检索结果-内蒙古大学图书馆

Congress on Evolutionary Computation

作者： Wenjing Hong Cheng Chen Zexuan Zhu Ke Tang National Engineering Laboratory for Big Data System Computing Technology Shenzhen University Shenzhen China Research Institute of Trustworthy Autonomous Systems Southern University of Science and Technology Shenzhen China Department of Computer Science and Engineering Southern University of Science and Technology Shenzhen China

ISBN: (数字)9798350308365

ISBN: (纸本)9798350308372

Messenger RNA (mRNA) vaccines have emerged as highly effective strategies in the prophylaxis and treatment of diseases. mRNA design, a key to the success of mRNA vaccines, in-volves finding optimal codons and increasing secondary structure stability to lengthen mRNA half-life, ultimately enhancing protein expression. Despite receiving widespread attention, most methods primarily rely on manual design, which is time-consuming and labor-intensive. While optimization approaches can alleviate this issue, existing methods still exhibit critical limitations caused by conflicts between codon usage and mRNA structural stability, compounded by the vast design space of mRNA resulting from the presence of synonymous codons. In this paper, a novel multi-objective evolutionary optimization-based mRNA design method is proposed. We first formulate the mRNA design problem as a multi-objective optimization problem and then develop an Elite Archive-Assisted Multi-Objective Evolutionary algorithm for mRNA Design, namely EAA-MOED, by incorporating a novel elite archive-assisted method into a weighted optimization framework to improve search efficiency. Experimental studies, involving two state-of-the-art mRNA design methods and five well-known MOEAs, show the competitiveness of the proposed EAA-MOED in mRNA design.

关键词： Measurement Design methodology Evolutionary computation Search problems Linear programming Stability analysis Vaccines

来源：评论

学校读者我要写书评

暂无评论

Algorithmic Stability of Heavy-Tailed Stochastic Gradient Descent on Least Squares 34

Algorithmic Stability of Heavy-Tailed Stochastic Gradient De...

引用

34th International Conference onAlgorithmic Learning Theory, ALT 2023

作者： Raj, Anant Barsbey, Melih Gürbüzbalaban, Mert Zhu, Lingjiong Şimşekli, Umut Coordinated Science Laboratory University of Illinois Urbana-Champaign Inria Ecole Normale Supérieure PSL Research University Paris France Department of Computer Engineering Bogaziçi University Istanbul Turkey Department of Management Science and Information Systems Rutgers University PiscatawayNJ United States Department of Mathematics Florida State University TallahasseeFL United States Inria CNRS Ecole Normale Supérieure PSL Research University Paris France

Recent studies have shown that heavy tails can emerge in stochastic optimization and that the heaviness of the tails have links to the generalization error. While these studies have shed light on interesting aspects of the generalization behavior in modern settings, they relied on strong topological and statistical regularity assumptions, which are hard to verify in practice. Furthermore, it has been empirically illustrated that the relation between heavy tails and generalization might not always be monotonic in practice, contrary to the conclusions of existing theory. In this study, we establish novel links between the tail behavior and generalization properties of stochastic gradient descent (SGD), through the lens of algorithmic stability. We consider a quadratic optimization problem and use a heavy-tailed stochastic differential equation (and its Euler discretization) as a proxy for modeling the heavy-tailed behavior emerging in SGD. We then prove uniform stability bounds, which reveal the following outcomes: (i) Without making any exotic assumptions, we show that SGD will not be stable if the stability is measured with the squared-loss x 7→ x2, whereas it in turn becomes stable if the stability is instead measured with a surrogate loss x 7→ |x|p with some p © 2023 A. Raj, M. Barsbey, M. Gürbüzbalaban, L. Zhu & U. Şimşekli.

关键词： Gradient methods

来源：评论

学校读者我要写书评

暂无评论

Impact of waveguide imperfections on lasing near a stationary inflection point

Impact of waveguide imperfections on lasing near a stationar...

引用

2023 Active Photonic Platforms, APP 2023

作者： Furman, Nathaniel Herrero-Parareda, Albert Thompson, Bradley Brown, Dean Gibson, Ricky Vitebskiy, Ilya Bedford, Robert Capolino, Filippo Department of Electrical Engineering and Computer Science University of California IrvineCA92617 United States Air Force Research Laboratory Sensors Directorate Wright-Patterson Air Force Base OH45433 United States Air Force Research Laboratory Materials and Manufacturing Directorate Wright-Patterson Air Force Base OH45433 United States UES Inc. DaytonOH45432 United States

ISBN: (纸本)9781510665088

We investigate the effect of fabrication tolerances on photonic multimode waveguides operating in the vicinity of a third-order exceptional point degeneracy (EPD), known as a stationary inflection point (SIP). An EPD is a point in the parameter space where two or more Bloch eigenmodes coalesce in an infinite periodic waveguide, and at an SIP three modes coalesce to form the frozen mode. Waveguides operating near an SIP exhibit slow-light behavior in finite-length waveguides with anomalous cubic scaling of the group delay with waveguide length. The frozen mode facilitates stronger light-matter interactions in active media, resulting in a significant increase in the effective gain within the cavity. However, systems operating near an EPD are also exceptionally sensitive to fabrication deviations. In this work, we explore wave propagation and the impact of various fabrication imperfections in analytic models and in fabricated photonic chips for three mirrorless devices operating near an SIP. To advance the concept of the SIP laser, we also analyze how the addition of gain and loss affects the SIP performance. Our results show that while minor deviations from the ideal parameters can prevent perfect mode coalescence at the EPD, the frozen mode remains resilient to small perturbations and a significant degree of mode degeneracy prevails. These findings provide critical insights into the design and fabrication of passive and active photonic devices operating near high-order EPDs, paving the way for their practical implementation in a wide range of applications. © 2023 SPIE.

关键词： Fabrication

来源：评论

学校读者我要写书评

暂无评论

Roles of UAVs, Robots, and AI in Earthquake Disaster Management 7

Roles of UAVs, Robots, and AI in Earthquake Disaster Managem...

引用

7th International Youth Conference on Radio Electronics, electrical and Power engineering, REEPE 2025

作者： Abdalzaher, Mohamed S. Shaaban, Mostafa Krichen, Moez Aburukba, Raafat American University of Sharjah Electrical Engineering Department Sharjah United Arab Emirates National Research Institute of Astronomy and Geophysics Seismology Department Cairo Egypt Energy Water and Sustainable Environment Research Center American University of Sharjah United Arab Emirates Al-Baha University Al-Baha Saudi Arabia ReDCAD Laboratory University of Sfax Sfax Tunisia American University of Sharjah Computer Science and Engineering Department Sharjah United Arab Emirates

ISBN: (纸本)9798331531836

Earthquakes are among the most destructive natural disasters, causing widespread loss of life, property, and infrastructure. Effective disaster management is essential to minimize their impact. This article examines the critical roles of Unmanned Aerial Vehicles (UAVs), robots, and Artificial Intelligence (AI) in earthquake disaster management across the phases of pre-disaster preparedness, real-time response, during the event, and after the disaster. The paper explores various types of UAVs and robots utilized for earthquake monitoring, highlights their advantages, and addresses the challenges associated with their deployment. Besides, the types of robots are also addressed to contribute to earthquake disaster management and to alleviating the consequences. By leveraging these advanced technologies, disaster response and recovery efforts can be significantly enhanced, offering new possibilities for resilience and efficiency in managing earthquake-related emergencies. © 2025 IEEE.

关键词： Disaster prevention

来源：评论

学校读者我要写书评

暂无评论

IEEE T-BIOM Editorial Board Changes

IEEE Transactions on Biometrics, Behavior, and Identity Scie...

引用

IEEE Transactions on Biometrics, Behavior, and Identity science 2024年第1期7卷 1-2页

作者： Naser Damer Weihong Deng Jianjiang Feng Vishal M. Patel Ajita Rattani Mark Nixon Smart Living & Biometric Technologies Fraunhofer Institute for Computer Graphics Research IGD Darmstadt Germany Pattern Recognition and Intelligent System Laboratory Beijing University of Posts and Telecommunications Beijing China Department of Automation Tsinghua University Beijing China Department of Electrical and Computer Engineering Johns Hopkins University Baltimore MD USA Department of Computer Science and Engineering College of Engineering University of North Texas Denton TX USA

来源：评论

学校读者我要写书评

暂无评论

Leveraging ASR Pretrained Conformers for Speaker Verification through Transfer Learning and Knowledge Distillation

arXiv

引用

arXiv 2023年

作者： Cai, Danwei Li, Ming The Department of Electrical and Computer Engineering Duke University DurhamNC27705 United States The Suzhou Municipal Key Laboratory of Multimodal Intelligent Systems Data Science Research Center Duke Kunshan University Kunshan China

This paper focuses on the application of Conformers in speaker verification. Conformers, initially designed for Automatic Speech Recognition (ASR), excel at modeling both local and global contexts within speech signals effectively. Previous research has established that ASR and speaker verification tasks can naturally complement each other. Building on this synergistic relationship, this study introduces three strategies for leveraging ASR-pretrained Conformers in speaker verification: (1) Transfer learning: We use a pretrained ASR Conformer encoder to initialize the speaker embedding network, thereby enhancing model generalization and mitigating the risk of overfitting. (2) Knowledge distillation: We distill the complex capabilities of an ASR Conformer into a speaker verification model. This not only allows for flexibility in the student mode's network architecture but also incorporates frame-level ASR distillation loss as an auxiliary task to reinforce speaker verification. (3) Parameter-efficient transfer learning with speaker adaptation: A lightweight speaker adaptation module is proposed to convert ASR-derived features into speaker-specific embeddings, without altering the core architecture of the original ASR Conformer. This strategy facilitates the concurrent execution of ASR and speaker verification tasks within a singular model. Experiments were conducted on VoxCeleb datasets. The results are compelling: models employing ASR pretraining and knowledge distillation significantly outperform standard Conformers. Specifically, the best model using the ASR pretraining method achieved a 0.43% equal error rate (EER) on the VoxCeleb1-O test trial, while the knowledge distillation approach yielded a 0.38% EER. Furthermore, by adding a mere 4.92 million parameters to a 130.94 million-parameter ASR Conformer encoder, the speaker adaptation approach achieved a 0.45% EER, enabling parallel speech recognition and speaker verification within a single ASR Conformer encoder.

关键词： Transfer learning

来源：评论

学校读者我要写书评

暂无评论

DiffDVC: Accurate Event Detection for Dense Video Captioning via Diffusion Models 39

DiffDVC: Accurate Event Detection for Dense Video Captioning...

引用

39th Annual AAAI Conference on Artificial Intelligence, AAAI 2025

作者： Chen, Wei Niu, Jianwei Liu, Xuefeng Wang, Zhendong Tang, Shaojie Zhu, Guogang State Key Laboratory of Virtual Reality Technology and Systems School of Computer Science and Engineering Beihang University Beijing China Zhongguancun Laboratory Beijing China Zhengzhou University Research Institute of Industrial Technology Zhengzhou University Zhengzhou China Department of Management Science and Systems University at Buffalo BuffaloNY United States

ISBN: (纸本)157735897X

Dense video captioning (DVC) aims to describe multiple events within a video, and its performance is greatly affected by the accuracy of video event detection. Video event detection involves predicting the proposal boundaries (start and end times) and the classification score of each event in a video. Recently, a few methods have applied diffusion models originally designed for image object detection to detect events in DVC. These methods add noise to the ground-truth event proposal boundaries, and subsequently learn the denoising process. However, these methods often overlook the fundamental differences between videos and images. We observe that, whereas in images the important information for object classification is normally around the boundaries of the ground-truth boxes, in videos the key information for event classification is typically centered in the middle of ground-truth event proposals. As a result, the classification module in these existing diffusion models becomes insensitive to boundary changes introduced by the added noise, leading to suboptimal performance. This paper introduces DiffDVC, an innovative diffusion model for DVC. The core of DiffDVC is a boundary-sensitive detector. The detector increases the sensitivity of the classification module to boundary changes by focusing on frames within a specific range around the start and end times of noisy event proposals. Additionally, this range is dynamically adjusted to suit different event proposals. Comprehensive experiments on ActivityNet-1.3, ActivityNet Captions, and YouCook2 datasets show DiffDVC achieving superior performance. Copyright © 2025, Association for the Advancement of Artificial Intelligence (***). All rights reserved.

关键词：

来源：评论

学校读者我要写书评

暂无评论

Once read is enough: domain-specific pretraining-free language models with cluster-guided sparse experts for long-tail domain knowledge 24

Once read is enough: domain-specific pretraining-free langua...

引用

Proceedings of the 38th International Conference on Neural Information Processing systems

作者： Fang Dong Mengyi Chen Jixian Zhou Yubin Shi Yixuan Chen Mingzhi Dong Yujiang Wang Dongsheng Li Xiaochen Yang Rui Zhu Robert Dick Qin Lv Fan Yang Tun Lu Ning Gu Li Shang China and Shanghai Key Laboratory of Data Science School of Computer Science Fudan University Oxford Suzhou Centre for Advanced Research Suzhou China University of Bath and China and Shanghai Key Laboratory of Data Science School of Computer Science Fudan University Microsoft Research Asia Shanghai China School of Mathematics & Statistics University of Glasgow Bayes Business School City St George's University of London Department of Electrical Engineering and Computer Science University of Michigan Department of Computer Science University of Colorado Boulder School of Microelectronics Fudan University

ISBN: (纸本)9798331314385

Language models (LMs) only pretrained on a general and massive corpus usually cannot attain satisfying performance on domain-specific downstream tasks, and hence, applying domain-specific pretraining to LMs is a common and indispensable practice. However, domain-specific pretraining can be costly and time-consuming, hindering LMs' deployment in real-world applications. In this work, we consider the incapability to memorize domain-specific knowledge embedded in the general corpus with rare occurrences and "long-tail" distributions as the leading cause for pretrained LMs' inferior downstream performance. Analysis of Neural Tangent Kernels (NTKs) reveals that those long-tail data are commonly overlooked in the model's gradient updates and, consequently, are not effectively memorized, leading to poor domain-specific downstream performance. Based on the intuition that data with similar semantic meaning are closer in the embedding space, we devise a Cluster-guided Sparse Expert (CSE) layer to actively learn long-tail domain knowledge typically neglected in previous pretrained LMs. During pretraining, a CSE layer efficiently clusters domain knowledge together and assigns long-tail knowledge to designate extra experts. CSE is also a lightweight structure that only needs to be incorporated in several deep layers. With our training strategy, we found that during pretraining, data of long-tail knowledge gradually formulate isolated, "outlier" clusters in an LM's representation spaces, especially in deeper layers. Our experimental results show that only pretraining CSE-based LMs is enough to achieve superior performance than regularly pretrained-finetuned LMs on various downstream tasks, implying the prospects of domain-specific-pretraining-free language models.

关键词：

来源：评论

学校读者我要写书评

暂无评论

Train faster, perform better: modular adaptive training in over-parameterized models 23

Train faster, perform better: modular adaptive training in o...

引用

Proceedings of the 37th International Conference on Neural Information Processing systems

作者： Yubin Shi Yixuan Chen Mingzhi Dong Xiaochen Yang Dongsheng Li Yujiang Wang Robert Dick Qin Lv Yingying Zhao Fan Yang Tun Lu Ning Gu Li Shang China and Shanghai Key Laboratory of Data Science School of Computer Science Fudan University School of Mathematics Statistics University of Glasgow Microsoft Research Asia Shanghai China Department of Engineering Science University of Oxford Department of Electrical Engineering and Computer Science University of Michigan Department of Computer Science University of Colorado Boulder School of Microelectronics Fudan University

Despite their prevalence in deep-learning communities, over-parameterized models convey high demands of computational costs for proper training. This work studies the fine-grained, modular-level learning dynamics of over-parameterized models to attain a more efficient and fruitful training strategy. Empirical evidence reveals that when scaling down into network modules, such as heads in self-attention models, we can observe varying learning patterns implicitly associated with each module's trainability. To describe such modular-level learning capabilities, we introduce a novel concept dubbed modular neural tangent kernel (mNTK), and we demonstrate that the quality of a module's learning is tightly associated with its mNTK's principal eigenvalue λmax. A large λmax indicates that the module learns features with better convergence, while those miniature ones may impact generalization negatively. Inspired by the discovery, we propose a novel training strategy termed Modular Adaptive Training (MAT) to update those modules with their λmax exceeding a dynamic threshold selectively, concentrating the model on learning common features and ignoring those superfluous ones. Unlike most existing training schemes with a complete BP cycle across all network modules, MAT can significantly save computations by its partially-updating strategy and can further improve performance. Experiments show that MAT nearly halves the computational cost of model training and outperforms the accuracy of baselines.

关键词：

来源：评论

学校读者我要写书评

暂无评论

Day-ahead Power Forecasting Model for a Photovoltaic Plant in Macao Based on Weather Classification Using SVM/PCC/LM-ANN

Day-ahead Power Forecasting Model for a Photovoltaic Plant i...

引用

2021 IEEE Sustainable Power and Energy Conference, iSPEC 2021

作者： Zhou, Zhipeng Liu, Li Dai, Ning Yi University of Macau State Key Laboratory of Internet of Things for Smart City Department of Electrical and Computer Engineering China University of Macau Zhuhai Um Science Technology Research Institute State Key Laboratory of Internet of Things for Smart City Department of Electrical and Computer Engineering China

ISBN: (纸本)9781665414395

With the growing demand for clean energy, the world's installed solar energy capacity has increased substantially in the last few years. But the power output of photovoltaic (PV) panels varies greatly under different weather conditions. To improve PV power stations' prediction accuracy, this paper designs a forecasting system that uses artificial neural networks (ANNs) optimized by Levenberg-Marquardt (LM) algorithm based on weather classification for 1-day ahead hourly forecasting. In this process, first, the weather is divided into three types: A - sunny, B - cloudy and C - rainy by using support vector machine (SVM) method. Then, the correlation between meteorological factors and PV power output is analyzed through Pearson correlation coefficient (PCC) method in order to select the forecasting model's input. Finally, the corresponding LM-ANN forecasting sub-models are established under each weather type. After applying the trained three sub-models to a small PV power plant in Macao, the results prove that the proposed forecasting system achieves better prediction effect than traditional backpropagation ANN model. © 2021 IEEE.

关键词： Neural networks

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：