检索结果-内蒙古大学图书馆

arXiv 2025年

作者： Li, Mingrui Liu, Shuhong Deng, Tianchen Wang, Hongyu the Department of Computer Science Dalian University of Technology Dalian China the Department of Mechano-informatics Information Science and Technology The University of Tokyo Tokyo Japan Institute of Medical Robotics Department of Automation Shanghai Jiao Tong University Key Laboratory of System Control and Information Processing Ministry of Education China

Gaussian SLAM systems excel in real-time rendering and fine-grained reconstruction compared to NeRF-based systems. However, their reliance on extensive keyframes is impractical for deployment in real-world robotic systems, which typically operate under sparse-view conditions that can result in substantial holes in the map. To address these challenges, we introduce DenseSplat, the first SLAM system that effectively combines the advantages of NeRF and 3DGS. DenseSplat utilizes sparse keyframes and NeRF priors for initializing primitives that densely populate maps and seamlessly fill gaps. It also implements geometry-aware primitive sampling and pruning strategies to manage granularity and enhance rendering efficiency. Moreover, DenseSplat integrates loop closure and bundle adjustment, significantly enhancing frame-to-frame tracking accuracy. Extensive experiments on multiple large-scale datasets demonstrate that DenseSplat achieves superior performance in tracking and mapping compared to current state-of-the-art methods. © 2025, CC BY.

关键词： Real time systems

来源：评论

学校读者我要写书评

暂无评论

Ant Colony Optimization for Service Composition in Multi-cloud Environments

Ant Colony Optimization for Service Composition in Multi-clo...

引用

International Conference on Computing and Networking technology (ICCNT)

作者： Mahesh Maurya G Vidhya Faiz Akram M Saravanan K Venkata Ramana S Rajakumari Department of Computer Engineering K.C. College of Engineering Mith Bunder road Kopri Thane East Maharashtra India Department of Computer Science and Engineering Sona College of Technology Salem Tamilnadu India Department of Computer Science Faculty of Computing and Informatics Jimma Institute of Technology Jimma University Jimma Oromia Region Ethiopia Department of ECE Sri Eshwar College of Engineering Coimbatore Tamilnadu India Department of CSE QIS College of Engineering and Technology Ongole Andrapradesh India Department of Mathematics R.M.D. Engineering College

ISBN: (数字)9798350370249

ISBN: (纸本)9798350370270

the Ant Colony Optimization (ACO) algorithm is a famous metaheuristic technique that has been successfully applied in various optimization issues. It mimics the foraging conduct of ants to locate the shortest path among meals assets and their colony. Recently, ACO has been increasingly more applied for provider composition in multi-cloud environments. This involves the composition of offerings from different cloud carriers to satisfy the service necessities of a person. By way of using ACO, services can be correctly selected and composed from a couple of carriers, considering elements including cost, excellent, and availability. ACO has proven promising outcomes in enhancing the overall performance and reliability of carrier composition in multi-cloud environments, making it a precious technique in todays tremendously distributed computing landscape.

关键词： Measurement Ant colony optimization Costs Shape Metaheuristics MIMICs Quality of service Reliability Distributed computing

来源：评论

学校读者我要写书评

暂无评论

OMPO: a unified framework for RL under policy and dynamics shifts 24

OMPO: a unified framework for RL under policy and dynamics s...

引用

Proceedings of the 41st International Conference on Machine Learning

作者： Yu Luo Tianying Ji Fuchun Sun Jianwei Zhang Huazhe Xu Xianyuan Zhan Department of Computer Science and Technology Tsinghua University Department of Informatics University of Hamburg Institute for Interdisciplinary Information Sciences Tsinghua University and Shanghai Qi Zhi Institute and Shanghai Artificial Intelligence Laboratory Shanghai Artificial Intelligence Laboratory and Institute for AI Industry Research Tsinghua University

Training reinforcement learning policies using environment interaction data collected from varying policies or dynamics presents a fundamental challenge. Existing works often overlook the distribution discrepancies induced by policy or dynamics shifts, or rely on specialized algorithms with task priors, thus often resulting in suboptimal policy performances and high learning variances. In this paper, we identify a unified strategy for online RL policy learning under diverse settings of policy and dynamics shifts: transition occupancy matching. In light of this, we introduce a surrogate policy learning objective by considering the transition occupancy discrepancies and then cast it into a tractable min-max optimization problem through dual reformulation. Our method, dubbed Occupancy-Matching Policy Optimization (OMPO), features a specialized actor-critic structure equipped with a distribution discriminator and a small-size local buffer. We conduct extensive experiments based on the OpenAI Gym, Meta-World, and Panda Robots environments, encompassing policy shifts under stationary and non-stationary dynamics, as well as domain adaption. The results demonstrate that OMPO outperforms the specialized baselines from different categories in all settings. We also find that OMPO exhibits particularly strong performance when combined with domain randomization, highlighting its potential in RL-based robotics applications

关键词：

来源：评论

学校读者我要写书评

暂无评论

Seizing serendipity: exploiting the value of past success in off-policy actor-critic 24

Seizing serendipity: exploiting the value of past success in...

引用

Proceedings of the 41st International Conference on Machine Learning

作者： Tianying Ji Yu Luo Fuchun Sun Xianyuan Zhan Jianwei Zhang Huazhe Xu Department of Computer Science and Technology Tsinghua University Institute for AI Industry Research Tsinghua University and Shanghai Artificial Intelligence Laboratory Department of Informatics University of Hamburg Shanghai Artificial Intelligence Laboratory and Institute for Interdisciplinary Information Sciences Tsinghua University and Shanghai Qi Zhi Institute

Learning high-quality Q-value functions plays a key role in the success of many modern off-policy deep reinforcement learning (RL) algorithms. Previous works primarily focus on addressing the value overestimation issue, an outcome of adopting function approximators and off-policy learning. Deviating from the common viewpoint, we observe that Q-values are often underestimated in the latter stage of the RL training process, potentially hindering policy learning and reducing sample efficiency. We find that such a long-neglected phenomenon is often related to the use of inferior actions from the current policy in Bellman updates as compared to the more optimal action samples in the replay buffer. To address this issue, our insight is to incorporate sufficient exploitation of past successes while maintaining exploration optimism. We propose the Blended Exploitation and Exploration (BEE) operator, a simple yet effective approach that updates Q-value using both historical best-performing actions and the current policy. Based on BEE, the resulting practical algorithm BAC outperforms state-of-the-art methods in over 50 continuous control tasks and achieves strong performance in failure-prone scenarios and real-world robot tasks. Benchmark results and videos are available at https://***/BEE/.

关键词：

来源：评论

学校读者我要写书评

暂无评论

Psychological Stress Classification Using Extreme Gradient Boosting Algorithm

Psychological Stress Classification Using Extreme Gradient B...

引用

International Conference on Information and Communication technology Convergence (ICTC)

作者： Maria Mahardini Sakanti Viacheslav Siniaev Aurelia Amaris Win-Jet Luo Suhartono C. Bambang Dwi Kuncoro Graduate Institute of Precision Manufacturing National Chin-Yi University of Technology Taichung Taichung Taiwan Institute of Applied Computer Science ITMO University St. Petersburg Russian Federation Department of Informatics Engineering Faculty of Computer Science Universitas Brawijaya Malang Indonesia Ph.D. Program Prospective Technology of Electrical Engineering and Computer Science National Chin-Yi University of Technology Taichung Taichung Taiwan Department of Refrigeration Air Conditioning and Energy Engineering National Chin-Yi University of Technology Taichung Taiwan

ISBN: (数字)9798350364637

ISBN: (纸本)9798350364644

Stress, as a reaction to threatening situations, can raise heart rate and result in serious conditions that might cause significant damage or even be life-threatening. Traditional methods for evaluating stress, which rely on subjective self-reporting and clinical assessments, often suffer from biases and inconsistencies. Artificial intelligence models have been explored to predict stress levels more accurately. This paper investigates the application of Extreme Gradient Boosting in classifying psychological stress using the WESAD dataset, which includes parameters such as acceleration, electrocardiogram, electromyography, electrodermal activity, temperature, and respiration. The dataset was balanced and sampled to create a manageable subset for experimental. Extreme Gradient Boosting was chosen for its efficiency and scalability in handling complex datasets. The model was trained and validated, achieving a 95% accuracy in predicting stress levels. This study highlights the potential of integrating Extreme Gradient Boosting models into wearable devices for real-time stress monitoring. Future work involves optimizing the model to utilize fewer sensors without decreasing accuracy, ensuring it can be integrated into portable/wearable systems using tiny microcontrollers.

关键词： Accuracy Microcontrollers Psychology Human factors Predictive models Boosting Prediction algorithms Real-time systems Sensors Wearable devices

来源：评论

学校读者我要写书评

暂无评论

Asynchronous Deep Reinforcement Learning for Collaborative Task Computing and On-Demand Resource Allocation in Vehicular Edge Computing

引用

IEEE Transactions on Intelligent Transportation Systems 2023年第12期24卷 15513-15526页

作者： Liu, Lei Feng, Jie Mu, Xuanyu Pei, Qingqi Lan, Dapeng Xiao, Ming Xidian University Guangzhou Institute of Technology Guangzhou510555 China Ministry of Education Shanghai201804 China Xidian University State Key Laboratory of Integrated Services Networks School of Telecommunication Engineering Shaanxi Xi'an710071 China Xi'an University of Posts and Telecommunications Shaanxi Key Laboratory of Information Communication Network and Security Shaanxi Xi'an710121 China Xidian University State Key Laboratory of Integrated Services Networks Xi'an710071 China University of Oslo Department of Informatics Oslo0373 Norway Royal Institute of Technology School of Electrical Engineering and Computer Science Stockholm10044 Sweden

Vehicular Edge Computing (VEC) is enjoying a surge in research interest due to the remarkable potential to reduce response delay and alleviate bandwidth pressure. Facing the ever-growing service applications in VEC, how to effectively aggregate and flexibly schedule ubiquitous network resources for implementing diverse tasks and meeting differentiated demands from numerous vehicular users remains haunting. Toward this end, we investigate collaborative task computing and on-demand resource allocation. The collaborative computing framework in VEC is provided to support deep collaboration and intelligent management of heterogeneous resources widely distributed in vehicles, edge servers and cloud. Based on this framework, the joint optimization problem of distributed task offloading and multi-resource management is formulated with the aim to maximize the system utility by making the optimal task and resource scheduling policy, the novelty of which lies in the exploration of available vehicle resources and the consideration of service migration. In view of the dynamics, randomness and time-variant of vehicular networks, the asynchronous deep reinforcement algorithm is leveraged to find the optimal solution. Extensive simulation experiments are implemented to demonstrate the superiority of our proposed algorithm in terms of response latency compared with full offloading and random offloading. © 2000-2011 IEEE.

关键词： Reinforcement learning

来源：评论

学校读者我要写书评

暂无评论

Offline-boosted actor-critic: adaptively blending optimal historical behaviors in deep off-policy RL 24

Offline-boosted actor-critic: adaptively blending optimal hi...

引用

Proceedings of the 41st International Conference on Machine Learning

Off-policy reinforcement learning (RL) has achieved notable success in tackling many complex real-world tasks, by leveraging previously collected data for policy learning. However, most existing off-policy RL algorithms fail to maximally exploit the information in the replay buffer, limiting sample efficiency and policy performance. In this work, we discover that concurrently training an offline RL policy based on the shared online replay buffer can sometimes outperform the original online learning policy, though the occurrence of such performance gains remains uncertain. This motivates a new possibility of harnessing the emergent outperforming offline optimal policy to improve online policy learning. Based on this insight, we present Offline-Boosted Actor-Critic (OBAC), a model-free online RL framework that elegantly identifies the outperforming offline policy through value comparison, and uses it as an adaptive constraint to guarantee stronger policy learning performance. Our experiments demonstrate that OBAC outperforms other popular model-free RL baselines and rivals advanced model-based RL methods in terms of sample efficiency and asymptotic performance across 53 tasks spanning 6 task suites.

关键词：

来源：评论

学校读者我要写书评

暂无评论

Chat Application Utilizing Automatic Translation Built on The Microsoft Azure Platform

引用

Procedia computer Science 2024年 245卷 1101-1110页

作者： Yulius Denny Prabowo Raymond Wangsa Putra Martinus Maslim BINUS Online Learning Computer Science Department Bina Nusantara University Jakarta 11480 Indonesia Institute of Information Management School of Management National Cheng Kung Unversity Tainan Taiwan Informatics Department Faculty of Industrial Technology Universitas Atma Jaya Yogyakarta Daerah Istimewa Yogyakarta Indonesia

Language serves as a means of interpersonal communication. There exist various categories of language, such as 'dialect,' 'slang,' 'jargon,' and 'lingo.' In addition to linguistic proficiency, it is essential to possess fundamental language skills to effectively utilize a language for communication purposes. The four skills include listening, speaking, reading, and writing. With the advancement of technology, we utilize many media to communicate in different languages. The most fundamental form of communication involves speaking and listening, whereas other mediums such as smartphones require the ability to read and write to send and receive messages. Only a select few individuals can proficiently perform four different abilities simultaneously at an advanced level. In addition to that, there is also a linguistic barrier, requiring the use of translation software. If someone engages in conversation with another person who faces a language barrier, they will require two essential tools: translation software and a means of communication. Therefore, this study aims to explore the development of a chat program that can effectively eliminate language barriers. Consequently, this software was effectively constructed utilizing the Microsoft Azure platform and offers three language options as the basis for translation.

关键词： chat program Microsoft Azure translation

来源：评论

学校读者我要写书评

暂无评论

Real Time Prediction Model for Air Pollution and Air Quality Index based on Machine Learning

Real Time Prediction Model for Air Pollution and Air Quality...

引用

Advanced Innovations in Smart Cities (ICAISC), International Conference on

作者： Rakesh Kumar Bhisham Sharma Shashi Shekhar Imed Ben Dhaou Saurabh Singhal Department of Computer Engineering & Applications GLA University Mathura India Chitkara University Institute of Engineering and Technology Chitkara University Punjab India Department of Computer Science Hekma School of Engineering Computing and Informatics Dar Al-Hekma University Jeddah Saudi Arabia

Controlling air pollution is a difficult issue for governments in densely populated and developing nations. The burning of fossil fuels, industrial parameters and traffic assume critical parts in contamination of air. There is distinctive particulate matter which decide the nature of the air however among all the particulate matter, consideration towards particulate matter (PM 2.5) is become a necessity. In this paper we detect the PM value using image processing technology. Image processing uses edge detection and depth estimation techniques to get the contaminated regions of the picture. Accordingly, image processing is used to detect air pollution. It detects and quantifies contamination in the air with the image features like time, day/night, outdoor conditions for determining the correlation. The proposal uses the learning model based on these parameters to predict PM level on collected photos. High-level of PM can cause major issues on individuals’ wellbeing. As a result, regulating it by just being vigilant on its overall visibility is critical. This paper proposes a method for identifying and evaluating PM contamination by distinguishing six image features: transmission, sky perfection and shading, complete and neighborhood picture difference, and picture entropy. To assess the association between PM level and numerous elements, we also analyze the time, terrain, and climate state of each image. We created a relapse model based on these data to forecast PM2.5 levels in a specific city.

关键词： Technological innovation Smart cities Atmospheric modeling Predictive models Air pollution Real-time systems Indexes

来源：评论

学校读者我要写书评

暂无评论

Enhancing Safety and Independence: A Smart Switch Solution for Elderly Individuals Living Independently

Enhancing Safety and Independence: A Smart Switch Solution f...

引用

International Conference on System Modeling & Advancement in Research Trends (SMART)

作者： Prakash Babu Nazia Ahmad Sajithunisa Hussain Remya P. George Ravi Shankar Shukla Vishal Jain Dept of IT Acharya Institute of Technology Bangalore India Computer Science Department The Applied College Imam Abdulrahman Bin Faisal University Saudi Arabia Department of Computer Science College of Computing and Informatics Saudi Electronic University Saudi Arabia Sharda University India

As the world's population ages, more and more people will need assistance with daily tasks, such as making sure their homes are safe and secure for themselves and their loved ones. This study presents a new smart switch technology that aims to make houses safer, more independent, and more comfortable for older people. The smart switch uses the Internet of Things (IoT) and an intuitive user interface to make it easy for the elderly to control their household appliances, which in turn reduces the likelihood of accidents and allows caregivers to remotely monitor their well-being. This switch has been meticulously designed to meet the unique physical and cognitive requirements of older users by actively involving those users in the design process. Initial research shows that users' home environment management skills significantly increase, which has a favorable effect on their everyday lives and overall health. By improving the lives of elderly people who live independently, the study does double duty: it advances smart home technology and meets a pressing social need.

关键词：

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：