检索结果-内蒙古大学图书馆

Exploring data driven soliton and rogue waves in symmetric and spatio-temporal potentials using PINN and SC-PINN methods

引用

The European Physical Journal Plus 2025年第3期140卷 1-27页

作者： Anand, R. Manikandan, K. Serikbayev, N. Centre for Computational Modeling Chennai Institute of Technology Chennai India Center for Nonlinear and Complex Networks SRM TRP Engineering College Tiruchirappalli India Center for Research SRM Institute of Science and Technology - Tiruchirappalli Tiruchirappalli India Department of General and Theoretical Physics L. N. Gumilyov Eurasian National University Astana Kazakhstan

In this study, we present a deep learning (DL) framework for solving the nonlinear Schrödinger equation with $$\mathcal{P}\mathcal{T}$$ -symmetric potentials using strongly constrained physics-informed neural networks (SC-PINNs). We focus on three types of physically compelling potentials, namely $$\mathcal{P}\mathcal{T}$$ -symmetric rational, Jacobian-periodic, and spatio-temporal dependent. SC-PINNs extend the standard physics-informed neural networks (PINNs) by incorporating compound derivative information into the soft constraints, along with an adaptive weight mechanism to accelerate loss function convergence. This enhancement significantly improves the training efficiency compared to standard PINNs. We employ SC-PINNs to approximate soliton and rogue wave solutions of the system under investigation. Additionally, we uncover the impact of various factors on the neural network’s performance, including five different nonlinear activation functions: ReLU, sigmoid, sech, tanh, and sine. Our results reveal that the SC-PINNs method achieves faster convergence and lower errors compared to traditional PINNs. Notably, when using the sine activation function for the three distinct potentials mentioned above, SC-PINNs reduced errors to the order of $$10^{-7}$$ , $$10^{-6}$$ , and $$10^{-4}$$ , effectively capturing complex physical features for highly accurate predictions. Furthermore, we analyze the effect of $$\mathcal{P}\mathcal{T}$$ -symmetric potential parameters on the obtained approximated solutions. The results demonstrate that our DL model successfully approximates soliton and rogue wave solutions of the considered system with high accuracy, outperforming traditional DL algorithms.

关键词：

来源：评论

学校读者我要写书评

暂无评论

HIGH ACCURACY TECHNIQUES BASED ADAPTIVE FINITE ELEMENT METHODS FOR ELLIPTIC PDES

arXiv

引用

arXiv 2025年

作者： Xiao, Jingjing Liu, Ying Yi, Nianyu School of Mathematics Shandong University Jinan250100 China School of Science Xi’an University of Technology Xi’an710048 China School of Mathematics and Computational Science Xiangtan University Hunan National Center for Applied Mathematics Hunan Key Laboratory for Computation and Simulation in Science and Engineering Xiangtan411105 China

This paper aims to develop an efficient adaptive finite element method for the second-order elliptic problem. Although the theory for adaptive finite element methods based on residual-type a posteriori error estimator and bisection refinement has been well established, in practical computations, the use of non-asymptotic exact of error estimator and the excessive number of adaptive iteration steps often lead to inefficiency of the adaptive algorithm. We propose an efficient adaptive finite element method based on high-accuracy techniques including the superconvergence recovery technique and high-quality mesh optimization. The centroidal Voronoi Delaunay triangulation mesh optimization is embedded in the mesh adaption to provide high-quality mesh, and then assure that the superconvergence property of the recovered gradient and the asymptotical exactness of the error estimator. A tailored adaptive strategy, which could generate high-quality meshes with a target number of vertices, is developed to ensure the adaptive computation process terminated within 7 steps. The effectiveness and robustness of the adaptive algorithm is numerically demonstrated. Copyright © 2025, The Authors. All rights reserved.

关键词： Triangulation

来源：评论

学校读者我要写书评

暂无评论

Co2 Adsorption and Activation on Auo(Co2)N−/+ (N = 1 – 3) Clusters: A Theoretical Study

SSRN

引用

SSRN 2025年

作者： Huang, Wei Zhao, Wenbao Guo, Zonghui Du, Shihu Tian, Jincheng Zhang, Ruoying Han, Haiyan Zhao, Zhi Pei, Wei Shi, Ruili Xie, Hua School of Mathematics and Physics Science and Engineering Hebei Computational Optical Imaging and Photoelectric Detection Technology Innovation Center Hebei International Joint Research Center for Computational Optical Imaging and Intelligent Sensing Hebei University of Engineering Handan056038 China School of Chemistry and Chemical Engineering Shandong University Jinan250100 China State Key Laboratory of Molecular Reaction Dynamics Dalian Institute of Chemical Physics Chinese Academy of Sciences Dalian116023 China College of Physics Science and Technology Yangzhou University Yangzhou225009 China

The geometric and electronic properties of AuO(CO2)n−/+ (n = 1 − 3) clusters were studied using density functional theory (DFT). Our results reveal that the coordination of the AuO unit with CO2 significantly alters CO2 geometry, generating diverse motifs. In the anionic clusters, strong Au coordination induces pronounced distortions of the CO2 moiety, resulting in CO3-like configurations. Whereas the cationic clusters retain near-linear CO2 geometries with localized charge distributions. The binding energies of both cluster types show a gradual decreasing trend, due to electrostatic stabilization. Natural population analysis (NPA) further demonstrates that the charged state and cluster size play crucial roles in modulating the electronic structure of CO2, with anions favoring covalent activation and cations maintaining CO2 linearity via electrostatic interactions. These findings enhance our fundamental understanding of CO2 activation on gold-based catalysts and provide valuable insights for the rational design of efficient catalysts for CO2 reduction and conversion processes. © 2025, The Authors. All rights reserved.

关键词： Transition metal oxides

来源：评论

学校读者我要写书评

暂无评论

MFF-YOLO: An Improved YOLO Algorithm Based on Multi-Scale Semantic Feature Fusion

引用

Tsinghua science and Technology 2025年第5期30卷 2097-2113页

作者： Junsan Zhang Chenyang Xu Shigen Shen Jie Zhu Peiying Zhang Qingdao Institute of Software College of Computer Science and Technology China University of Petroleum (East China) Qingdao China School of Information Engineering Huzhou University Huzhou China Hebei Key Laboratory of Machine Learning and Computational Intelligence College of Mathematics and Information Science Hebei University Baoding China National Research Center for Information Science and Technology Tsinghua University Beijing China

The YOLOv5 algorithm is widely used in edge computing systems for object detection. However, the limited computing resources of embedded devices and the large model size of existing deep learning based methods increase the difficulty of real-time object detection on edge devices. To address this issue, we propose a smaller, less computationally intensive, and more accurate algorithm for object detection. Multi-scale Feature Fusion-YOLO (MFF-YOLO) is built on top of the YOLOv5s framework, but it contains substantial improvements to YOLOv5s. First, we design the MFF module to improve the feature propagation path in the feature pyramid, which further integrates the semantic information from different paths of feature layers. Then, a large convolution-kernel module is used in the bottleneck. The structure enlarges the receptive field and preserves shallow semantic information, which overcomes the performance limitation arising from uneven propagation in Feature Pyramid Networks (FPN). In addition, a multi-branch downsampling method based on depthwise separable convolutions and a bottleneck structure with deformable convolutions are designed to reduce the complexity of the backbone network and minimize the real-time performance loss caused by the increased model complexity. The experimental results on PASCAL VOC and MS COCO datasets show that, compared with YOLOv5s, MFF-YOLO reduces the number of parameters by 7% and the number of FLoating point Operations Per second (FLOPs) by 11.8%. The mAP@0.5 has improved by 3.7% and 5.5%, and the mAP@0.5:0.95 has improved by 6.5% and 6.2%, respetively. Furthermore, compared with YOLOv7-tiny, PP-YOLO-tiny, and other mainstream methods, MFF-YOLO has achieved better results on multiple indicators.

关键词： YOLO Deformable models Accuracy computational modeling Image edge detection Semantics Neural networks Feature extraction Real-time systems Edge computing

来源：评论

学校读者我要写书评

暂无评论

Information-Maximized Soft Variable Discretization for Self-Supervised Image Representation Learning

arXiv

引用

arXiv 2025年

作者： Niu, Chuang Xia, Wenjun Shan, Hongming Wang, Ge Department of Biomedical Engineering Center for Biotechnology and Interdisciplinary Studies Rensselaer Polytechnic Institute TroyNY12180 United States Institute of Science and Technology for Brain-inspired Intelligence and MOE Frontiers Center for Brain Science Key Laboratory of Computational Neuroscience and Brain-Inspired Intelligence Fudan University Shanghai200433 China Shanghai Center for Brain Science and Brain-inspired Technology Shanghai201210 China

Self-supervised learning (SSL) has emerged as a crucial technique in image processing, encoding, and understanding, especially for developing today’s vision foundation models that utilize large-scale datasets without annotations to enhance various downstream tasks. This study introduces a novel SSL approach, Information-Maximized Soft Variable Discretization (IMSVD), for image representation learning. Specifically, IMSVD softly discretizes each variable in the latent space, enabling the estimation of their probability distributions over training batches and allowing the learning process to be directly guided by information measures. Motivated by the MultiView assumption, we propose an information-theoretic objective function to learn transform-invariant, non-travail, and redundancy-minimized representation features. We then derive a joint-cross entropy loss function for self-supervised image representation learning, which theoretically enjoys superiority over the existing methods in reducing feature redundancy. Notably, our non-contrastive IMSVD method statistically performs contrastive learning. Extensive experimental results demonstrate the effectiveness of IMSVD on various downstream tasks in terms of both accuracy and efficiency. Thanks to our variable discretization, the embedding features optimized by IMSVD offer unique explainability at the variable level. IMSVD has the potential to be adapted to other learning paradigms. Our code is publicly available at https://***/niuchuangnn/IMSVD. © 2025, CC BY.

关键词： Self-supervised learning

来源：评论

学校读者我要写书评

暂无评论

Dynamic Broad Metric Learning

IEEE Transactions on Artificial Intelligence

引用

IEEE Transactions on Artificial Intelligence 2025年

作者： Hu, Xiaoman Chen, C. L. Philip Zhang, Tong South China University of Technology Guangdong Provincial Key Laboratory of Computational AI Models and Cognitive Intelligence The School of Computer Science and Engineering Guangzhou510006 China Pazhou Lab Guangzhou510335 China Engineering Research Center of the Ministry of Education on Health Intelligent Perception and Paralleled Digital-Human Guangzhou China

Many existing metric learning methods are based on fixed similarity constraints. However, the quality of fixed similarity constraints is usually hard to guarantee, and inflexible constraints also limit the performance of metric learning. Moreover, when new training samples are acquired, retraining the model is usually inefficient for many classical metric learning methods. This paper proposes a novel Dynamic Broad Metric Learning (DynBML) method to overcome the above limitations. DynBML trains a broad network with dynamic similarity constraints between samples and their target points to learn a discriminative data transformation with high representation ability. To overcome the weaknesses of fixed constraints, DynBML dynamically updates the target points to produce new constraints during the learning process. A Force-Directed Reference Points Generation (FD-RPG) algorithm is designed to generate reference points from the currently learned space, which encourage the new target points to be more discriminative. A new objective function is proposed to simultaneously learn the new target points and train the data transformation with the guidance of the reference points. Additionally, to avoid retraining the model from scratch when new data arrive, the incremental learning scheme of DynBML (I-DynBML) is developed, which efficiently updates the trained model of DynBML by generating and satisfying new constraints of new input. DynBML is evaluated on ten datasets to demonstrate its superiority. Experiment results show that the dynamic constraints constructed in DynBML improve the learning performance. The effectiveness and efficiency of the incremental learning scheme of DynBML are also verified. © 2020 IEEE.

关键词： Contrastive Learning

来源：评论

学校读者我要写书评

暂无评论

Atomic-vibration-induced nonlinear electronic polarization for terahertz detection

引用

Physical Review B 2025年第8期111卷 085201-085201页

作者： Zhi Li Hussain Irshad Yu Gu Miriding Mutailipu Toshiaki Iitaka Haibin Su Xiangang Wan School of Material Science and Engineering Nanjing University of Science and Technology Nanjing 210094 China Research Center for Crystal Materials Xinjiang Technical Institute of Physics and Chemistry Chinese Academy of Sciences Urumqi 830011 China Discrete Event Simulation Research Team RIKEN Center for Computational Science 2-1 Hirosawa Wako Saitama 351-0198 Japan Department of Chemistry The Hong Kong University of Science and Technology Hong Kong China National Laboratory of Solid State Microstructures and School of Physics Nanjing University Nanjing China Collaborative Innovation Center of Advanced Microstructures Nanjing University Nanjing China

Terahertz (THz) has broad applications for imaging, spectroscopy, and high-data-rate communication. However, THz detection at room temperature remains a significant challenge. By model Hamiltonian including electron-phonon coupling and photon-electron interaction within dipole approximation, we demonstrate that atomic displacement A⃗(t) driven by THz field E⃗(t) can induce static nonlinear electronic polarization in semiconductors without spatial inversion symmetry by nonlinear optic effect (αijkEjEk), Raman effect (βijkAjEk), and nonlinear phononic effect (γijkAjAk). Our numerical calculations also reveals that nonlinear phononic effect dominates the change of static polarization in ferroelectric semiconductor BaTiO3 irradiated by THz field. This work presents a design principle for efficient THz detection at room temperature by monitoring the change of surface polarization induced by THz-driven lattice vibration.

关键词： Lattice vibrations

来源：评论

学校读者我要写书评

暂无评论

Robustness of Vacancy-Bound Non-Abelian Anyons in the Kitaev Model in a Magnetic Field

arXiv

引用

arXiv 2025年

作者： Xiao, Bo Alvarez, Gonzalo Halász, Gábor B. Materials Science and Technology Division Oak Ridge National Laboratory Oak RidgeTN37831 United States Quantum Science Center Oak RidgeTN37831 United States Computational Science and Engineering Division Oak Ridge National Laboratory Oak RidgeTN37831 United States

The search for non-Abelian anyons in quantum spin liquids (QSLs) is crucial for advancing fault-tolerant topological quantum computation. In the exactly solvable Kitaev honeycomb model, nonmagnetic spin vacancies are known to bind emergent gauge fluxes of the QSL ground state, which in turn become non-Abelian anyons for an infinitesimal magnetic field. Here, we investigate how this approach for stabilizing non-Abelian anyons at spin vacancies extends to a finite magnetic field represented by a proper Zeeman term. Specifically, we use large-scale density-matrix renormalization group (DMRG) simulations to compute the vacancy-anyon binding energy as a function of magnetic field for both the ferromagnetic (FM) and antiferromagnetic (AFM) Kitaev models. We find that, while the inclusion of the field weakens anyon binding in both cases, there is a pronounced difference in the finite-field behavior;the binding energy remains finite within the entire QSL phase for the FM Kitaev model but approaches zero already inside this phase for the AFM Kitaev model. To reliably compute a very small binding energy that is three orders of magnitude below the magnetic interaction strength, we also introduce a refined definition for the binding energy and an extrapolation scheme for its accurate extraction through exclusively ground-state properties using DMRG. © 2025, CC BY.

关键词： Ground state

来源：评论

学校读者我要写书评

暂无评论

Key Clues Guided Video Character Social Relationship Recognition Enhanced by LLM

Key Clues Guided Video Character Social Relationship Recogni...

引用

International Conference on Acoustics, Speech, and Signal Processing (ICASSP)

作者： Wenlong Dong Qing Zhu Qirong Mao School of Computer Science and Communication Engineering Jiangsu University China Jiangsu Engineering Research Center of Big Data Ubiquitous Perception and Intelligent Agriculture Applications China Provincial Key Laboratory of Computational Intelligence and New Technologies in Low-Altitude Digital Agriculture Zhenjiang China

ISBN: (数字)9798350368741

ISBN: (纸本)9798350368758

Video Character Social Relationship Recognition (VCSRR) requires a comprehensive consideration about spatio-temporal and multi-modal clues in videos. Most existing methods mainly focus on integrating multi-modal clues and modeling interactions among characters. However, they fail to discover key clues in the complex video data or fully understand the clues related to social relationships. In this article, we propose a novel Large Language Model Enhanced Key Clues Selection (LE-KCS) framework to address the aforementioned issues. The core of LE-KCS is to mine multi-scale key clues from the perspectives of time, space and multi-modality, then transfer the knowledge about social relationships of the Large Language Model to VCSRR for understanding the selected clues. We evaluated LE-KCS on the MovieGraphs dataset and the experimental results indicate that our proposed LE-KCS achieves state-of-the-art performance.

关键词： Accuracy Large language models Speech recognition Signal processing Acoustics Character recognition Speech processing

来源：评论

学校读者我要写书评

暂无评论

Global Enhanced Frame Prompt Tuning for Sound Event Detection

Global Enhanced Frame Prompt Tuning for Sound Event Detectio...

引用

International Conference on Acoustics, Speech, and Signal Processing (ICASSP)

作者： Shiyu Yu Lijian Gao Qirong Mao School of Computer Science and Communication Engineering Jiangsu University Zhenjiang China Jiangsu Engineering Research Center of Big Data Ubiquitous Perception and Intelligent Agriculture Applications Provincial Key Laboratory of Computational Intelligence and New Technologies in Low-Altitude Digital Agriculture Zhenjiang China

ISBN: (数字)9798350368741

ISBN: (纸本)9798350368758

Sound Event Detection (SED) often employs pre-trained models to address data scarcity issues. However, existing systems usually treat the pretrained models as frozen feature extractors, resulting in suboptimal efficiency, or fully fine-tune the pretrained models, which requires substantial computational resources. To fully leverage the knowledge from pretrained models, we propose a novel Global Enhanced Frame Prompt Tuning (GE-FPT) framework, providing global and local insights tailored for SED tasks. Additionally, Frame Prompt Tuning (FPT) is proposed in our GE-FPT to effectively explore local temporal information, i.e., temporal details and context, which is essential for SED tasks, and in particular, for precise event boundary detection. Extensive experiments claim that our approach significantly outperforms full fine-tuning methods while substantially reducing computational costs. Our system achieves new state-of-the-art results, with PSDS1/PSDS2 scores of 0.628/0.845 on the DCASE2023 Challenge Task4 dataset. The source code is publicly available 1 .

关键词： Event detection computational modeling Source coding Signal processing Feature extraction Data models Acoustics computational efficiency Speech processing Tuning

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：