process are performed on the power transmission and crosstalk of the waveguide crossings and then the values calculated are compared with other studies in the literature. The best performing adjusted crossing is optimally designed by applying the following metaheuristics: Particle Swarm Optimization (PSO), Grey Wolf Optimizer (GWO), and a hybrid algorithm combining the Greedy Randomized Adaptive Search Procedure and Simulated Annealing (GRASP-SA). Transmission efficiency results higher than 97% and crosstalk below the -50 dB over a wide wavelength range, with a footprint of $\mathbf{4.9}\times \mathbf{4.9}\ \boldsymbol{\mu} \mathbf{m}^{\mathbf{2}}$ .

关键词： Performance evaluation Power measurement Metaheuristics Crosstalk Power transmission Simulated annealing Silicon Planar waveguides Optical waveguides Photonics

来源：评论

学校读者我要写书评

暂无评论

Air-Ground Integrated Online Federated Learning Under Unreliable Communication

引用

IEEE Transactions on cognitive Communications and Networking 2025年

作者： Jing, Yuqian Qu, Yuben Wu, Tao Dong, Chao Guo, Song Wu, Qihui Ministry of Industry and Information Technology Key Laboratory of Dynamic Cognitive System of Electromagnetic Spectrum Space China Nanjing University of Aeronautics and Astronautics College of Electronic and Information Engineering Nanjing211106 China Hong Kong Polytechnic University Department of Computing Hong Kong Hong Kong National University of Defense Technology Anhui Hefei230009 China Hong Kong University of Science and Technology Department of Computer Science and Engineering Hong Kong999077 Hong Kong

In recent years, unmanned aerial vehicles (UAVs) equipped with intelligent computing modules for machine learning (ML) have garnered attention. Federated learning (FL), as an emerging distributed learning paradigm, aligns seamlessly with this scenario. UAVs train local models while a terrestrial base station (BS) aggregates the global model, forming an air-ground integrated federated learning (AGIFL) system. However, under air-ground integrated networks with unreliable wireless links, the online retraining of ML models with new samples is promising but challenging in AGIFL. To this end, we study how to enhance the performance of air-ground integrated online FL (AGIOFL) considering packet transmission error. Specifically, we formulate a joint optimization problem for sample selection and client scheduling, aiming to minimize the loss function that maximizes the performance of AGIOFL. To solve the aforementioned problem without explicit expression, we first derive a closed-form expression for the expected convergence rate of the FL algorithm to quantify the impact of AGIOFL and unreliable wireless transmission. Based on this relationship, we reformulate the optimization problem into a mixed-integer nonlinear programming problem. Subsequently, we develop an alternating iterative optimization algorithm (OFL-PE) combining branch-and-bound with the Dinkelbach algorithm to solve this problem. Simulations indicate that OFL-PE converges within 10 iterations only, with a maximum error of less than 7% compared to the exhaustive brute force approach. In addition, we construct a proof-of-concept system based on widely used UAV embedded computers, and experimental results show that OFL-PE demonstrates superior accuracy, convergence, and robustness compared to three existing FL algorithms. © 2015 IEEE.

关键词： Federated learning

来源：评论

学校读者我要写书评

暂无评论

Fireworks Image Classification with Deep Learning 26

Fireworks Image Classification with Deep Learning

引用

26th International Conference on Technologies and Applications of Artificial Intelligence, TAAI 2021

作者： Chang, Chih-Lung Tseng, Hsin-Ming Chu, Hsueh-Ting Asia University Department of Computer Science and Information Engineering Taichung Taiwan Asia University Ph.D. Program in Artificial Intelligence Taichung Taiwan

ISBN: (纸本)9781665408257

Before the advent of deep learning, traditional image recognition used algorithms to find features and then classify images using classical machine learning algorithms. But it is difficult to define features for variable types of images. Instead, Representation learning is a way to allow a system to discover the representations of feature for image processing. Deep-learning algorithms attempt to learn multiple levels of representation and play the key role of modern Representation learning. When deep learning neural networks recognize images, the shallow layers usually extract lower-level features, then start with intermediate-level features, and finally the full images. This study uses CNN (Convolutional Neural Network) to classify fireworks images, and the trained modules are used to evaluate the accuracy and suitability. © 2021 IEEE.

关键词： Image classification

来源：评论

学校读者我要写书评

暂无评论

On the Impact of Knowledge Distillation for Model Interpretability

arXiv

引用

arXiv 2023年

作者： Han, Hyeongrok Kim, Siwon Choi, Hyun-Soo Yoon, Sungroh Department of Electrical and Computer Engineering Seoul National University Seoul Korea Republic of Department of Computer Science and Engineering Seoul National University of Science and Technology Seoul Korea Republic of ZIOVISION Inc. Chuncheon Korea Republic of Interdisciplinary Program in Artificial Intelligence Seoul National University Seoul Korea Republic of

Several recent studies have elucidated why knowledge distillation (KD) improves model performance. However, few have researched the other advantages of KD in addition to its improving model performance. In this study, we have attempted to show that KD enhances the interpretability as well as the accuracy of models. We measured the number of concept detectors identified in network dissection for a quantitative comparison of model interpretability. We attributed the improvement in interpretability to the class-similarity information transferred from the teacher to student models. First, we confirmed the transfer of class-similarity information from the teacher to student model via logit distillation. Then, we analyzed how class-similarity information affects model interpretability in terms of its presence or absence and degree of similarity information. We conducted various quantitative and qualitative experiments and examined the results on different datasets, different KD methods, and according to different measures of interpretability. Our research showed that KD models by large models could be used more reliably in various fields. © 2023, CC BY-NC-SA.

关键词： Distillation

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：