检索结果-内蒙古大学图书馆

Adaptive Convolutional Neural Network for Image Super-resolution

学校读者我要写书评

暂无评论

arXiv 2024年

作者： Tian, Chunwei Zhang, Xuanyu Wang, Tao Zhang, Yongjun Zhu, Qi Lin, Chia-Wen The School of Software Northwestern Polytechnical University Xi’an710129 China The National Engineering Laboratory for Integrated Aero-Space-Ground-Ocean Big Data Application Technology Xi’an710129 China The Yangtze River Delta Research Institute Northwestern Polytechnical University Taicang215400 China The School of Computer Science Northwestern Polytechnical University Xi’an710129 China The College of Computer Science and Technology Guizhou University Guiyang550025 China The School of Artificial Intelligence Nanjing University of Aeronautics and Astronautics Nanjing210016 China The Department of Electrical Engineering The Institute of Communications Engineering National Tsing Hua University Taiwan

Convolutional neural networks can automatically learn features via deep network architectures and given input samples. However, the robustness of obtained models may face challenges in varying scenes. bigger differences in network architecture are beneficial to extract more diversified structural information to strengthen the robustness of an obtained superresolution model. In this paper, we proposed a adaptive convolutional neural network for image super-resolution (ADSRNet). To capture more information, ADSRNet is implemented by a heterogeneous parallel network. The upper network can enhance relation of context information, salient information relation of a kernel mapping and relations of shallow and deep layers to improve performance of image super-resolution. That can strengthen adaptability of an obtained super-resolution model for different scenes. The lower network utilizes a symmetric architecture to enhance relations of different layers to mine more structural information, which is complementary with a upper network for image super-resolution. The relevant experimental results show that the proposed ADSRNet is effective to deal with image resolving. Codes are obtained at https://***/hellloxiaotian/ADSRNet. Copyright © 2024, The Authors. All rights reserved.

关键词： Convolutional neural networks

Learning discrete representations via constrained clustering for effective and efficient dense retrieval

学校读者我要写书评

暂无评论

arXiv 2021年

作者： Zhan, Jingtao Mao, Jiaxin Liu, Yiqun Guo, Jiafeng Zhang, Min Ma, Shaoping Department of Computer Science and Technology Institute for Artificial Intelligence Beijing National Research Center for Information Science and Technology Tsinghua University Beijing100084 China Beijing Key Laboratory of Big Data Management and Analysis Methods Gaoling School of Artificial Intelligence Renmin University of China Beijing100872 China CAS Key Lab of Network Data Science and Technology Institute of Computing Technology Chinese Academy of Sciences Beijing China

Dense Retrieval (DR) has achieved state-of-the-art first-stage ranking effectiveness. However, the efficiency of most existing DR models is limited by the large memory cost of storing dense vectors and the time-consuming nearest neighbor search (NNS) in vector space. Therefore, we present RepCONC, a novel retrieval model that learns discrete Representations via CONstrained Clustering. RepCONC jointly trains dual-encoders and the Product Quantization (PQ) method to learn discrete document representations and enables fast approximate NNS with compact indexes. It models quantization as a constrained clustering process, which requires the document embeddings to be uniformly clustered around the quantization centroids and supports end-to-end optimization of the quantization method and dual-encoders. We theoretically demonstrate the importance of the uniform clustering constraint in RepCONC and derive an efficient approximate solution for constrained clustering by reducing it to an instance of the optimal transport problem. Besides constrained clustering, RepCONC further adopts a vector-based inverted file system (IVF) to support highly efficient vector search on CPUs. Extensive experiments on two popular ad-hoc retrieval benchmarks show that RepCONC achieves better ranking effectiveness than competitive vector quantization baselines under different compression ratio settings. It also substantially outperforms a wide range of existing retrieval models in terms of retrieval effectiveness, memory efficiency, and time efficiency. Copyright © 2021, The Authors. All rights reserved.

关键词： Signal encoding

MetaGCD: Learning to Continually Learn in Generalized Category Discovery

学校读者我要写书评

暂无评论

arXiv 2023年

作者： Wu, Yanan Chi, Zhixiang Wang, Yang Feng, Songhe Key Laboratory of Big Data & Artificial Intelligence in Transportation Ministry of Education Beijing Jiaotong University Beijing100044 China School of Computer and Information Technology Beijing Jiaotong University Beijing100044 China Department of Electrical and Computer Engineering University of Toronto TorontoM5G1V7 Canada Department of Computer Science and Software Engineering Concordia University MontrealH3G2J1 Canada

In this paper, we consider a real-world scenario where a model that is trained on pre-defined classes continually encounters unlabeled data that contains both known and novel classes. The goal is to continually discover novel classes while maintaining the performance in known classes. We name the setting Continual Generalized Category Discovery (C-GCD). Existing methods for novel class discovery cannot directly handle the C-GCD setting due to some unrealistic assumptions, such as the unlabeled data only containing novel classes. Furthermore, they fail to discover novel classes in a continual fashion. In this work, we lift all these assumptions and propose an approach, called MetaGCD, to learn how to incrementally discover with less forgetting. Our proposed method uses a meta-learning framework and leverages the offline labeled data to simulate the testing incremental learning process. A meta-objective is defined to revolve around two conflicting learning objectives to achieve novel class discovery without forgetting. Furthermore, a soft neighborhood-based contrastive network is proposed to discriminate uncorrelated images while attracting correlated images. We build strong baselines and conduct extensive experiments on three widely used benchmarks to demonstrate the superiority of our method. Our code is available at https://***/ynanwu/MetaGCD. © 2023, CC0.

关键词： Learning systems

Interpreting dense retrieval as mixture of topics

学校读者我要写书评

暂无评论

arXiv 2021年

Dense Retrieval (DR) reaches state-of-the-art results in first-stage retrieval, but little is known about the mechanisms that contribute to its success. Therefore, in this work, we conduct an interpretation study of recently proposed DR models. Specifically, we first discretize the embeddings output by the document and query encoders. Based on the discrete representations, we analyze the attribution of input tokens. Both qualitative and quantitative experiments are carried out on public test collections. Results suggest that DR models pay attention to different aspects of input and extract various high-level topic representations. Therefore, we can regard the representations learned by DR models as a mixture of high-level topics. Copyright © 2021, The Authors. All rights reserved.

关键词： Mixtures

FuXi Weather: A data-to-forecast machine learning system for global weather

学校读者我要写书评

暂无评论

arXiv 2024年

作者： Sun, Xiuyu Zhong, Xiaohui Xu, Xiaoze Huang, Yuanqing Li, Hao Neelin, J. David Chen, Deliang Feng, Jie Han, Wei Wu, Libo Qi, Yuan Shanghai Academy of Artificial Intelligence for Science Shanghai200232 China Artificial Intelligence Innovation and Incubation Institute Fudan University Shanghai200433 China School of Atmospheric Physics Nanjing University of Information Science and Technology Nanjing210044 China Earth System Modeling and Prediction Centre China Meteorological Administration Beijing100081 China Department of Atmospheric and Oceanic Sciences University of California Los Angeles90095 United States Ministry of Education Key Laboratory for Earth System Modeling Department of Earth System Science Tsinghua University Beijing100084 China University of Gothenburg Gothenburg Sweden Department of Atmospheric and Oceanic Sciences Institute of Atmospheric Sciences Fudan University Shanghai200433 China School of Data Science Fudan University Shanghai200433 China Institute for Big Data Fudan University Shanghai200433 China MOE Laboratory for National Development and Intelligent Governance Fudan University Shanghai200433 China

Weather forecasting traditionally relies on numerical weather prediction (NWP) systems that integrates global observational systems, data assimilation (DA), and forecasting models. Despite steady improvements in forecast accuracy over recent decades, further advances are increasingly constrained by high computational costs, the underutilization of vast observational datasets, and the challenges of obtaining finer resolution. These limitations, alongside the uneven distribution of observational networks, result in global disparities in forecast accuracy, leaving some regions vulnerable to extreme weather. Recent advances in machine learning present a promising alternative, providing more efficient and accurate forecasts using the same initial conditions as NWP. However, current machine learning models still depend on the initial conditions generated by NWP systems, which require extensive computational resources and expertise. Here we introduce FuXi Weather, a machine learning weather forecasting system that assimilates data from multiple satellites. Operating on a 6-hourly DA and forecast cycle, FuXi Weather generates reliable and accurate 10-day global weather forecasts at a spatial resolution of 0.25◦. FuXi Weather is the first system to achieve all-grid, all-surface, all-channel, and all-sky DA and forecasting, extending skillful forecast lead times beyond those of the European Centre for Medium-range Weather Forecasts (ECMWF) high-resolution forecasts (HRES) while using significantly fewer observations. FuXi Weather consistently outperforms ECMWF HRES in observation-sparse regions, such as central Africa, demonstrating its potential to improve forecasts where observational infrastructure is limited. © 2024, CC BY.

关键词： Weather forecasting

A parameter adaptive differential evolution based on depth information

学校读者我要写书评

暂无评论

A parameter adaptive differential evolution based on depth i...

作者： Meng, Zhenyu Yang, Cheng Meng, Fanjia Chen, Yuxin Lin, Fang Institute of Artificial Intelligence Fujian University of Technology Fuzhou China Fujian Provincial Key Laboratory of Big Data Mining and Applications Fujian University of Technology Fuzhou China Intelligent Information Processing Research Center Fujian University of Technology Fuzhou China Guanzhuang Central Primary School of Zhangqiu District Jinan China

Differential Evolution (DE) was an easy-coding and efficient stochastic algorithm for global optimization, and the whole optimization process simulates biological evolution. Superior individuals of the population that were suitable for the environment were retained during the evolution, and consequently the tolerable solutions could be obtained in the end. Despite the excellent performance of DE algorithm, there were still some shortcomings. For example, the general performance of DE depended largely on mutation strategy and control parameters, how to design the appropriate control parameters and mutation strategy were difficult tasks. Here a novel DE variant was proposed to overcome these shortcomings. By incorporating the depth information of previous generations of populations, a better diversity of trial vector candidates could be secured during the evolution process. Moreover, the thought that successful parameters should be retained to guide the update of themselves during the evolution was also incorporated into the novel algorithm. The optimization performance of the new proposed DE variant was verified under CEC 2013 test suit containing 28 benchmarks, and the results showed its competitiveness with several state-of-the-art DE variants. © 2020 - IOS Press and the authors. All rights reserved.

关键词： Global optimization

A heterogeneous group CNN for image super-resolution

学校读者我要写书评

暂无评论

arXiv 2022年

作者： Tian, Chunwei Zhang, Yanning Zuo, Wangmeng Lin, Chia-Wen Zhang, David Yuan, Yixuan The School of Software Northwestern Polytechnical University Shaanxi Xi’an710129 China The National Engineering Laboratory for Integrated Aero-Space-Ground-Ocean Big Data Application Technology Shaanxi Xi’an710129 China The School of Computer Science Northwestern Polytechnical University The National Engineering Laboratory for Integrated Aero-Space-Ground-Ocean Big Data Application Technology Shaanxi Xi’an710129 China The School of Computer Science and Technology Harbin Institute of Technology Heilongjiang Harbin150001 China The Department of Electrical Engineering The Institute of Communications Engineering National Tsing Hua University Hsinchu Taiwan Guangdong Shenzhen518172 China The Shenzhen Institute of Artificial Intelligence and Robotics for Society Shenzhen China The Department of Electronic Engineering The Chinese University of Hong Kong Hong Kong

Convolutional neural networks (CNNs) have obtained remarkable performance via deep architectures. However, these CNNs often achieve poor robustness for image superresolution (SR) under complex scenes. In this paper, we present a heterogeneous group SR CNN (HGSRCNN) via leveraging structure information of different types to obtain a high-quality image. Specifically, each heterogeneous group block (HGB) of HGSRCNN uses a heterogeneous architecture containing a symmetric group convolutional block and a complementary convolutional block in a parallel way to enhance internal and external relations of different channels for facilitating richer low-frequency structure information of different types. To prevent appearance of obtained redundant features, a refinement block with signal enhancements in a serial way is designed to filter useless information. To prevent loss of original information, a multi-level enhancement mechanism guides a CNN to achieve a symmetric architecture for promoting expressive ability of HGSRCNN. Besides, a parallel up-sampling mechanism is developed to train a blind SR model. Extensive experiments illustrate that the proposed HGSRCNN has obtained excellent SR performance in terms of both quantitative and qualitative analysis. Codes can be accessed at https://***/hellloxiaotian/HGSRCNN. © 2022, CC BY-NC-SA.

关键词： Convolutional neural networks

JGC-IAGCL: Fusing joint graph convolution and intent-aware graph contrastive learning for explainable recommendation

学校读者我要写书评

暂无评论

Information Fusion 2025年 123卷

作者： Yang, Zhi Lin, Chuan Qin, Yongbin Huang, Ruizhang Chen, Yanping Qin, Jiwei State Key Laboratory of Public Big Data Guizhou University Guiyang Guizhou550025 China Text Computing & Cognitive Intelligence Engineering Research Center of National Education Ministry College of Computer Science and Technology Guizhou University Guiyang Guizhou550025 China College of Information Science and Engineering Xinjiang University Xinjiang Uygur Autonomous Region Urumqi830046 China

Graph contrastive learning (GCL) enhances recommendation accuracy by leveraging self-supervised features to refine node representations from large-scale unlabeled data. Traditional GCL-based recommendation models typically construct contrastive views via graph augmentation (e.g., stochastic node/edge dropout) or embedding-space perturbation, aiming to maximize representation consistency. However, these methods struggle to effectively model and interpret user preferences and consumption intents, limiting explainability and recommendation performance. To address these challenges, we propose JGC-IAGCL (Joint Graph Convolution and Intent-Aware Graph Contrastive Learning), an explainable recommendation framework. JGC-IAGCL integrates joint graph convolution to capture implicit user preferences and employs intent-aware graph contrastive learning to extract explicit user intents from user–item interactions. By fusing these features, our method generates evenly distributed, intent-propensity-aware user/item representations. Theoretical analysis shows that JGC-IAGCL mitigates popularity bias while enhancing the exposure of long-tail items. Extensive experiments on four highly sparse public datasets validate its effectiveness, demonstrating superior recommendation accuracy and improved interpretability. © 2025

关键词： Self-supervised learning

A parameter adaptive de algorithm on real-parameter optimization

学校读者我要写书评

暂无评论

A parameter adaptive de algorithm on real-parameter optimiza...

作者： Pan, Jeng-Shyang Yang, Cheng Meng, Fanjia Chen, Yuxin Meng, Zhenyu College of Computer Science and Engineering Shandong University of Science and Technology Qingdao China Institute of Artificial Intelligence Fujian University of Technology Fuzhou China Guanzhuang Central Primary School of Zhangqiu District Jinan China Fujian Provincial Key Laboratory of Big Data Mining and Applications Fujian University of Technology Fuzhou China

Differential Evolution (DE) algorithm generates a population of individuals by encoding with a floating point vector, and it is a simple and effective population-based stochastic optimization algorithm for global optimization of continuous space. Because of its excellent performance, DE variants can be applied in a wide range of applications in science and engineering. However, the performance of DE is sensitive to the choice of trial vector generation strategy and the associated control parameters. Therefore, it is necessary to choose appropriate mutation strategy and control parameters when tackling optimization applications. In this paper, an adaptive update mechanism is proposed to update control parameters F and Cr. The experimental results are verified on the CEC 2013 test suite which contains 28 benchmark functions for the evaluation of single objective real parameter optimization. The proposed algorithm is compared with jDE, iwPSO and ccPSO, and experiment results show its good performance. © 2020 - IOS Press and the authors. All rights reserved.

关键词： Global optimization