检索结果-内蒙古大学图书馆

Prompt-guided consistency learning for multi-label classification with incomplete labels

Neural networks : the official journal of the International Neural Network Society 2025年 190卷 107604页

作者： Shouwen Wang Qian Wan Zihan Zhang Zhigang Zeng School of Artificial Intelligence and Automation Huazhong University of Science and Technology Wuhan 430074 China Key Lab of Image Processing and Intelligent Control Ministry of Education Wuhan 430074 China. Independent Researcher Wuhan 430074 China. Key Lab of Image Processing and Intelligent Control Ministry of Education Wuhan 430074 China. Electronic address: zgzeng@***.

Addressing insufficient supervision and improving model generalization are essential for multi-label classification with incomplete annotations, i.e., partial and single positive labels. Recent studies incorporate pseudo-labels to provide additional supervision and enhance model generalization. However, the noise in pseudo-labels generated by the model tends to accumulate, resulting in confirmation bias during training. Self-correction methods, commonly used approaches for mitigating confirmation bias, rely on model predictions but remain susceptible to confirmation bias caused by visual confusion, including both visual ambiguity and similarity. To reduce visual confusion, we propose a prompt-guided consistency learning (PGCL) framework designed for two incomplete labeling settings. Specifically, we introduce an intra-category supervised contrastive loss, which imposes consistency constraints on reliable positive class samples in the feature space of each category, rather than across the feature space of all categories, as in traditional inter-category supervised contrastive loss. Building on this, the distinction between true positive and visual confusion samples for each category is enhanced through label-level contrasting of the same category. Additionally, we develop a class-specific semantic decoupling module that leverages CLIP's strong vision-language alignment capability, since the proposed contrastive loss requires high-quality label-level representations as contrastive samples. Extensive experimental results on multiple datasets demonstrate that our method can effectively address the problems of two incomplete labeling settings and achieve state-of-the-art performance.

关键词： Confirmation bias Contrastive learning Incomplete labels Multi-label classification Pseudo-labeling Semantic decoupling

来源：评论

学校读者我要写书评

暂无评论

Self-Organized Criticality in C. Elegans Neural Network

Self-Organized Criticality in C. Elegans Neural Network

引用

Chinese control Conference (CCC)

作者： Xiao Gu Jian Liu Ye Yuan Peng Zhao Kuankuan Xin Tao Fang Department of Automation Shanghai Jiao Tong University Shanghai China Key Laboratory of System Control and Information Processing Ministry of Education Shanghai China Shanghai Engineering Research Centre of Intelligent Control and Management Shanghai China Institute of Machine Intelligence University of Shanghai for Science and Technology Shanghai China Queensland Brain Institute The University of Queensland Brisbane QLD Australia

The self-organized criticality is a fundamental property for organisms to perform normal functional activities. Extensive large-scale biological neural network researches have found the existence of self-organized criticality through avalanche phenomenon, which is typical characteristic of self-organized criticality. However, such research method is not suitable for small-scale biological neural network, because number of neurons in small-scale biological neural network is not large enough to support the emergence of avalanche phenomenon. It has been demonstrated that branching parameter can also be used to describe the propagation of activity in the neural network. Here, in this paper, the C. elegans neural network composed of 302 neurons is selected to investigate the existence of self-organized criticality using branching parameter. Based on the obtained connectome of C. elegans, a directed graph model is constructed to simulate the gradient potential of neurons and the activity of C. elegans, then, the existence of self-organized criticality in the neural network of C. elegans is proved by using branching parameter method. Meanwhile, the result is further confirmed by analyzing the calcium activity data of C. elegans obtained from biological experiments.

关键词：

来源：评论

学校读者我要写书评

暂无评论

Joint Association, Beamforming, and Resource Allocation for Multi-IRS Enabled MU-MISO Systems With RSMA

arXiv

引用

arXiv 2024年

作者： Wang, Chunjie Zhang, Xuhui Xing, Huijun Xue, Liang Wang, Shuqiang Shen, Yanyan Yang, Bo Guan, Xinping Shenzhen Institute of Advanced Technology Chinese Academy of Sciences Shenzhen518055 China School of Information and Electrical Engineering Hebei University of Engineering Handan056038 China Hebei Key Laboratory of Security and Protection Information Sensing and Processing Handan056038 China Shenzhen Future Network of Intelligence Institute The School of Science and Engineering The Guangdong Provincial Key Laboratory of Future Networks of Intelligence The Chinese University of Hong Kong Guangdong Shenzhen518172 China Department of Electrical and Electronic Engineering Imperial College London LondonSW7 2AZ United Kingdom Shenzhen Future Network of Intelligence Institute The Guangdong Provincial Key Laboratory of Future Networks of Intelligence The Chinese University of Hong Kong Guangdong Shenzhen518172 China Department of Automation The Key Laboratory of System Control and Information Processing Ministry of Education Shanghai Jiao Tong University Shanghai200240 China

Intelligent reflecting surface (IRS) and rate-splitting multiple access (RSMA) technologies are at the forefront of enhancing spectrum and energy efficiency in the next generation multi-antenna communication systems. This paper explores a RSMA system with multiple IRSs, and proposes two purpose-driven scheduling schemes, i.e., the exhaustive IRS-aided (EIA) and opportunistic IRS-aided (OIA) schemes. The aim is to optimize the system weighted energy efficiency (EE) under the above two schemes, respectively. Specifically, the Dinkelbach, branch and bound, successive convex approximation, and the semidefinite relaxation methods are exploited within the alternating optimization framework to obtain effective solutions to the considered problems. The numerical findings indicate that the EIA scheme exhibits better performance compared to the OIA scheme in diverse scenarios when considering the weighted EE, and the proposed algorithm demonstrates superior performance in comparison to the baseline algorithms. Copyright © 2024, The Authors. All rights reserved.

关键词： Beamforming

来源：评论

学校读者我要写书评

暂无评论

Fully Circuit Implementation of a two-layer Memristive Neural Network for Pattern Recognition

Fully Circuit Implementation of a two-layer Memristive Neura...

引用

International Conference on Information Science and Technology (ICIST)

作者： Mian Li Xiaoping Wang Zhanfei Chen School of Artificial Intelligence and Automation Huazhong University of Science and Technology Wuhan China School of Artificial Intelligence and Automation Key Laboratory of Image Processing and Intelligent Control of Education Ministry of China Huazhong University of Science and Technology Wuhan China

ISBN: (纸本)9781665412667;9781665429412

In this paper, a fully circuit implementation of Memristive Neural Network (MNN) is proposed. The forward calculation of the network is based on winner-take-all (WTA) mechanism. The weight updating is achieved through the difference of pre-spike and post-spike, which is more close to the biological weight adjustment mechanism. The network is implemented in a full circuit without additional control units. The designed circuit consists of four modules. The memristive crossbar array module with one-memristor (1M) unit structure can effectively calculate the vector-matrix multiplication with only one step. The switch S is replaced by the transistor in the designed leaky-integrate-and-fire (LIF) module, which can control the integration and leakage of the membrane voltage and realize the lateral inhibition between output neurons. Connecting the integrated monostable trigger and the difference circuit, the post-spike generating module can output the required post-spike. The signal switch module realizes the switching of signals connected to memristors by using voltage-controlled switches. The combination of two modules validly realizes weight updating. The functions of four modules are verified separately. We performed a simulation experiment of 5×3 pixels image classification based on the designed circuit in PSPICE. The circuit output results and the high classification accuracy prove the circuit can be effectively applied in pattern recognition. The noise experiment shows the robustness of the designed circuit.

关键词： Training Neurons Switches Robustness Pattern recognition Integrated circuit modeling Task analysis

来源：评论

学校读者我要写书评

暂无评论

A multi-stage semi-supervised learning for ankle fracture classification on CT images

arXiv

引用

arXiv 2024年

作者： Liu, Hongzhi Li, Guicheng Nie, Jiacheng Tang, Hui Yang, Chunfeng Feng, Qianjin Xu, Hailin Chen, Yang The School of Computer Science and Engineering Southeast University Nanjing210096 China The Peking University People's Hospital Beijing100000 China The Guangdong Provincial Key Laboratory of Medical Image Processing School of Biomedical Engineering Southern Medical University Guangzhou510515 China The Key Laboratory of New Generation Artificial Intelligence Technology and its Interdisciplinary Applications Southeast University Ministry of Education China Jiangsu Provincial Joint International Research Laboratory of Medical Information Processing School of Computer Science and Engineering Southeast University Nanjing210096 China Jiangsu Key Laboratory of Molecular and Functional Imaging Department of Radiology Zhongda Hospital Southeast University Nanjing210009 China

Because of the complicated mechanism of ankle injury, it is very difficult to diagnose ankle fracture in clinic. In order to simplify the process of fracture diagnosis, an automatic diagnosis model of ankle fracture was proposed. Firstly, a tibia-fibula segmentation network is proposed for the joint tibiofibular region of the ankle joint, and the corresponding segmentation dataset is established on the basis of fracture data. Secondly, the image registration method is used to register the bone segmentation mask with the normal bone mask. Finally, a semi-supervised classifier is constructed to make full use of a large number of unlabeled data to classify ankle fractures. Experiments show that the proposed method can segment fractures with fracture lines accurately and has better performance than the general method. At the same time, this method is superior to classification network in several indexes. Copyright © 2024, The Authors. All rights reserved.

关键词： image classification

来源：评论

学校读者我要写书评

暂无评论

MMTrans-MT: A Framework for Multimodal Emotion Recognition Using Multitask Learning

MMTrans-MT: A Framework for Multimodal Emotion Recognition U...

引用

International Workshop on Advanced Computational intelligence (IWACI)

作者： Jinrui Shen Jiahao Zheng Xiaoping Wang the School of Artificial Intelligence and Automation Huazhong University of Science and Technology Wuhan China the School of Artificial Intelligence and Automation the Key Laboratory of Image Processing and Intelligent Control of Education Ministry of China Huazhong University of Science and Technology Wuhan China

With the development of deep learning, emotion recognition tasks are more inclined to use multimodal data and adequate supervised information to improve accuracy. In this work, MMTrans-MT (Multimodal Transformer-Multitask), the framework for multimodal emotion recognition using multitask learning is proposed. It has three modules: modalities representation module, multimodal fusion module, and multitask output module. Three modalities, i.e, words, audio and video, are comprehensively utilized to carry out emotion recognition by a simple but efficient fusion model based on Transformer. As for multitask learning, the two tasks are defined as categorical emotion classification and dimensional emotion regression. Considering a potential mapping relationship between two kinds of emotion model, multitask learning is adopted to make the two tasks promote each other and improve recognition accuracy. We conduct experiments on CMU-MOSEI and IEMOCAP datasets. Comprehensive experiments show that the accuracy of recognition using multimodal information is higher than that using unimodal information. Adopting multitask learning promotes the performance of emotion recognition.

关键词： Learning systems Deep learning Emotion recognition Correlation Computational modeling Task analysis Computational intelligence

来源：评论

学校读者我要写书评

暂无评论

Emission Evaluation for Diesel Vehicles Under Typical Operating Conditions

Emission Evaluation for Diesel Vehicles Under Typical Operat...

引用

International Conference on Advanced Robotics and Mechatronics (ICARM)

作者： Jiaren Li Zhenyi Xu Yang Cao Xiushan Xia Yu Kang AHU-IAI AI Joint Laboratory Anhui University Hefei China Key Laboratory of System Control and Information Processing Ministry of Education Shanghai Institute of Artificial Intelligence Hefei Comprehensive National Science Center Hefei China Department of Automation University of Science and Technology of China Hefei China Institute of Advanced Technology University of Science and Technology of China Hefei China

China‘s ambient air quality has improved significantly as a result of the formulation and implementation of the Action Plan for the Prevention and control of Air Pollution and the Three-Year Action Plan for Winning the Blue Sky Defense War, but the air pollution problem in key regions remains severe, diesel truck pollution has not been effectively addressed, and mobile sources are an important source of nitrogen oxide emissions. The existing diesel vehicle emission model is primarily divided based on average vehicle speed and VSP binning methods, and it lacks an instantaneous vehicle speed model. This paper collected data from 10 diesel vehicles in order to accurately evaluate driving conditions and promote energy saving and emission reduction of heavy-duty diesel vehicles. Then we extracted 51,925 kinematic fragment data after cleaning abnormal data and built a kinematic feature evaluation system of diesel vehicles closely related to NOx emission. The sample data of kinematic segments were clustered into two groups of low-speed conditions and medium-speed conditions using the standardized mutual information to extract the characteristics with strong correlation with NOx and the vehicle driving conditions assessment model of cluster analysis. From the clustering results, the high NOx emission is completely classified into the low-speed working condition, which matches the conclusion that the high NOx emission region appears in the low-speed acceleration region reached by previous researchers through experimental studies, indicating that the kinematic feature evaluation model of diesel vehicles established in this paper is reasonable and effective.

关键词：

来源：评论

学校读者我要写书评

暂无评论

Cascaded Aggregation Convolution Network for Saliency Detection Of Stored-Grain Pests

SSRN

引用

SSRN 2023年

作者： Yu, Junwei Zhai, Fupin Liu, Nan Key Laboratory of Grain Information Processing and Control Ministry of Education Zhengzhou450001 China School of Artificial Intelligence and Big Data Henan University of Technology Zhengzhou450001 China College of Information Scinence and Engineering Henan University of Technology Zhengzhou450001 China PLA SSF Information Engineering University Zhengzhou450001 China

The behaviors exhibited by stored grain pests, including feeding, respiration, excretion, and reproduction, engender multifarious threats to grain storage. Consequently, it is imperative to diligently undertake the task of detecting and effectively controlling these pests. At present, Great progress has made in both traditional methods and deep learning techniques for salient object detection. It is still a challenge, especially for small objects and non-salient images, biological vision. This paper presents a grain pests detection method based on cascaded strous convolution . We collet(collect) a novel dataset GrainPest, which comprises 500 images covering various salient pests in grains. We develop a simple and effective framework named cascaded aggregation convolution network(CACNet). CACNet outperforms various state-of-the-art object detection baselines on all datasets tested, making it a robust, general framework that can help facilitate future research in COD. Finally, we conduct a large-scale COD study, evaluating 13 cutting-edge models, providing some interesting findings, and showing several potential applications. Our research offers the community an opportunity to explore more in this new field. © 2023, The Authors. All rights reserved.

关键词： Convolution

来源：评论

学校读者我要写书评

暂无评论

Machine Learning-Assisted to Study the Effect of N:P Ratio on Algae Growth to Forecasting the Bloom with Dielectric Spectroscopy Technique

SSRN

引用

SSRN 2023年

作者： Wang, Shuhang Bingke, Yang Qu, Yihe Zhang, Dayong Chen, Junyi Russel, Mohammad Biswas, Kishore National Engineering Laboratory for Lake Pollution Control and Ecological Restoration Chinese Research Academy of Environmental Beijing100012 China School of Ocean Science and Technology Key laboratory of Industrial Ecology and Environmental Engineering Ministry of Education Dalian University of Technology Panjin124221 China Artificial Intelligence–Natural Language Processing School of Informatics Peking University Beijing100085 China

Real-time online monitoring of algal biomass has emerged as a significant challenge in contemporary marine and freshwater environmental conservation efforts aimed at safeguarding the aquatic ecosystem. Given the pivotal role that algae play in maintaining a delicate balance of nutrients and other environmental factors, the application of a machine learning-assisted dielectric spectroscopy approach represents a novel and intelligent technique for mitigating the current challenges associated with monitoring algal *** work optimized a machine learning -assisted real-time dielectric spectroscopy approach to study the effect of N:P ratios on Desmodesmus sp. growth to forecasting the blooming. As a primary feature, various N:P concentrations (1:1, 2:1, 3:1, 1:2, and 1:3) were utilized to characterize and compare the dielectric (S11(dB) & Ε'') and conventional UV-vis spectroscopy properties (OD686 & Chl-a). The dielectric measurements of S11(dB) & Ε'' are consistent with the OD686 & Chl-a value. Highest Desmodesmus sp. OD686 0.87,0.76 recorded at N:P 3:1,2:1 ratio respectively. The machine learning based PCA and linear regression model were applied and found at 0-8 days incubation time Desmodesmus sp. growth biomass negatively corelated with S11(dB) & Ε" and positively correlatedly with OD686 & Chl-a respectively, where R2 values >0.95. While the measured Ε"& S11(dB) -7.956 &-0.6775, the linear regression model was applied to predicted and forecasting OD686 for 8days and 20days was 0.817, 0.716;2.136,1.926 respectively, at the ratios N:P (3:1) and RMES value were 0.0047 & 0.0319. This analysis revealed that the machine learning- assisted model has high accuracy. Thus, Machine learning-assisted dielectric spectroscopy method could be the potential tools to real-time monitoring the algal cell biomass or forecasting the blooming to control the N:P ratios to protect the aquatic ecosystem. © 2023, The Authors. All rights reserved.

关键词： Forecasting

来源：评论

学校读者我要写书评

暂无评论

Symplectic Wigner Distribution in the Linear Canonical Transform Domain: Theory and Application

arXiv

引用

arXiv 2025年

作者： He, Yangfan Zhang, Zhichao School of Communication and Artificial Intelligence School of Integrated Circuits Nanjing Institute of Technology Nanjing211167 China School of Mathematics and Statistics Nanjing University of Information Science and Technology Nanjing210044 China Key Laboratory of System Control and Information Processing Ministry of Education Shanghai Jiao Tong University Shanghai200240 China Key Laboratory of Computational Science Application of Hainan Province Hainan Normal University Haikou571158 China Key Laboratory of Numerical Simulation of Sichuan Provincial Universities School of Mathematics and Information Sciences Neijiang Normal University Neijiang641000 China

This paper devotes to combine the chirp basis function transformation and symplectic coordinates transformation to yield a novel Wigner distribution (WD) associated with the linear canonical transform (LCT), named as the symplectic WD in the LCT domain (SWDL). It incorporates the merits of the symplectic WD (SWD) and the WD in the LCT domain (WDL), achieving stronger capability in the linear frequency-modulated (LFM) signal frequency rate feature extraction while maintaining the same level of computational complexity. Some essential properties of the SWDL are derived, including marginal distributions, energy conservations, unique reconstruction, Moyal formula, complex conjugate symmetry, time reversal symmetry, scaling property, time translation property, frequency modulation property, and time translation and frequency modulation property. Heisenberg’s uncertainty principles of the SWDL are formulated, giving rise to three kinds of lower bounds attainable respectively by Gaussian enveloped complex exponential signal, Gaussian signal and Gaussian enveloped chirp signal. The optimal symplectic matrices corresponding to the highest time-frequency resolution are generated by solving the lower bound optimization (minimization) problem. The time-frequency resolution of the SWDL is compared with those of the SWD and WDL to demonstrate its superiority in LFM signals time-frequency energy concentration. A synthesis example is also carried out to verify the feasibility and reliability of the theoretical analysis. Copyright © 2025, The Authors. All rights reserved.

关键词： Wigner-Ville distribution

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：