检索结果-内蒙古大学图书馆

第33届中国控制与决策会议

作者： Ruijing Cui Jianbin Sun Jiajun Cheng Kewei Yang College of Systems Engineering National University of Defense Technology Information Research Center of Military Science PLA Academy of Military Science

ISBN: (纸本)9781665431293

The Global Innovation Index(GII) is employed to rank the innovation capability for all countries and economies in the world,which was proposed by Cornell University,the INSEAD,and the World Intellectual Property *** analyze the development status of innovation capability of a country or economy at any time,this paper proposes a pairwise comparison-based global innovation capability ranking prediction approach named *** experimental results show that:(1) The ranking of innovation capability can be predicted with PairwiseGIIRank effectively.(2) Compared with other classifiers,the approach performs the best when the k-nearest neighbors algorithm is used as its internal classifier.(3) To achieve the desired predictive effect,at least 5 indicators are *** ranking of innovation capability of countries and economies around the world can be predicted with PairwiseGIIRank before the GII report is *** provides decision-making support for decision-makers to adjust their innovation capability development policies.

关键词： Global Innovation Index(GⅡ) Innovation capability Pairwise comparison Classification Decision making

来源：评论

学校读者我要写书评

暂无评论

Enhancing Complex Causality Extraction via Improved Subtask Interaction and Knowledge Fusion

arXiv

引用

arXiv 2024年

作者： Gao, Jinglong Lu, Chen Ding, Xiao Li, Zhongyang Liu, Ting Qin, Bing Research Center for Social Computing and Information Retrieval Harbin Institute of Technology China National Key Laboratory of Information Systems Engineering Nanjing China Huawei Cloud Shenzhen China

Event Causality Extraction (ECE) aims at extracting causal event pairs from texts. Despite ChatGPT’s recent success, fine-tuning small models remains the best approach for the ECE task. However, existing fine-tuning based ECE methods cannot address all three key challenges in ECE simultaneously: 1) Complex Causality Extraction, where multiple causal-effect pairs occur within a single sentence;2) Subtask Interaction, which involves modeling the mutual dependence between the two subtasks of ECE, i.e., extracting events and identifying the causal relationship between extracted events;and 3) Knowledge Fusion, which requires effectively fusing the knowledge in two modalities, i.e., the expressive pretrained language models and the structured knowledge graphs. In this paper, we propose a unified ECE framework (UniCE) to address all three issues in ECE simultaneously. Specifically, we design a subtask interaction mechanism to enable mutual interaction between the two ECE subtasks. Besides, we design a knowledge fusion mechanism to fuse knowledge in the two modalities. Furthermore, we employ separate decoders for each subtask to facilitate complex causality extraction. Experiments on three benchmark datasets demonstrate that our method achieves state-of-the-art performance and outperforms ChatGPT with a margin of at least 30% F1-score. More importantly, our model can also be used to effectively improve the ECE performance of ChatGPT via in-context learning. © 2024, CC BY.

关键词： Knowledge graph

来源：评论

学校读者我要写书评

暂无评论

Opportunities and Challenges from Using Animal Videos in Reinforcement Learning for Navigation

引用

IFAC-PapersOnLine 2023年第2期56卷 9056-9061页

作者： Vittorio Giammarino James Queeney Lucas C. Carstensen Michael E. Hasselmo Ioannis Ch Paschalidis Division of Systems Engineering Boston University Boston MA 02215 USA Center for Systems Neuroscience Department of Psychological and Brain Sciences and Graduate Program for Neuroscience Boston University Boston MA 02215 USA College of Engineering and Hariri Institute for Computing & Computational Science & Engineering Boston University Boston MA 02215 USA

We investigate the use of animal videos (observations) to improve Reinforcement Learning (RL) efficiency and performance in navigation tasks with sparse rewards. Motivated by theoretical considerations, we make use of weighted policy optimization for off-policy RL and describe the main challenges when learning from animal videos. We propose solutions and test our ideas on a 2D navigation task. We show how the use of animal videos improves performance over RL algorithms that do not leverage such observations.

关键词： Learning for Control Reinforcement Learning Deep Learning RL with demonstrations Imitation Learning Bio-inspired Control

来源：评论

学校读者我要写书评

暂无评论

Automatic Diagnosis of Pectus Excavatum from CT Images Using a Joint CNN-LSTM Model 9

Automatic Diagnosis of Pectus Excavatum from CT Images Using...

引用

9th International Conference on Control, Decision and information Technologies, CoDIT 2023

作者： Liao, Yizhi Zhou, Haiyu Xie, Longhan Cai, Siqi School of Computing and Information Systems University of Melbourne Melbourne Australia Guangdong Lung Cancer Institute Guangdong Provincial People's Hospital Division of Thoracic Surgery Guangzhou China Shien-Ming Wu School of Intelligent Engineering South China University of Technology Guangzhou China National University of Singapore Department of Electrical and Computer Engineering Singapore

ISBN: (纸本)9798350311402

Pectus excavatum (PE) is one of the most common congenital sternal deformities. Accurate preoperative diagnosis of PE is of great significance for subsequent correction and improvement of the patient's quality of life. However, current diagnostic methods rely on the calculation of some PE indices, which is a heavy workload for physical therapists and suffers from measurement errors. To address this issue, we propose an end-to-end automatic assessment of PE, which features a cascaded structure of CNN and LSTM. Specifically, the high-level feature representations of CT images are extracted by the pretrained CNN and then processed through the LSTM layers for classification. In addition, we build up a medical image dataset for PE diagnosis by collecting chest CT images of 42 subjects. Results on this dataset show that the proposed CNN-LSTM framework achieves a relatively high accuracy of 90.20%, which provides a new perspective for the automatic diagnosis of PE in clinics. © 2023 IEEE.

关键词： Computerized tomography

来源：评论

学校读者我要写书评

暂无评论

DCPI-Depth: Explicitly Infusing Dense Correspondence Prior to Unsupervised Monocular Depth Estimation

arXiv

引用

arXiv 2024年

作者： Zhang, Mengtan Feng, Yi Chen, Qijun Fan, Rui College of Electronics & Information Engineering Shanghai Research Institute for Intelligent Autonomous Systems Shanghai Institute of Intelligent Science and Technology The State Key Laboratory of Intelligent Autonomous Systems Frontiers Science Center for Intelligent Autonomous Systems Tongji University Shanghai201804 China

There has been a recent surge of interest in learning to perceive depth from monocular videos in an unsupervised fashion. A key challenge in this field is achieving robust and accurate depth estimation in challenging scenarios, particularly in regions with weak textures or where dynamic objects are present. This study makes three major contributions by delving deeply into dense correspondence priors to provide existing frameworks with explicit geometric constraints. The first novelty is a contextual-geometric depth consistency loss, which employs depth maps triangulated from dense correspondences based on estimated ego-motion to guide the learning of depth perception from contextual information, since explicitly triangulated depth maps capture accurate relative distances among pixels. The second novelty arises from the observation that there exists an explicit, deducible relationship between optical flow divergence and depth gradient. A differential property correlation loss is, therefore, designed to refine depth estimation with a specific emphasis on local variations. The third novelty is a bidirectional stream co-adjustment strategy that enhances the interaction between rigid and optical flows, encouraging the former towards more accurate correspondence and making the latter more adaptable across various scenarios under the static scene hypotheses. DCPI-Depth, a framework that incorporates all these innovative components and couples two bidirectional and collaborative streams, achieves state-of-the-art performance and generalizability across multiple public datasets, outperforming all existing prior arts. Specifically, it demonstrates accurate depth estimation in texture-less and dynamic regions, and shows more reasonable smoothness. Our source code will be publicly available at ***/DCPI-Depth upon publication. Copyright © 2024, The Authors. All rights reserved.

关键词： Optical flows

来源：评论

学校读者我要写书评

暂无评论

Gridless Bayesian Inference for DOA Estimation with Coprime Array

Gridless Bayesian Inference for DOA Estimation with Coprime ...

引用

2023 IEEE International Radar Conference, RADAR 2023

作者： Liu, Ying Zhang, Zongyu Zhou, Chengwei Shi, Zhiguo Zhejiang University College of Control Science and Engineering Hangzhou310027 China Zhejiang University College of Information Science and Electronic Engineering Hangzhou310027 China Key Laboratory of Collaborative Sensing and Autonomous Unmanned Systems of Zhejiang Province Hangzhou310015 China Zhejiang University International Joint Innovation Center Haining314400 China

ISBN: (纸本)9781665482783

In this paper, we propose a gridless Bayesian inference based direction-of-arrival (DOA) estimation algorithm with coprime array to overcome the basis mismatch arising in conventional on-grid sparse signal reconstruction based algorithms. Under Bayesian framework, the distributions of DOAs are assumed as mixtures of von Mises models yielding the continuous parameterization of the direction information, while the signals power is governed by a Bernoulli-Laplace model turning the sources number estimation into binary sequence detection. The Bayesian model is formulated on the second-order virtual coarray signals and the variational Bayesian inference is performed to realized accurate DOA estimation with increased number of degrees of freedom (DOFs). Simulation results show the effectiveness of the proposed algorithm especially at low signal-to-noise ratio (SNR). © 2023 IEEE.

关键词： Binary sequences

来源：评论

学校读者我要写书评

暂无评论

Self-Propelled Capsule Endoscope and Tracking Control of Its Wireless Power Supply Coil 1

Self-Propelled Capsule Endoscope and Tracking Control of Its...

引用

1st International Conference on Robotics, engineering, Science, and Technology, RESTCON 2024

作者： Ito, Takahiro Fujita, Kohei Miyauchi, Kenji Murakami, Sunao Ohtake, Hiroshi Katsuta, Chihiro Karasawa, Kaoru Tanaka, Akane Matsuda, Hiroshi Kimura, Toshihiro Kutsumi, Hiromu Computer Science and Systems Engineering Kyushu Institute of Technology Iizuka Japan Tokyo University of Agriculture and Technology Division of Animal Life Science Fuchu-shi Japan Savior Inc Yokohama Japan Center for Clinical Research and Advanced Medicine Establishment Shiga University of Medical Science Otsu-Shi Japan

ISBN: (纸本)9798350348842

We have developed the self-propelled capsule endoscope. Although it is the same size as a conventional capsule endoscope, with a diameter of 11 mm and a length of 31 mm, it can travel inside the gastrointestinal tract. A running mechanism that uses vibrations allows it to run without any protrusions on the outer shell. Therefore, there is no risk of harming the inside of the body. This is a noteworthy feature of this capsule. Experiments have confirmed that this capsule can travel in the intestines of living dogs. Because the power supply to this capsule needs to continue stably for over an hour, we developed a wireless power supply method to the internal capsule. Conventional wireless power supply with a fixed power transmission coil had the problem that power supply would be interrupted if the capsule moved. Therefore, we investigated stabilizing the power supply by moving the power transmission coil in X-Y to follow the capsule in response to increases and decreases in the power supply. We were able to confirm the feasibility through experiments. This technology can be expected to perform endoscopic examinations with less burden on both patients and doctors, and can also be applied to power supply to medical devices inside the body. © 2024 IEEE.

关键词： Endoscopy

来源：评论

学校读者我要写书评

暂无评论

Secure Status Updates for Internet of Drones: A Deep Q-Learning-Based Antenna Selection Approach

Secure Status Updates for Internet of Drones: A Deep Q-Learn...

引用

International Wireless Communications and Mobile Computing Conference, IWCMC

作者： Yuquan Xiao Qinghe Du Chen Lu Yizhuo Wang School of Information and Communications Engineering Xi’an Jiaotong University Xi’an China Shaanxi Smart Networks and Ubiquitous Access Research Center Xi’an China National Simulation Education Center for Communications and Information Systems Xi’an China School of Information and Communication Engineering Shenzhen Institute of Information Technology Shenzhen China

ISBN: (数字)9798350361261

ISBN: (纸本)9798350361278

Status update applications are very common in the internet of drones (IoD), where some of status update information are privacy-sensitive. How these information can be efficiently and securely transmitted remains as an open challenging issue. Motivated by this concern, we in this paper consider using the transmit antenna selection (TAS) technique to guarantee secure status updates over IoD downlink networks, in which only the drones corresponding to the highest channel gain on each antenna can become the candidate receivers, and then the base station (BS) selects one antenna to transmit the associated status update information with wiretap coding to the corresponding candidate. The mobility of drones causes the channel gain is time-varying, and thus one natural question arises, i.e., what is the optimal antenna selection scheme in terms of the long-term network-wide freshness performance. To answer this question, the weighted sum of average age-of-information (AoI) minimization problem is formulated, and we propose a deep Q-learning-based antenna selection approach to solve it. It is worth mentioning that, in light of the try-error mechanism of Q-learning, no prior knowledge of the wireless environments is required for our proposal in contrast to the traditional schemes. Finally, the numerical results verify that our proposal can further reduce the weighted sum of AoI as compared with the state-of-the-art max-weight scheme.

关键词： Q-learning Wireless networks Surveillance Transmitting antennas Receiving antennas Minimization Downlink

来源：评论

学校读者我要写书评

暂无评论

Deep Network-based Slow Feature Analysis for Human Fall Detection

Deep Network-based Slow Feature Analysis for Human Fall Dete...

引用

2022 International Conference on Computational Modelling, Simulation and Optimization, ICCMSO 2022

作者： Pramanik, Anima Venkatagiri, Kavya Sarkar, Sobhan Pal, Sankar K. IIT Kharagpur Department of Industrial & Systems Engineering WB Kharagpur721302 India IIT Kharagpur Department of Mining and Safety Engineering Kharagpur721302 India IIM Ranchi Information Systems & Business Analytics Jharkhand Ranchi834008 India Indian Statistical Institute Center for Soft Computing Research West Bengal Kolkata700108 India

ISBN: (纸本)9798350332889

One of the most concerning safety hazards for elderly people is abnormal falls in public places. Vision-based fall detection using ambient cameras is a popular non-intrusive solution. Recent research uses Slow Feature Analysis (SFA), which can learn the slow invariant varying shape features obtained from input signals and is efficient. Another recent famous approach in motion detection is deep learning. However, the fall event in actual cases is diverse, resulting in complications in the detection task. Additionally, it is difficult to acquire fall-related data;hence, simulation is done on fall events to generate a training dataset, resulting in smaller data. Considering these complications, we have presented a novel method by combining SFA, deep learning models, namely Convolutional Neural Network (CNN) and Long Short Term Memory (LSTM), and rule-base. CNN is used to extract the object region, thereby reducing the region of interest (RoI). Two shape features, such as aspect ratio and area of RoI are considered as input to the LSTM for retrieving the temporal information which is further used for rule generation, thereby increasing the detection accuracy. The efficacy of the proposed method for various features, such as aspect ratio, area, and aspect ra tio+area is demonstrated over the UR Fall data with an accuracy of 95.2%, 93.8%, and 96.36%, respectively. © 2022 IEEE.

关键词： Long short-term memory

来源：评论

学校读者我要写书评

暂无评论

Superior electrocatalytic negative electrode with tailored nitrogen functional group for vanadium redox flow battery

引用

Journal of Energy Chemistry 2023年第3期78卷 148-157,I0005页

作者： Min Gu Kang Wook Ahn Joonhee Kang Shin Ae Song Kiyoung Kim Ju Young Woo Yong-Cheol Jeong Bonwook Koo Dae Soo Jung Sung Nam Lim Micro/Nano Scale Manufacturing Group Korea Institute of Industrial Technology143 HanggaulroSangrok-guAnsan-siGyeonggi-do 15588Republic of Korea Department of Energy Systems Engineering Soonchunhyang University22 Soonchunhyang-roShinchang-myeonAsan-siChungcheongnam-do 31538Republic of Korea Department of Nanoenergy Engineering Pusan National UniversityBusan 46241Republic of Korea Department of Nano Fusion Technology Pusan National UniversityBusan 46241Republic of Korea Energy Storage Materials Center Energy&Environmental DivisionKorea Institute of Ceramic Engineering&Technology(KICET)101 Soho-roJinju-siGyeongsangnam-doRepublic of Korea

Development of electrodes with high electrocatalytic activity and stability is essential for solving problems that still restrict the extensive application of vanadium redox flow batteries(VRFBs).Here,we designed a novel negative electrode with superior electrocatalytic activity by tailoring nitrogen functional groups,such as newly formed nitro and pyridinic-N transformed to pyridonic-N,from the prenitrogen-doped *** was experimentally confirmed that an electrode with pyridonic-N and nitro fuctional groups(tailored nitrogen-doped graphite felt,TNGF) has superior electrocatalytic acivity with enhanced electron and mass *** functional theory calulations demonstrated the pyridonic-N and nitro functional groups promoted the adsorption,charge transfer,and bond formation with the vanadium species,which is consistent with expermental *** addition,the V2+/V3+redox reaction mechanism on pyridonic-N and nitro functional groups was estabilised based on density functional theory(DFT) *** TNGF was applied to a VRFB,it enabled enhanced-electrolyte utilization and energy efficiencies(EE) of 57.9% and 64.6%,respectively,at a current density of 250 mA cm^(-2).These results are 18.6% and 8.9% higher than those of VRFB with electrode containing graphitic-N and pyridinicN ***,TNGF-based VRFB still operated with an EE of 59% at a high current density of300 mA cm^(-2).The TNGF-based VRFB exhibited stable cycling performance without noticeable decay of EE over 450 charge-discharge cycles at a current density of 250 mA cm^(-2).The results of this study suggest that introducing pyridonic-N and nitro groups on the electrode is effective for improving the electrochemical performance of VRFBs.

关键词： Tailored nitrogen functional group Pyridonic-N Nitro Negative electrode Density functional theory Vanadium redoxﬂow batteries

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：