检索结果-内蒙古大学图书馆

On-Line Conference (ONCON), Industrial Electronics Society Annual

作者： Yingbo Luo Meibao Yao Xueming Xiao Bo Zheng Intelligent Robotics Lab (IRL) School of Artificial Intelligence Jilin University Changchun China Engineering Research Center of Knowledge-Driven Human-Machine Intelligence Ministry of Education China CVIR Lab Changchun University of Science and Technology Changchun China Shanghai Aerospace Control Technology Institute Shanghai China

Mobile robots are widely used in industrial manufacturing, agricultural production, health care, home services, transportation, and other fields. As the most basic and important research topic in the field of mobile robots, path planning technology has always been the focus of attention. To solve the problems caused by the traditional A* algorithm that the path length is too long and the trajectory is not smooth, This paper proposes an improved A* algorithm based on Ant Colony Optimization. Firstly, use the A* algorithm to roughly find a constrained shortest path. Secondly, extract the key nodes of the shortest path and calculate the distribution of obstacles between these key nodes. Finally, use the improved ant colony algorithm to optimize the searched path to generate the final path. Simulation experiments show that the time complexity and space complexity of the improved A* algorithm are reasonable. The generated path length is shorter, the total turning angle is smaller, the trajectory is smoother, and it is more suitable for the real environment. Compared with the traditional A* algorithm, it has obvious advantages.

关键词：

来源：评论

学校读者我要写书评

暂无评论

A Diversity-Enhanced knowledge Distillation Model for Practical Math Word Problem Solving

arXiv

引用

arXiv 2025年

作者： Zhang, Yi Zhou, Guangyou Xie, Zhiwen Ma, Jinjin Huang, Jimmy Xiangji Faculty of Artificial Intelligence in Education Central China Normal University Hubei Wuhan430079 China School of Information and Safety Engineering Zhongnan University of Economics and Law Hubei Wuhan430073 China Hubei Provincial Key Laboratory of Artificial Intelligence and Smart Learning School of Computer Central China Normal University Hubei Wuhan430079 China Information Retrieval and Knowledge Management Research Lab York University Toronto Canada

Math Word Problem (MWP) solving is a critical task in natural language processing, has garnered significant research interest in recent years. Various recent studies heavily rely on Seq2Seq models and their extensions (e.g., Seq2Tree and Graph2Tree) to generate mathematical equations. While effective, these models struggle to generate diverse but counterpart solution equations, limiting their generalization across various math problem scenarios. In this paper, we introduce a novel Diversity-enhanced knowledge Distillation (DivKD) model for practical MWP solving. Our approach proposes an adaptive diversity distillation method, in which a student model learns diverse equations by selectively transferring high-quality knowledge from a teacher model. Additionally, we design a diversity prior-enhanced student model to better capture the diversity distribution of equations by incorporating a conditional variational auto-encoder. Extensive experiments on four MWP benchmark datasets demonstrate that our approach achieves higher answer accuracy than strong baselines while maintaining high efficiency for practical applications. Copyright © 2025, The Authors. All rights reserved.

关键词： Students

来源：评论

学校读者我要写书评

暂无评论

DC-MOT: Motion Deblurring and Compensation for Multi-Object Tracking in UAV Videos

DC-MOT: Motion Deblurring and Compensation for Multi-Object ...

引用

IEEE International Conference on Robotics and Automation (ICRA)

作者： Song Cheng Meibao Yao Xueming Xiao Intelligent Robotics Lab (IRL) School of Artificial Intelligence Jilin University Changchun China Engineering Research Center of Knowledge-Driven Human-Machine Intelligence Ministry of Education China CVIR lab Changchun University of Science and Technology China Key Lab of Optoelectronic Measurement and Optical Information Transmission Technology Ministry of Education China

In this paper, we propose a multi-object tracking framework for videos captured by UAVs, considering motion imperfection in the following two aspects: 1) motion blurring of objects due to high-speed motion of the UAV and the objects, deteriorating the performance of the detector; 2) motion coupling of the global movement of the UAV camera with the object motion, resulting in the nonlinearity of objects trajectories in adjacent frames and further more difficult to predict. For motion blurring, this paper proposes a hybrid deblurring module that deals with the blurred frames while retaining the clear frames, trading off between video tracking performance and spatio-temporal consistency. For motion coupling, we proposed a motion compensation module to align adjacent frames by feature matching, and the corrected target position is obtained in the next frame to alleviate the interference of camera movement with tracking. We evaluate the proposed methods on VisDrone dataset and validate that our framework achieves new state-of-the-art performance on UAV-based MOT systems.

关键词：

来源：评论

学校读者我要写书评

暂无评论

IKGN: Intention-aware knowledge Graph Network for POI Recommendation

IKGN: Intention-aware Knowledge Graph Network for POI Recomm...

引用

IEEE International Conference on Data Mining (ICDM)

作者： Xiaoyu Zhu Chenyang Bu Bingbing Dong Shengwei Ji Yi He Xindong Wu Key Laboratory of Knowledge Engineering with Big Data (the Ministry of Education of China) Hefei University of Technology School of Computer Science and Information Engineering Hefei University of Technology China School of Artificial Intelligence and Big Data Hefei University China Department of Computer Science Old Dominion University USA Research Center for Knowledge Engineering Zhejiang Lab China

Point-of-Interest (POI) recommendation, pivotal for guiding users to their next interested locale, grapples with the persistent challenge of data sparsity. Whereas knowledge graphs (KGs) have emerged as a favored tool to mitigate the issue, existing KG-based methods tend to overlook two crucial elements: the intention steering users’ location choices and the high-order topological structure within the KG. In this paper, we craft an Intention-aware knowledge Graph (IKG) that harmonizes users’ visit histories, movement trajectories, and location categories to model user intentions. Building upon IKG, our novel Intention-aware knowledge Graph Network (IKGN) delves deeper into the POI recommendation by weighing and propagating node embeddings through an attention mechanism, capturing the unique locational intent of each user. A sequential model like GRU is then employed to ensure a comprehensive representation of users’ short- and long-term location preferences. An empirical study on two real-world datasets validates the effectiveness of our proposed IKGN, with it markedly outshining seven benchmark rival models in both Recall and NDCG metrics. The code of IKGN is available at https://***/Jungle123456/IKGN.

关键词：

来源：评论

学校读者我要写书评

暂无评论

A Data Enhancement Strategy for Multi-Agent Cooperative Hunting based on Deep Reinforcement Learning

A Data Enhancement Strategy for Multi-Agent Cooperative Hunt...

引用

Industrial Cyber-Physical Systems (ICPS)

作者： Zhenkun Gao Xiaoyan Dai Meibao Yao Xueming Xiao Intelligent Robotics Lab (IRL) School of Artificial Intelligence Jilin University Changchun China Engineering Research Center of Knowledge-Driven Human-Machine Intelligence Ministry of Education China CVIR lab Changchun University of Science and Technology China Key Lab of Opto-electronic Measurement and Optical Information Transmission Technology Ministry of Education China

Cooperative hunting is a typical and significant scene to study multi-agent behaviors, where conventional control strategies are difficult to cope with, due to its high dimensionality of state space and locality of communication. Reinforcement learning provides a framework and a set of tools for this issue by trial-and-error interactions with the environment. Though promising, it often requires a large number of empirical sample data to learn effective hunting strategies, leading to low sample efficiency, understood as the training episodes required for the agent to learn effective behavior strategies. To improve the sampling efficiency, we propose a data enhancement strategy integrated in the execution (CTDE) training framework to train the multi-agent system. The data enhancement strategy is based on a state transfer dynamics model to generate additional predicted data, which we called dynamic prediction model, combined with the empirical data by interacting with the environment, for higher sample efficiency. The simulation results on the Webots platform show that our method outperforms some state-of-the-art methods, such as MAPPO, with high data sample efficiency.

关键词： Training Service robots Simulation Training data Reinforcement learning Predictive models Data models

来源：评论

学校读者我要写书评

暂无评论

Co-Optimization Method for Modular Robots Based on Evolutionary Algorithm

Co-Optimization Method for Modular Robots Based on Evolution...

引用

On-Line Conference (ONCON), Industrial Electronics Society Annual

作者： Peng Zhao Meibao Yao Bo Zheng Xueming Xiao Intelligent Robotics Lab (IRL) School of Artificial Intelligence Jilin University Changchun China Engineering Research Center of Knowledge-Driven Human-Machine Intelligence Ministry of Education China Shanghai Aerospace Control Technology Institute Shanghai China CVIR Lab Changchun University of Science and Technology Changchun China Key Lab of Optoelectronic Measurement and Optical Information Transmission Technology

Deploying robots instead of human labor in complex and hazardous environments is a highly effective strategy. However, current fixed-configuration robot platforms struggle to meet the demanding exploration requirements in non-structured environments. In contrast, reconfigurable modular robots are able to adjust their configurations in response to environmental and task requirements, thereby offering immense potential for tackling complex exploration tasks in unstructured environments. This versatility of reconfigurable modular robots allows them to better adapt to various environmental and task demands, such as using snake or caterpillar configurations to navigate narrow holes, or quadruped configurations to traverse rugged terrain. However, the current research on modular robots primarily focuses on optimizing either their own configurations or solely the motion control model, potentially limiting their adaptability to the environment and hindering the full realization of their reconfigurability. To address this limitation, this paper improves upon traditional EA and proposes a configuration-motion co-optimization method and a new encoding method for modular robots, which jointly models and optimizes both configuration and motion. The method includes dynamic encoding of configuration-motion and the design of compatible operators, effectively enhancing the adaptive capabilities of modular robots in complex environments. The method was validated using the Webots platform, and the results demonstrate successful generation of adaptable reconfiguration schemes for various terrains. Compared to traditional EA, this method has shown superior performance in handling diverse tasks.

关键词：

来源：评论

学校读者我要写书评

暂无评论

A Mars Multi-terrain Simulator Using a Modular Terrain Construction Framework

A Mars Multi-terrain Simulator Using a Modular Terrain Const...

引用

On-Line Conference (ONCON), Industrial Electronics Society Annual

作者： Yun Jiang Meibao Yao Bo Zheng Tao Hu Tao Cao Xueming Xiao Intelligent Robotics Lab (IRL) School of Artificial Intelligence Jilin University Changchun China Engineering Research Center of Knowledge-Driven Human-Machine Intelligence Ministry of Education China Shanghai Aerospace Control Technology Institute Shanghai China CVIR Lab Changchun University of Science and Technology Changchun China Key Lab of Optoelectronic Measurement and Optical Information Transmission Technology

Mars exploration tasks are inseparable from the support of robotic equipment. More comprehensive environmental awareness capabilities can improve the efficiency of robots on the surface of Mars. artificial neural networks have emerged as pivotal tools in enhancing the perceptual capabilities of robots. The cornerstone of effective neural network training lies in the availability of high-quality data. Regrettably, the unstructured terrains, particularly the Martian landscape, lack real-world datasets with high-quality annotations. To overcome this challenge, using simulators to generate large amounts of high-quality data for researchers is a good option. Nonetheless, the creation of these simulators requires complex details such as ground texture characteristics and rock distribution, which increases research costs. In response, we introduce a novel approach – the modular terrain construction framework and construct a Mars multi-terrain simulator based on it. This innovative simulator not only simplifies the simulation process but also grants users the flexibility to alter it to their specific research needs. One of the standout features of our simulator is its capacity to generate rich metadata, such as semantic and depth, which can be used seamlessly for a multitude of downstream applications. To validate the reliability of the data produced by our simulator, we have assessed its performance on a semantic segmentation task.

关键词：

来源：评论

学校读者我要写书评

暂无评论

LPL-SLAM: LiDAR SLAM with Plane and Line Optimization for Structural Environment

LPL-SLAM: LiDAR SLAM with Plane and Line Optimization for St...

引用

On-Line Conference (ONCON), Industrial Electronics Society Annual

作者： Huanfeng Zhao Meibao Yao Xueming Xiao Bo Zheng Tao Cao Tao Hu Intelligent Robotics Lab (IRL) School of Artificial Intelligence Jilin University Changchun China Engineering Research Center of Knowledge-Driven Human-Machine Intelligence Ministry of Education China CVIR Lab Changchun University of Science and Technology Changchun China Key Lab of Opto-Electronic Measurement and Optical Information Transmission Technology Shanghai Aerospace Control Technology Institute Shanghai China

In LiDAR SLAM, data association of sparse LiDAR scan and high computational complexity of point-to-model metrics limit real-time applications of joint optimization module like Bundle Adjustment in Visual SLAM, which leads to cumulative errors deteriorating performance. To address this drawback, in this paper, we propose LPL-SLAM with plane and line optimization for the structural environment. Since plane and line primitives are ubiquitous in man-made environments, we treat the primitives as landmarks to process scans. We introduce line and plane factors relying on line-to-line and plane-to-plane metrics to optimize keyframe poses, planes and lines. With the line and plane factors, our system avoids a large-scale optimization problem and yields accurate and consistent tracking poses. Experimental results demonstrate that LPL-SLAM outperforms the state-of-the-art algorithms and achieves real-time performance.

关键词：

来源：评论

学校读者我要写书评

暂无评论

M3TTS: Multi-Modal Text-to-Speech of Multi-Scale Style Control for Dubbing

SSRN

引用

SSRN 2023年

作者： Liu, Yan Wei, Li-Fang Qian, Xinyuan Zhang, Tian-Hao Chen, Song-Lu Yin, Xu-Cheng School of Computer and Communication Engineering University of Science and Technology Beijing Beijing100083 China Key Laboratory of Knowledge Automation for Industrial Processes of Ministry of Education Beijing100083 China USTB-EEasyTech Joint Lab of Artificial Intelligence University of Science and Technology Beijing Beijing100083 China University of Science and Technology Beijing China

Dubbing refers to the procedure of recording characters by professional voice actors in films and games. It is more expressive and immersive than conventional text-to-speech (TTS) technologies and requires synchronization and style consistency of audio and video. Previous dubbing methods use video to provide either a global style vector or a local prosody embedding, limiting the expressiveness of the predicted waveform. To generate more expressive audio with precise visual temporal alignment, we propose a multi-modal multi-scale expressive speech synthesis method, namely multi-modal multi-scale TTS (M3TTS), which introduces an auxiliary video input to provide style embedding. Specifically, M3TTS adopts a memory network to bridge heterogeneous modalities and further solves the training-inference style mismatch in conventional multi-modal TTS. To enhance the expressiveness of synthesized audio, a multi-scale style modeling scheme is used for recovering style characteristics at different scales. In addition, M3TTS can convert the style of speech by choosing different reference videos. We conduct extensive experiments on the public GRID corpus, where our proposed M3TTS can generate high-quality video-aligned speech. It also shows superior performance over the other comparable methods, both subjectively and objectively. © 2023, The Authors. All rights reserved.

关键词： Embeddings

来源：评论

学校读者我要写书评

暂无评论

Repetitive nonoverlapping sequential pattern mining

arXiv

引用

arXiv 2023年

作者： Geng, Meng Wu, Youxi Li, Yan Liu, Jing Fournier-Viger, Philippe Zhu, Xingquan Wu, Xindong School of Artificial Intelligence Hebei University of Technology Tianjin300401 China School of Economics and Management Hebei University of Technology Tianjin300401 China State Key Laboratory of Reliability and Intelligence of Electrical Equipment Hebei University of Technology Tianjin300401 China Shenzhen University Shenzhen China Department of Computer & Electrical Engineering and Computer Science Florida Atlantic University FL33431 United States Research Center for Knowledge Engineering The Zhejiang Lab Hangzhou311121 China

Sequential pattern mining (SPM) is an important branch of knowledge discovery that aims to mine frequent sub-sequences (patterns) in a sequential database. Various SPM methods have been investigated, and most of them are classical SPM methods, since these methods only consider whether or not a given pattern occurs within a sequence. Classical SPM can only find the common features of sequences, but it ignores the number of occurrences of the pattern in each sequence, i.e., the degree of interest of specific users. To solve this problem, this paper addresses the issue of repetitive nonoverlapping sequential pattern (RNP) mining and proposes the RNP-Miner algorithm. To reduce the number of candidate patterns, RNP-Miner adopts an itemset pattern join strategy. To improve the efficiency of support calculation, RNP-Miner utilizes the candidate support calculation algorithm based on the position dictionary. To validate the performance of RNP-Miner, 10 competitive algorithms and 20 sequence databases were selected. The experimental results verify that RNP-Miner outperforms the other algorithms, and using RNPs can achieve a better clustering performance than raw data and classical frequent patterns. All the algorithms were developed using the PyCharm environment and can be downloaded from https://***/wuc567/Pattern-Mining/tree/master/RNPMiner. Copyright © 2023, The Authors. All rights reserved.

关键词： Clustering 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：