检索结果-内蒙古大学图书馆

RPR-Net: A Point Cloud-based Rotation-aware Large Scale Place Recognition Network

学校读者我要写书评

暂无评论

arXiv 2021年

作者： Fan, Zhaoxin Song, Zhenbo Zhang, Wenping Liu, Hongyan He, Jun Du, Xiaoyong Key Laboratory of Data Engineering and Knowledge Engineering of MOE School of Information Renmin University of China Beijing100872 China Department of Management Science and Engineering Tsinghua University Beijing100084 China School of Computer Science and Engineering Nanjing University of Science and Technology Nanjing210094 China

Point cloud-based large scale place recognition is an important but challenging task for many applications such as Simultaneous Localization and Mapping (SLAM). Taking the task as a point cloud retrieval problem, previous methods have made delightful achievements. However, how to deal with catastrophic collapse caused by rotation problems is still under-explored. In this paper, to tackle the issue, we propose a novel Point Cloud-based Rotation-aware Large Scale Place Recognition Network (RPR-Net). In particular, to solve the problem, we propose to learn rotation-invariant features in three steps. First, we design three kinds of novel Rotation-Invariant Features (RIFs), which are low-level features that can hold the rotation-invariant property. Second, using these RIFs, we design an attentive module to learn rotation-invariant kernels. Third, we apply these kernels to previous point cloud features to generate new features, which is the well-known SO(3) mapping process. By doing so, high-level scene-specific rotation-invariant features can be learned. We call the above process an Attentive Rotation-Invariant Convolution (ARIConv). To achieve the place recognition goal, we build RPR-Net, which takes ARIConv as a basic unit to construct a dense network architecture. Then, powerful global descriptors used for retrieval-based place recognition can be sufficiently extracted from RPR-Net. Experimental results on prevalent datasets show that our method achieves comparable results to existing state-of-the-art place recognition models and significantly outperforms other rotation-invariant baseline models when solving rotation problems. Copyright © 2021, The Authors. All rights reserved.

关键词： Network architecture

A Privacy-preserving Location Recommendation Scheme without Trustworthy Entity

学校读者我要写书评

暂无评论

A Privacy-preserving Location Recommendation Scheme without ...

IEEE International Conference on Trust, Security and Privacy in Computing and Communications (TrustCom)

作者： Changli Zhou Jing Peng Ying Ma Qingfeng Jiang College of Computer Science and Technology Huaqiao University Xiamen PR China Key Laboratory of Data Mining and Intelligent Recommendation Fujian Province University Xiamen PR China Xiamen Key Laboratory of Data Security and Blockchain Technology Huaqiao University Xiamen China College of Computer Science and Engineering Changshu Institute of Technology Changshu PR China

ISBN: (纸本)9781665416597

The widespread application of location recommendation service brings convenience and social fun to users, but the misuse of user check-in data and location recommendation results can bring serious privacy leakage, because these check-in data and location recommendation results imply private information of users' interests and behavioral hobbies. To address the check-in privacy and recommendation result privacy leakage issues in existing location recommendation schemes, we first propose a location recommendation algorithm based on the similarity of user historical check-ins, and then we devise three location recommendation protocols on ciphertext to protect the privacy of user check-in data and recommendation results based on the homomorphic property of Paillier cryptosystem without assuming any trustworthy entity. These protocols calculate the ciphertexts of recommendation results at the location recommendation server(LSP) without learning any explicit check-in privacy of the user, and it also guarantees that the decryption server(DS) can only implement the decryption operations as requested by the user without being aware of any decrypted recommendation results, so that the privacy of user check-in data and recommendation results are both preserved solidly in our protocols. Although the above recommendation process is carried out without any user's explicit data, it is still effective enough to guarantee the accuracy of recommendation results. Finally, we prove the security, efficiency and accuracy of our scheme by extensive performance analysis.

关键词： Privacy data privacy Protocols Conferences Prediction algorithms Performance analysis Cryptography

Privileged graph distillation for cold start recommendation

学校读者我要写书评

暂无评论

arXiv 2021年

作者： Wang, Shuai Zhang, Kun Wu, Le Ma, Haiping Hong, Richang Wang, Meng Key Laboratory of Knowledge Engineering with Big Data Hefei University of Technology China School of Computer Science and Information Engineering Hefei University of Technology China Institute of Artificial Intelligence Hefei Comprehensive National Science Center China Anhui University China

The cold start problem in recommender systems is a long-standing challenge, which requires recommending to new users (items) based on attributes without any historical interaction records. In these recommendation systems, warm users (items) have privileged collaborative signals of interaction records compared to cold start users (items), and these Collaborative Filtering (CF) signals are shown to have competing performance for recommendation. Many researchers proposed to learn the correlation between collaborative signal embedding space and the attribute embedding space to improve the cold start recommendation, in which user and item categorical attributes are available in many online platforms. However, the cold start recommendation is still limited by two embedding spaces modeling and simple assumptions of space transformation. As user-item interaction behaviors and user (item) attributes naturally form a heterogeneous graph structure, in this paper, we propose a privileged graph distillation model (PGD). The teacher model is composed of a heterogeneous graph structure for warm users and items with privileged CF links. The student model is composed of an entity-attribute graph without CF links. Specifically, the teacher model can learn better embeddings of each entity by injecting complex higher-order relationships from the constructed heterogeneous graph. The student model can learn the distilled output with privileged CF embeddings from the teacher embeddings. Our proposed model is generally applicable to different cold start scenarios with new user, new item, or new user-new item. Finally, extensive experimental results on the real-world datasets clearly show the effectiveness of our proposed model on different types of cold start problems, with average 6.6%, 5.6%, and 17.1% improvement over state-of-the-art baselines on three datasets, respectively. © 2021, CC BY.

关键词： Embeddings

Indoor Sound Source Localization Algorithm Based on BP Neural Network 21

学校读者我要写书评

暂无评论

Indoor Sound Source Localization Algorithm Based on BP Neura...

21st IEEE International Conference on Communication Technology, ICCT 2021

作者： Wang, Lan Zhang, Kun Shen, Chong Wang, Chai Fu, Xixi Hainan University State Key Laboratory of Marine Resources Utilization in South China Sea Hainan Haikou570228 China Data Mining and Monitoring and Early Warning Technol. of Isl. Tourism Rsrc. of Key Lab. of the Min. of Culture AndTourism of the People's Rep. of China Hainan Tropical Ocean University Hainan Sanya572022 China College of Information and Communication Engineering Hainan University Hainan Haikou570228 China

ISBN: (纸本)9781665432061

Sound source localization is an important part of the perception of things around. Sound source localization can overcome the shortcomings of visual localization, and can also locate the invisible place. The application of sound source localization in indoor is the latest trend. Starting from the application of deep learning to indoor sound source localization, this paper focuses on the analysis and research of BP neural network applied to indoor sound source localization algorithm. In this paper, an off-line sampling scheme is used to construct the network structure with 7 neurons hidden in the layer, and the BP algorithm of LevenBerg-Marquardt is used as the training function, this algorithm can solve the traditional algorithm through the study of the physical properties of sound, set up the corresponding equation, and then solve, the process is complex, to solve the difficult problem. The simulation results show that the algorithm can be implemented in 100 square meters of the house, through sampling 400 sets of data for machine training, positioning error can be controlled in a few centimeters effect. © 2021 IEEE.

关键词： Deep learning

Size-fixed group discovery via multi-constrained graph pattern matching

学校读者我要写书评

暂无评论

Information Sciences 2025年 690卷

作者： Guliu Liu Lei Li Guanfeng Liu Xindong Wu Key Laboratory of Knowledge Engineering with Big Data (Hefei University of Technology) Ministry of Education 230601 Hefei China School of Computer Science and Information Engineering Hefei University of Technology 230601 Hefei China Department of Computing Macquarie University Sydney 2122 Australia

Multi-Constrained Graph Pattern Matching (MC-GPM) aims to match a pattern graph with multiple attribute constraints on its nodes and edges, and has garnered significant interest in various fields, including social-based e-commerce and trust-based group discovery. However, the existing MC-GPM methods do not consider situations where the number of each node in the pattern graph needs to be fixed, such as finding experts group with expert quantities and relations specified. In this paper, a Multi-Constrained Strong Simulation with the Fixed Number of Nodes (MCSS-FNN) matching model is proposed, and then a Trust-oriented Optimal Multi-constrained Path (TOMP) matching algorithm is designed for solving it. Additionally, two heuristic optimization strategies are designed, one for combinatorial testing and the other for edge matching, to enhance the efficiency of the TOMP algorithm. Empirical experiments are conducted on four real social network datasets, and the results demonstrate the effectiveness and efficiency of the proposed algorithm and optimization strategies.

关键词：

Solar PV Power Forecasting Approach Based on Hybrid Deep Neural Network 6th

学校读者我要写书评

暂无评论

Solar PV Power Forecasting Approach Based on Hybrid Deep Neu...

6th International Conference on Advanced Machine Learning Technologies and Applications, AMLTA 2021

作者： Chang, Kuo-Chi Omer, Abdalaziz Altayeb Ibrahim Chu, Kai-Chun Chang, Fu-Hsiang Wang, Hsiao-Chuan Lin, Yuh-Chung Nguyen, Trong-The Sung, Tien-Wen Fujian Provincial Key Laboratory of Big Data Mining and Applications Fujian University of Technology Fuzhou China School of Information Science and Engineering Fujian University of Technology Fuzhou China School of Management Fujian University of Technology Fuzhou China Department of Tourism Shih-Hsin University Taipei Taiwan Institute of Environmental Engineering National Taiwan University Taipei Taiwan College of Mechanical & Electrical Engineering National Taipei University of Technology Taipei Taiwan Department of Business Administration North Borneo University College Sabah Malaysia

ISBN: (纸本)9783030697167

There are incredibly renewable outlets such as solar and wind intermittent, and machine reliability is hard to sustain with the intolerable balance of green energy resources. Because of population growth and technological development, global energy demand is increasing exponentially. However, civilization must address a stable energy source for two benefits, a cost-effective and sustainable basis for future renewable energy production. Solar Energy is a platform for green energy and is a potential means of solving problems that face a possible energy system. The power output is primarily determined by the properties of the incoming radiation and solar panels. Precise and accurate knowledge of these variables provides a consistent model for predicting the solar future. Solar power forecasts would have a significant effect on the viability of massive solar energy projects. This study presents a new approach to predict the expected amount of energy to be produced from Photovoltaics using a hybrid deep neural network based on a combination of extended short-term memory networks (LSTM) and convolutional neural networks (CNN), taking into account the energy of electrons and photons as well as elements and climatic conditions as input data. © 2021, The Author(s), under exclusive license to Springer Nature Switzerland AG.

关键词： Solar energy

Doppler Shift and Channel Estimation for Intelligent Transparent Surface Assisted Communication Systems on High-Speed Railways

学校读者我要写书评

暂无评论

arXiv 2022年

作者： Wang, Yirun Wang, Gongpu He, Ruisi Ai, Bo Tellambura, Chintha The School of Information and Electronics Beijing Institute of Technology Beijing100081 China The Beijing Key Laboratory of Transportation Data Analysis and Mining Beijing Jiaotong University Beijing100044 China The State Key Laboratory of Rail Traffic Control and Safety Beijing Jiaotong University Beijing100044 China The Department of Electrical and Computer Engineering University of Alberta EdmontonABT6G 2V4 Canada

The critical distinction between the emerging intelligent transparent surface (ITS) and intelligent reflection surface (IRS) is that the incident signals can penetrate the ITS instead of being reflected, which enables the ITS to combat the severe signal penetration loss for high-speed railway (HSR) wireless communications. This paper thus investigates the channel estimation problem for an ITS-assisted HSR network where the ITS is embedded into the carriage window. We first formulate the channels as functions of physical parameters, and thus transform the channel estimation into a parameter recovery problem. Next, we design the first two pilot blocks within each frame and develop a serial low-complexity channel estimation algorithm. Specifically, the channel estimates are initially obtained, and each estimate is further expressed as the sum of its perfectly known value and the estimation error. By leveraging the relationship between channels for the two pilot blocks, we recover the Doppler shifts from the channel estimates, based on which we can further acquire other channel parameters. Moreover, the Cramér-Rao lower bound (CRLB) for each parameter is derived as a performance benchmark. Finally, we provide numerical results to establish the effectiveness of our proposed estimators. Copyright © 2022, The Authors. All rights reserved.

关键词： Doppler effect

FortisAVQA and MAVEN: a Benchmark dataset and Debiasing Framework for Robust Multimodal Reasoning

学校读者我要写书评

暂无评论

arXiv 2025年

作者： Ma, Jie Gao, Zhitao Chai, Qi Liu, Jun Wang, Pinghui Tao, Jing Su, Zhou Ministry of Education of Key Laboratory for Intelligent Networks and Network Security School of Cyber Science and Engineering Xi’an Jiaotong University Shaanxi Xi’an710049 China Shannxi Provincial Key Laboratory of Big Data Knowledge Engineering School of Computer Science and Technology Xi’an Jiaotong University Shaanxi Xi’an710049 China Guangdong Guangzhou510000 China

Audio-Visual Question Answering (AVQA) is a challenging multimodal reasoning task requiring intelligent systems to answer natural language queries based on paired audio-video inputs accurately. However, existing AVQA approaches often suffer from overfitting to dataset biases, leading to poor robustness. Moreover, current datasets may not effectively diagnose these methods. To address these challenges, we first introduce a novel dataset, FortisAVQA, constructed in two stages: (1) rephrasing questions in the test split of the public MUSIC-AVQA dataset and (2) introducing distribution shifts across questions. The first stage expands the test space with greater diversity, while the second enables a refined robustness evaluation across rare, frequent, and overall question distributions. Second, we introduce a robust Multimodal Audio-Visual Epistemic Network (MAVEN) that leverages a multifaceted cycle collaborative debiasing strategy to mitigate bias learning. Experimental results demonstrate that our architecture achieves state-of-the-art performance on FortisAVQA, with a notable improvement of 7.81%. Extensive ablation studies on both datasets validate the effectiveness of our debiasing components. Additionally, our evaluation reveals the limited robustness of existing multimodal QA methods. We also verify the plug-and-play capability of our strategy by integrating it with various baseline models across both datasets. Our dataset and code are available at https://***/reml-group/fortisavqa. Copyright © 2025, The Authors. All rights reserved.

关键词：

Q&A driven conversational recommendation integrating multiple interest modeling

学校读者我要写书评

暂无评论

Neurocomputing 2025年 647卷

作者： Haiping Zhu Yuqi Sun Shuwei Che Yan Chen School of Computer Science and Technology Xi’an Jiaotong University Xi’an 710049 China Shaanxi Province Key Laboratory of Big Data Knowledge Engineering Xi’an Jiaotong University Xi’an 710049 China Ministry of Education Key Laboratory of Intelligent Networks and Network Security Xi’an Jiaotong University Xi’an 710049 China

The conversational recommendation aims to provide users with recommendations in an interaction form of “System Ask-User Respond”. Existing studies rarely consider combining users’ multi-interest at the attribute level for preference modeling, and conduct inefficient conversational recommendation strategy learning, which affects the recommendation performance and conversation performance. To this end, we proposed a Q-A driven conversational recommendation method integrating multiple interest modeling. Specifically, we integrate the user’s positive and negative feedback to model a session dynamic graph, then use the signed graph convolutional network for graph representation learning, and we model multiple interest sequences based on the attention mechanism, then obtain the user interest state representation by fusing of sequence representations to solve the problem of insufficient user preference representation. Besides, we proposed a personalized decision space optimization method to narrow the range of action candidates and train the model with a multi-agent reinforcement learning method integrating hierarchical decision debiasing to solve the problem of poor conversational recommendation strategy learning effect. Experimental results on three public datasets, LastFM, Yelp, and Book, show that compared with existing conversational recommendation methods, our method demonstrates consistent performance improvement across all datasets. In addition, the results of ablation experiments verify the effectiveness of each component in our method.

关键词：