检索结果-内蒙古大学图书馆

第34届中国控制与决策会议

作者： Xin Li Xiaoli Li Kang Wang Faculty of Information Technology Beijing University of Technology Beijing Key Laboratory of Computational Intelligence and Intelligent System Engineering Research Center of Digital Community Ministry of Education

Decomposition-based multi-objective optimization aliorithms have been shown to be successful in solving multi-objective optimization problems(MOPs).Since the shape is unknown before the search,the classical decomposition method relies on a set of uniformly distributed reference vectors to divide the objective space into multiple ***,the the distribution of reference vectors produced by the Das and Dennis systematic approach is difficult to maintain the diversity of(extremely) concave and(extremely) convex *** remedy this issue,a multi-objective evolutionary algorithm based on cone heart point adjustment vectors is proposed to solve concave and convex problems(CHP-MOEA).In this paper,the concavity or convexity is first according to the intermediate objective ***,the distribution of reference vectors are adjusted according to the cone heart point on the basis of the traditional ***,experimental results on 31 benchmark problems show that the proposed algorithm exhibits better performance than other state-of-the-art algorithms.

关键词： Cone heart point Concavity or convexity Multi-objective evolutionary algorithm

来源：评论

学校读者我要写书评

暂无评论

TRAIL: Trust-Aware Client Scheduling for Semi-Decentralized Federated Learning 39

TRAIL: Trust-Aware Client Scheduling for Semi-Decentralized ...

引用

39th Annual AAAI Conference on Artificial intelligence, AAAI 2025

作者： Hu, Gangqiang Lu, Jianfeng Han, Jianmin Cao, Shuqin Liu, Jing Fu, Hao School of Computer Science and Technology Zhejiang Normal University China School of Computer Science and Technology Wuhan University of Science and Technology China Key Laboratory of Social Computing and Cognitive Intelligence Dalian University of Technology Ministry of Education China Hubei Province Key Laboratory of Intelligent Information Processing and Real-time Industrial System Wuhan University of Science and Technology China

ISBN: (纸本)157735897X

Due to the sensitivity of data, Federated Learning (FL) is employed to enable distributed machine learning while safeguarding data privacy and accommodating the requirements of various devices. However, in the context of semi-decentralized FL, clients’ communication and training states are dynamic. This variability arises from local training fluctuations, heterogeneous data distributions, and intermittent client participation. Most existing studies primarily focus on stable client states, neglecting the dynamic challenges inherent in real-world scenarios. To tackle this issue, we propose a TRust-Aware clIent scheduLing mechanism called TRAIL, which assesses client states and contributions, enhancing model training efficiency through selective client participation. We focus on a semi-decentralized FL framework where edge servers and clients train a shared global model using unreliable intra-cluster model aggregation and inter-cluster model consensus. First, we propose an adaptive hidden semi-Markov model to estimate clients’ communication states and contributions. Next, we address a client-server association optimization problem to minimize global training loss. Using convergence analysis, we propose a greedy client scheduling algorithm. Finally, our experiments conducted on real-world datasets demonstrate that TRAIL outperforms state-of-the-art baselines, achieving an improvement of 8.7% in test accuracy and a reduction of 15.3% in training loss. Copyright © 2025, Association for the Advancement of Artificial intelligence (***). All rights reserved.

关键词： Federated learning

来源：评论

学校读者我要写书评

暂无评论

Data-driven Optimal Control for Acid Production Process with Flue Gas

Data-driven Optimal Control for Acid Production Process with...

引用

第33届中国过程控制会议

作者： Minghua Liu Xiaoli Li Kang Wang Meiqi Wang Faculty of Information Technology Beijing University of Technology Beijing Key Laboratory of Computational Intelligence and Intelligent System Engineering Research Center of Digital Community Ministry of Education

The optimal control of acid production process with flue gas is an effective method to reduce energy consumption and ***,how to improve the SO conversion rate is still a challenge faced by the current acid production process with flue *** solve this problem,a data-driven optimization control method is ***,BP neural network is used to establish the expression relationship between SO conversion rate and inlet temperature of converter,and the optimized setting model of operation process is ***,particle swarm optimization algorithm is used to solve the optimization setting model,and the results are applied to the industrial ***,taking the first layer inlet temperature of converter as an example,PID controller is used to track and control *** results show that the method can significantly improve the conversion rate of SO in the converter and meet the national *** the same time,the PID controller has better control performance to the first layer inlet temperature of converter and strong anti-interference ability.

关键词： Data-driven method acid production converter inlet temperature particle swarm optimization optimal control

来源：评论

学校读者我要写书评

暂无评论

Research on Attention-Based Predictive Model in Multi-Agent Reinforcement Learning

Research on Attention-Based Predictive Model in Multi-Agent ...

引用

2022 Chinese Automation Congress, CAC 2022

作者： Hong, Weixin Wu, Huaiyu Shen, Zhen Xiong, Gang Hu, Bin Dong, Xisong Wang, Fei-Yue The State Key Laboratory for Management and Control of Complex Systems Institute of Automation Chinese Academy of Sciences Beijing100190 China School of Artificial Intelligence University of Chinese Academy of Sciences Beijing100049 China The Beijing Engineering Research Center of Intelligent Systems and Technology Institute of Automation Chinese Academy of Sciences Beijing100190 China The Guangdong Engineering Research Center of 3D Printing and Intelligent Manufacturing Cloud Computing Center Chinese Academy of Sciences Dongguan523808 China

ISBN: (纸本)9781665465335

Agent actions planning is a challenging problem in multi-agent reinforcement learning. Recent methods typically build their predictive models by full connection layers, but the shortage of utilizing action and observation information in complex multi-agent environment limits the performance of those methods. To address this issue, we propose a novel predictive model based on self-attention mechanism which can effectively use information in the environment. Specifically, we use the MADDPG algorithm to make a dataset where the input is the observations and actions and the output is the predicted observations and rewards. We also introduce the multi-head attention mechanism into the predictive model so that the predictive model can better combine different information for prediction. We perform extensive experiments to assess the proposed method and compare it with the representative baseline approach. Results demonstrate that our method outperforms competitors with higher accuracy in different tasks. Our implementation will be publicly available. © 2022 IEEE.

关键词： Reinforcement learning

来源：评论

学校读者我要写书评

暂无评论

FedCross: Intertemporal Federated Learning Under Evolutionary Games 39

FedCross: Intertemporal Federated Learning Under Evolutionar...

引用

39th Annual AAAI Conference on Artificial intelligence, AAAI 2025

作者： Lu, Jianfeng Zhang, Ying Jia, Riheng Cao, Shuqin Liu, Jing Fu, Hao School of Computer Science and Technology Wuhan University of Science and Technology China Key Laboratory of Social Computing and Cognitive Intelligence (Dalian University of Technology) Ministry of Education China School of Computer Science and Technology Zhejiang Normal University China Hubei Province Key Laboratory of Intelligent Information Processing and Real-time Industrial System Wuhan University of Science and Technology China

ISBN: (纸本)157735897X

Federated Learning (FL) mitigates privacy leakage in decentralized machine learning by allowing multiple clients to train collaboratively locally. However, dynamic mobile networks with high mobility, intermittent connectivity, and bandwidth limitation severely hinder model updates to the cloud server. Although previous studies have typically addressed user mobility issue through task reassignment or predictive modeling, frequent migrations may result in high communication overhead. Addressing this challenge involves not only dealing with resource constraints, but also finding ways to mitigate the challenges posed by user migrations. We therefore propose a intertemporal incentive framework, FedCross, which ensures the continuity of FL tasks by migrating interrupted training tasks to feasible mobile devices. FedCross comprises two distinct stages: Specifically, in Stage 1, we address the task allocation problem across regions under resource constraints by employing a multi-objective migration algorithm to quantify the optimal task receivers. Moreover, we adopt evolutionary game theory to capture the dynamic decision-making of users, forecasting the evolution of user proportions across different regions to mitigate frequent migrations. In Stage 2, we utilize a procurement auction mechanism to allocate rewards among base stations, ensuring that those providing high-quality models receive optimal compensation. This approach incentivizes sustained user participation, thereby ensuring the overall feasibility of FedCross. Finally, experimental results validate the theoretical soundness of FedCross and demonstrate its significant reduction in communication overhead. Copyright © 2025, Association for the Advancement of Artificial intelligence (***). All rights reserved.

关键词： Federated learning

来源：评论

学校读者我要写书评

暂无评论

Traj-LLM: A New Exploration for Empowering Trajectory Prediction with Pre-trained Large Language Models

arXiv

引用

arXiv 2024年

作者： Lan, Zhengxing Li, Hongbo Liu, Lingshan Fan, Bo Lv, Yisheng Ren, Yilong Cui, Zhiyong The School of Transportation Science and Engineering Beihang University Beijing100191 China The State Key Lab of Intelligent Transportation System Beijing100191 China The Research Institute for Frontier Science Beihang University Beijing100191 China The Beijing Key Laboratory of Traffic Engineering College of Metropolitan Transportation Beijing University of Technology Beijing100124 China The State Key Laboratory of Multimodal Artificial Intelligence Systems Institute of Automation Chinese Academy of Sciences Beijing100190 China The School of Artificial Intelligence University of Chinese Academy of Sciences Beijing100049 China The Zhongguancun Laboratory Beijing100194 China The Beihang Hangzhou Innovation Institute Yuhang Hangzhou310023 China

Predicting the future trajectories of dynamic traffic actors is a cornerstone task in autonomous driving. Though existing notable efforts have resulted in impressive performance improvements, a gap persists in scene cognitive and understanding of the complex traffic semantics. This paper proposes Traj-LLM, the first to investigate the potential of using Large Language Models (LLMs) without explicit prompt engineering to generate future motion from agents’ past/observed trajectories and scene semantics. Traj-LLM starts with sparse context joint coding to dissect the agent and scene features into a form that LLMs understand. On this basis, we innovatively explore LLMs’ powerful comprehension abilities to capture a spectrum of high-level scene knowledge and interactive information. Emulating the human-like lane focus cognitive function and enhancing Traj-LLM’s scene comprehension, we introduce lane-aware probabilistic learning powered by the pioneering Mamba module. Finally, a multi-modal Laplace decoder is designed to achieve scene-compliant multi-modal predictions. Extensive experiments manifest that Traj-LLM, fortified by LLMs’ strong prior knowledge and understanding prowess, together with lane-aware probability learning, outstrips state-of-the-art methods across evaluation metrics. Moreover, the few-shot analysis further substantiates Traj-LLM’s performance, wherein with just 50% of the dataset, it outperforms the majority of benchmarks relying on complete data utilization. This study explores equipping the trajectory prediction task with advanced capabilities inherent in LLMs, furnishing a more universal and adaptable solution for forecasting agent motion in a new way. Copyright © 2024, The Authors. All rights reserved.

关键词： Semantics

来源：评论

学校读者我要写书评

暂无评论

RobustPFL: Robust Personalized Federated Learning

引用

IEEE Transactions on Dependable and Secure computing 2025年

作者： Chen, Guorong Wang, Wei Wu, Yufang Li, Chao Xu, Guangquan Ji, Shouling Li, Tao Shen, Meng Han, Yufei Xi'an Jiaotong University Ministry of Education Key Lab for Intelligent Networks and Network Security Xi'an China Beijing Jiaotong University 3 Shangyuancun Beijing Key Laboratory of Security and Privacy in Intelligent Transportation Beijing100044 China Tianjin University Tianjin Key Laboratory of Advanced Networking College of Intelligence and Computing Tianjin300350 China Zhejiang University College of Computer Science and Technology Hangzhou310027 China ITAI Haihe Lab Tianjin China Nankai University College of Computer Science Tianjin300350 China Beijing Institute of Technology School of Cyberspace Security Beijing100081 China Peng Cheng Laboratory Cyberspace Security Research Center Shenzhen518066 China INRIA Bretagne Rennes35042 France

Conventional federated learning (FL) coordinated by a central server focuses on training a global model and protecting the privacy of clients' training data by storing it locally. However, the statistical heterogeneity hinders the global model from adapting to the non-IID distributions among clients. Moreover, untrusted and unreliable central servers and malicious clients may compromise model integrity and availability, thus degrading the robustness of FL. To address these challenges, we present RobustPFL, a decentralized personalized federated learning (PFL) approach that combines α-based Layer-position Normalized Similarity (α-LNS) and local collaborative training to improve personalized performance while utilizing a blockchain-based committee mechanism to coordinate the aggregation process, thereby achieving high personalized accuracy and robustness. Extensive experiments show that our RobustPFL approach outperforms multiple algorithms, including Local training, FedAvg, FedReptile, Per-FedAvg, FedBN, and SPFL, on MNIST, CIFAR10, EMNIST, and N-BaIoT datasets in four non-IID settings. We also evaluate RobustPFL's effectiveness against attacks - poisoning attacks and free-riding attacks. Particularly, for three prevalent poisoning attacks (backdoor, label flipping, and model poisoning attacks), we compare non-defensive (FedAvg) and defensive (Krum, trimmed mean, Bulyan, FedBN, FLAME, and FangTrmean) methods with our proposed RobustPFL. The results show that our approach achieves significant defensive effects. © 2004-2012 IEEE.

关键词： Federated learning

来源：评论

学校读者我要写书评

暂无评论

intelligent prediction of wastewater treatment blower bearing temperature based on random event agent model

引用

Scientia Sinica Technologica 2025年第5期55卷 879-892页

作者： Wu, XiaoLong Jiang, WeiYu Han, HongGui Li, Tong Zhao, DeZheng Faculty of Information Technology Beijing Key Laboratory of Computational Intelligence and Intelligent System Beijing University of Technology Beijing100124 China Beijing Drainage Group Co Ltd. Beijing100044 China CLP Intelligent Technology Co Ltd. Beijing102200 China

The bearing temperature of a wastewater treatment blower is a critical parameter for assessing operational status. It provides essential insights into the bearing’s working condition and lubricate state. Excessive temperatures can lead to bearing damage, lubrication failure, or even blower shutdown. To achieve an accurate prediction of blower bearing temperature, this study proposes an intelligent prediction method based on a random event agent model. First, a primary model using a fuzzy neural network is established. This model is dynamically adjusted using data from outlet air pressure, high-speed shaft vibration, and inlet air temperature, allowing precise temperature predictions under normal conditions without the influence of random events. Second, a random event agent model (REAM) is designed, incorporating fuzzy rules to describe the impact of different random events on blower bearing temperature. By combining the REAM with the primary model, accurate temperature predictions are achieved even during random events. The proposed model was applied to predict the blower bearing temperatures of wastewater treatment blowers. Its performance, before and after random event triggers, was evaluated through experiments. Results demonstrate that this method enhances the accuracy of blower bearing temperature predictions under random events. © 2025 Science Press. All rights reserved.

关键词： Fuzzy neural networks

来源：评论

学校读者我要写书评

暂无评论

Labyrinth Morphological Modeling and its Application on Unreferenced Segmentation Assessment

SSRN

引用

SSRN 2022年

作者： Li, Xiaoguang Zhu, Ziyao Yin, Hongxia Zhao, Pengfei Lv, Han Tang, Ruowei Qin, Yating Zhuo, Li Wang, Zhenchang Beijing Key Laboratory of Computational Intelligence and Intelligent System Beijing University of Technology Beijing100124 China Department of Radiology Beijing Friendship Hospital Capital Medical University Beijing100050 China

The labyrinth of the inner ear is an important auditory and balance sensory organ, and is closely related to tinnitus, deafness, vertigo and Ménière diseases. Deep learning effectively improves the performance of the labyrinth segmentation. However, due to the limited number of labeled samples, it is difficult to evaluated the quality of segmentation with referenced metrics on large-scale unlabeled testing set. In addition, the complex morphology of the labyrinth makes it difficult to quantitatively described for further analysis. In this paper, a variational autoencoder (VAE) based morphological modeling of labyrinth is proposed. Firstly, a two-stage pose alignment strategy is introduced to align the segmented inner ear labyrinths. Then, an energy-adaptive spatial and inter-slice dimensionality reduction strategy is adopted to extract compact morphological features via VAE. Finally, a statistical model of the compact feature in the latent space is established to represent the morphology distribution of the labyrinths. Based on feature similarity of latent vectors, an reference-free quality evaluation for segmentation of the labyrinth is explored. The experimental results show that the consistency between the proposed method and the Dice similarity coefficient (DSC) reaches 0.78. In additional, further analysis showed that the model also has high potential to apply to morphological analysis, such as anomaly detection, of the labyrinths. © 2022, The Authors. All rights reserved.

关键词： Anomaly detection

来源：评论

学校读者我要写书评

暂无评论

Soft-Sensing Method of Effluent Ammonia Nitrogen Based on Emotional Neural Network

Soft-Sensing Method of Effluent Ammonia Nitrogen Based on Em...

引用

Chinese Automation Congress (CAC)

作者： Hongye Zhang Cuili Yang Junfei Qiao Beijing Key Laboratory of Computational Intelligence and Intelligent System Beijing University of Technology Beijing China

ISBN: (纸本)9781665426480

In the process of wastewater treatment process (WWTP), the concentration of effluent ammonia nitrogen is one of the important indexes to test whether the water quality reaches the standard. Soft sensing technology is a method which can use some easily measurable auxiliary variables to predict key variables online. In this paper, we designed a soft sensing model of effluent NH 4 -N based on radial basis emotional neural network (RBENN), and then a feature selection (FS) method of sticky binary particle swarm optimization (SBPSO) algorithm based on mutual information (MI) and symmetric uncertainty (SU) is proposed to obtain the appropriate feature variables while balancing correlation and redundancy, and then the network parameters are optimized by comprehensive learning particle swarm optimization (CLPSO) algorithm. Simulation results show that the proposed method effectively improves the prediction performance of the model.

关键词： Training Soft sensors Neural networks Effluents Predictive models Feature extraction Prediction algorithms

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：