检索结果-内蒙古大学图书馆

Dual-path transformer network: Direct context-aware modeling for end-to-end monaural speech separation

学校读者我要写书评

暂无评论

arXiv 2020年

作者： Chen, Jingjing Mao, Qirong Liu, Dong School of Computer Science and Communication Engineering Jiangsu University China Jiangsu Engineering Research Center of big data ubiquitous perception and intelligent agriculture applications Zhenjiang China

The dominant speech separation models are based on complex recurrent or convolution neural network that model speech sequences indirectly conditioning on context, such as passing information through many intermediate states in recurrent neural network, leading to suboptimal separation performance. In this paper, we propose a dual-path transformer network (DPTNet) for end-to-end speech separation, which introduces direct context-awareness in the modeling for speech sequences. By introduces a improved transformer, elements in speech sequences can interact directly, which enables DPTNet can model for the speech sequences with direct context-awareness. The improved transformer in our approach learns the order information of the speech sequences without positional encodings by incorporating a recurrent neural network into the original transformer. In addition, the structure of dual paths makes our model efficient for extremely long speech sequence modeling. Extensive experiments on benchmark datasets show that our approach outperforms the current state-of-the-arts (20.6 dB SDR on the public WSj0-2mix data corpus). Copyright © 2020, The Authors. All rights reserved.

关键词： Separation

Statistical Detection of Collective data Fraud

学校读者我要写书评

暂无评论

arXiv 2020年

作者： Wang, Ruoyu Sun, Daniel Li, Guoqiang School of Software Shanghai Jiao Tong University China School of Computer Science and Engineering University of New South Wales Australia Data61 CSIRO Australia

—Statistical divergence is widely applied in multimedia processing, basically due to regularity and explainable features displayed in data. However, in a broader range of data realm, these advantages may not out-stand, and therefore a more general approach is required. In data detection, statistical divergence can be used as an similarity measurement based on collective features. In this paper, we present a collective detection technique based on statistical divergence. The technique extracts distribution similarities among data collections, and then uses the statistical divergence to detect collective anomalies. Our technique continuously evaluates metrics as evolving features and calculates adaptive threshold to meet the best mathematical expectation. To illustrate details of the technique and explore its efficiency, we case-studied a real world problem of click farming detection against malicious online sellers. The evaluation shows that these techniques provided efficient classifiers. They were also sufficiently sensitive to a much smaller magnitude of data alteration, compared with real world malicious behaviours. Thus, it is applicable in the real world. Copyright © 2020, The Authors. All rights reserved.

关键词： Statistics

Mass flow rates effect on the performance of PV/T bi-fluid hybrid collector (single and simultaneous modes) 2nd

学校读者我要写书评

暂无评论

Mass flow rates effect on the performance of PV/T bi-fluid h...

2nd International Conference on Electronic engineering and Renewable Energy, ICEERE 2020

作者： El Manssouri, Oussama El Fouas, Chaimae Hajji, Bekkay Rabhi, Abdelhamid Tina, Giuseppe Marco Gagliano, Antonio Renewable Energy Embedded System and Data Processing Laboratory National School of Applied Sciences Mohamed First University Oujda Morocco Modelization Information and Systems Laboratory Picardie Jules Verne University Amiens France DIEEI: Department of Electric Electronic and Computer Engineering University of Catania Catania Italy

ISBN: (纸本)9789811562587

A hybrid photovoltaic/thermal collector (PV/T) is used to produce simultaneously electrical and thermal energy from absorbed solar irradiation. The research to date has tended to focus on either bi-fluids (water and air) as the working fluid to supply energy needs for different applications. The purpose of this work is to test the performances of PV/T at different operating modes of fluid e.g. the air mode, the water mode, and the simultaneous mode (water& air). Furthermore the effect of mass flow rate as a key parameter for better electrical and thermal performances has been investigated. The PV/T performances were assessed based on a dynamic numerical model. An energy balance equations have been established for each layer, then implemented in MATLAB software. The results show that thermal efficiency in the simultaneous mode (air & water) is better compared to others modes. The thermal efficiencies for independently fluid condition have ranged from approximately 20 to 48%, and increased to a maximum efficiency of near to 68% for the case of the simultaneously fluids. This result indicates that the optimum mass flow rates for air and water are 0.035 kg/s and 0.007 kg/s respectively. Therefore, the theoretical model developed of the independently and simultaneously operational modes is validated, evidencing a good fit between simulation results and the experimental data available in literature experimental. © Springer Nature Singapore Pte Ltd. 2021.

关键词： Air

Adaptive domain of dynamic distribution based on manifold space

学校读者我要写书评

暂无评论

Adaptive domain of dynamic distribution based on manifold sp...

IEEE International Conference on Big Knowledge (ICBK)

作者： Daoyuan Yu Xuegang Hu Key Laboratory of Knowledge Engineering with Big Data (Hefei University of Technology) Ministry of Education School of Computer Science and Information Engineering Hefei University of Technology Hefei China

ISBN: (数字)9781728181561

ISBN: (纸本)9781728181578

Domain adaption aims to use the source domain knowledge to assist the model learning. Most of the existing methods are based on the feature representation learning model, which are achieved by aligning the data distribution between two domains. However, in the process of feature representation learning, due to the diversity of distribution differences between domains, it faces the problems of degenerated feature transformation and unevaluated distribution alignment, which bring challenges to the existing research. Therefore, the dynamic distribution alignment factors of marginal distribution and conditional distribution are introduced to reduce the spatial distortion in the process of feature representation, and the domain adaptation is realized based on the dynamic manifold regularization constraints and structural risk minimization learning. In this paper, we propose the structural risk minimization and an improved dynamic manifold regularization constraint to solve these problems. Experimental results show that compared with traditional methods and deep level methods, the algorithm in this paper has a significant improvement in classification accuracy.

关键词： Manifolds Feature extraction Task analysis Distortion Adaptation models Risk management Kernel

SharDAG: Scaling DAG-Based Blockchains Via Adaptive Sharding 40

学校读者我要写书评

暂无评论

SharDAG: Scaling DAG-Based Blockchains Via Adaptive Sharding

40th IEEE International Conference on data engineering, ICDE 2024

作者： Cheng, Feng Xiao, Jiang Liu, Cunyang Zhang, Shijie Zhou, Yifan Li, Bo Li, Baochun Jin, Hai School of Computer Science and Technology Huazhong University of Science and Technology National Engineering Research Center for Big Data Technology and System Services Computing Technology and System Lab Cluster and Grid Computing Lab China Hong Kong University of Science and Technology Hong Kong University of Toronto Canada

ISBN: (纸本)9798350317152

Directed Acyclic Graph (DAG)-based blockchain (a.k.a distributed ledger) has become prevalent for supporting highly concurrent applications. Its inherent parallel data structure accelerates block generation significantly, shifting the bottleneck from performance to storage scalability. An intuitive solution is to apply state sharding that divides the entire ledger (i.e., transactions and states) into multiple shards. While each node only stores proportional transactions, it suffers from the challenges of storing and ensuring the processing consistency of cross-shard transactions. In this paper, we propose SharDAG, a new mechanism that leverages adaptive sharding for DAG-based blockchains to achieve high performance and strong consistency. The key idea of SharDAG is to exploit unique characteristics -silent assets -and design a lightweight processing mechanism based on avatar account caching. Furthermore, we design a Byzantine resilient cross-shard verification mechanism with a theoretically optimal number of participating nodes, which guarantees the consistency and security of avatar account aggregation. Our comprehensive evaluations on real-world workloads demonstrate that SharDAG presents up to 3.8 x throughput improvement compared to the state-of-the-art and reduces the storage overhead of cross-shard transactions. © 2024 IEEE.

关键词： Blockchain

Dynamic Interactional And Cooperative Network For Shield Machine

学校读者我要写书评

暂无评论

arXiv 2022年

作者： Gao, Dazhi Li, Rongyang Mao, Lingfeng Wang, Hongbo Ning prof, Huansheng The School of Computer and Communication Engineering University of Science and Technology Beijing Beijing100083 China Beijing Engineering Research Center for Cyberspace Data Analysis and Applications Beijing China

The shield machine (SM) is a complex mechanical device used for tunneling. However, the monitoring and deciding were mainly done by artificial experience during traditional construction, which brought some limitations, such as hidden mechanical failures, human operator error, and sensor anomalies. To deal with these challenges, many scholars have studied SM intelligent methods. Most of these methods only take SM into account but do not consider the SM operating environment. So, this paper discussed the relationship among SM, geological information, and control terminals. Then, according to the relationship, models were established for the control terminal tasks, including SM rate prediction and SM anomaly detection. The experimental results show that compared with baseline models, the proposed model in this paper perform better. In the proposed model, the R2 and MSE of rate prediction can reach 92.2%, and 0.0064 respectively. The abnormal data detection rate of the anomaly detection task is up to 98.2%. Copyright © 2022, The Authors. All rights reserved.

关键词： Anomaly detection

MBLLEn: Low-light image/video enhancement using CNNs 29

学校读者我要写书评

暂无评论

MBLLEn: Low-light image/video enhancement using CNNs

29th British Machine Vision Conference, BMVC 2018

作者： Lv, Feifan Lu, Feng Wu, Jianhua Lim, Chongsoon State Key Laboratory of Virtual Reality Technology and Systems School of Computer Science and Engineering Beihang University Beijing China Beijing Advanced Innovation Center for Big Data-Based Precision Medicine Beihang University Beijing China Panasonic R and D Center Singapore Singapore City Singapore

We present a deep learning based method for low-light image enhancement. This problem is challenging due to the difficulty in handling various factors simultaneously including brightness, contrast, artifacts and noise. To address this task, we propose the multi-branch low-light enhancement network (MBLLEN). The key idea is to extract rich features up to different levels, so that we can apply enhancement via multiple subnets and finally produce the output image via multi-branch fusion. In this manner, image quality is improved from different aspects. Through extensive experiments, our proposed MBLLEN is found to outperform the state-of-art techniques by a large margin. We additionally show that our method can be directly extended to handle low-light videos. © 2018. The copyright of this document resides with its authors.

关键词： Image enhancement

Explainable High-order Visual Question Reasoning: A New Benchmark and Knowledge-routed Network

学校读者我要写书评

暂无评论

arXiv 2019年

作者： Cao, Qingxing Li, Bailin Liang, Xiaodan Lin, Liang School of Intelligent Systems Engineering Sun Yat-sen University China DarkMatter AI Research School of Data and Computer Science Sun Yat-sen University China

Explanation and high-order reasoning capabilities are crucial for real-world visual question answering with diverse levels of inference complexity (e.g., what is the dog that is near the girl playing with?) and important for users to understand and diagnose the trustworthiness of the system. Current VQA benchmarks on natural images with only an accuracy metric end up pushing the models to exploit the dataset biases and cannot provide any interpretable justification, which severally hinders advances in high-level question answering. In this work, we propose a new HVQR benchmark for evaluating explainable and high-order visual question reasoning ability with three distinguishable merits: 1) the questions often contain one or two relationship triplets, which requires the model to have the ability of multistep reasoning to predict plausible answers;2) we provide an explicit evaluation on a multistep reasoning process that is constructed with image scene graphs and commonsense knowledge bases;and 3) each relationship triplet in a large-scale knowledge base only appears once among all questions, which poses challenges for existing networks that often attempt to overfit the knowledge base that already appears in the training set and enforces the models to handle unseen questions and knowledge fact usage. We also propose a new knowledge-routed modular network (KM-net) that incorporates the multistep reasoning process over a large knowledge base into visual question reasoning. An extensive dataset analysis and comparisons with existing models on the HVQR benchmark show that our benchmark provides explainable evaluations, comprehensive reasoning requirements and realistic challenges of VQA systems, as well as our KM-net’s superiority in terms of accuracy and explanation ability. Copyright © 2019, The Authors. All rights reserved.

关键词： Knowledge based systems

Unmanned Aerial Vehicle Classification and Detection Based on Deep Transfer Learning

学校读者我要写书评

暂无评论

Unmanned Aerial Vehicle Classification and Detection Based o...

Intelligent Computing and Human-computer Interaction (ICHCI), International Conference on

作者： Wei Meng Meng Tia School of computer and information engineering Henan University Henan Key Laboratory of Big Data Analysis and Processing Henan Engineering Laboratory of Spatial Information Processing Kaifeng China

Currently, major security incidents caused by the “unlicensed flying” of Unmanned Aerial Vehicle (UAV) have emerged one after another, which poses a grave threat to the security issues of public facilities and sensitive areas. Whether it can timely detect and prevent “unlicensed flying” of UAV has become a social concern. In response to this demand, the transfer learning method is adopted in this paper to conduct twoclassification and detection on UAV images. Image recognition technology based on transfer learning is an effective method to improve recognition accuracy by applying deep learning models to small samples. Different from the large number of training samples required by deep learning, transfer learning transfers the weights of the pre-trained deep neural network, and uses only small sample data to obtain good results in UAV image recognition. First of all, this paper proposes to construct a UAV data set according to different types of UAV shape structures, to perfect the classification and detection effect and the generalization ability of the model. Then, based on the transfer learning method, experimental comparison is made between three classic deep convolutional neural network classification models (Inception V3, ResNet 101 and VGG16) and two classic deep convolutional neural network detection models (Faster RCNN and SSD). Finally, an experimental evaluation is conducted on the collected UAV test data set. Compared with the traditional recognition model, the image classification model based on transfer learning employed in this paper has achieved important improvements in accuracy, recall and precision. Especially in the InceptionV3 model of transfer training, the recall reaches 96.98%. In addition, the image detection model based on transfer learning has achieved good detection results in accuracy, recall and F1-score.

关键词： Training Deep learning Image recognition Shape Transfer learning Buildings Unmanned aerial vehicles