检索结果-内蒙古大学图书馆

48th IEEE International Conference on Acoustics, Speech and Signal Processing, ICASSP 2023

作者： Tsai, Tsung-Han Wan, Wei-Chung National Central University Department of Electrical Engineering Taoyuan City Taiwan

ISBN: (纸本)9781728163277

Monocular Depth Estimation is a popular and challenging problem for many years. IR CNNs (Convolutional Neural Networks)-based method with encoder-decoder architecture is proposed and shows a reasonable result. In this paper, we propose a SE-Net-based module for the decoder part in the encoder-decoder architecture to improve the result. We proposed a DSE (decoder-Squeeze-and-Excitation) module to deal with the whole up-sampling process globally for the decoder part. We also include the Non-local Network space attention method to design the Non-Local decoder-Squeeze-and-Excitation (NL-DSE) module. The proposed NL-DSE module is installed and evaluated on the NYU Depth V2 dataset and achieves higher accuracy. Moreover, the design is independent of the encoder-decoder architecture and can be applied in the other encoder-decoder networks to have a more accurate network. © 2023 IEEE.

关键词： CNNs encoder-decoder monocular depth estimation SE-Net up-sampling recalibration

来源：评论

学校读者我要写书评

暂无评论

An Infrastructure Segmentation Neural Network for UAV Remote Sensing Images 5

An Infrastructure Segmentation Neural Network for UAV Remote...

引用

5th International Conference on Robotics, Intelligent Control and Artificial Intelligence, RICAI 2023

作者： Ma, Shuo Zhang, Xiaokang Fan, Haojia Li, Teng School of Control Science and Engineering Shandong University Jinan China

ISBN: (纸本)9798350357950

Infrastructure segmentation (bridges, dams, etc.) based on remote sensing images has been widely used in urban planning, land management, disaster monitoring and other monitoring fields. As a mature remote sensing monitoring platform, unmanned aerial vehicles (UA V) have become the main source of low-altitude remote sensing image acquisition of infrastructures. However, with the increasing data acquired by UAVs, the large data volume and high resolution of remote sensing images lead to challenges for the task of infrastructure segmentation. To solve the difficulties, this paper proposes a deep learning method for infrastructure segmentation based on adaptive hidden layer feature fusion, which specifically contains an encoder network, an adaptive layer selection mechanism, and a decoder network. In addition, this paper proposes a composite loss function containing two evaluation metrics, namely, boundary evaluation loss and Tversky loss, for training the proposed neural network. The validation of this paper is carried out on a real-world dataset, and the experimental results demonstrate the high accuracy and reliability of the proposed method on the infrastructure segmentation task. © 2023 IEEE.

关键词： adaptive layer selection deep learning encoder-decoder infrastructure segmentation

来源：评论

学校读者我要写书评

暂无评论

Robust multi-scale time series prediction for building carbon emissions with explainable deep learning

引用

ENERGY AND BUILDINGS 2024年 312卷

作者： Chen, Chao Guo, Jing Zhang, Limao Wu, Xiaoying Yang, Zheng Huazhong Univ Sci & Technol Sch Civil & Hydraul Engn 1037 Luoyu Rd Wuhan 430074 Hubei Peoples R China Huazhong Univ Sci & Technol Natl Ctr Technol Innovat Digital Construct 1037 Luoyu Rd Wuhan 430074 Hubei Peoples R China Glodon Corp Ltd Beijing 100000 Peoples R China Cent South Univ Sch Resources & Safety Engn Changsha 410006 Hunan Peoples R China

Accurate prediction dramatically enhances the effectiveness of carbon emission control, making it a valuable tool for achieving decarbonization goals. In this work, we propose a robust explainable encoder-decoder deep learning method for predicting multi-scale short-term carbon emissions and feature importance analysis to predict buildings ' emission patterns accurately. In the encoder, the initial inputs are transformed into a sequence to capture the time series and spatial information, where the model predicts the building emissions with an attention mechanism in the decoder. The hybrid model is trained and validated with a multi-scale dataset collected from a university. Results reveal that (1) After considering the buildings ' interactions, the R 2 values regarding individual building emissions predictions are 0.943, 0.938, 0.941, and 0.943, respectively, climbing to 0.952, 0.943, 0.943, and 0.950. (2) Benefiting from the proposed model ' s structure, the constructed model can accurately predict the community ' s emissions with the R 2 equal to 0.952 and keep excellent robustness after adding the white Gaussian noise data with all R 2 are above 0.919. (3) The historical hourly emissions, outdoor humidity and time of day are considered the most essential factors responsible for the building emissions. The novelty of this work lies in the multi-scale building emissions prediction that is conducted by the developed hybrid model simultaneously considering the impact of environmental factors and building interactions, allowing us to expand the model ' s application scenarios and provide valuable references for decarbonization decision-making.

关键词： encoder-decoder Time series prediction Building carbon emissions Explainable deep learning Decarbonization

来源：评论

学校读者我要写书评

暂无评论

High-Resolution Satellite Imagery Analysis for Terrain and Surface Data Extraction: Techniques and Applications 7

High-Resolution Satellite Imagery Analysis for Terrain and S...

引用

2023 7th International Conference On Computing, Communication, Control And Automation, ICCUBEA 2023

作者： Sukate, Sameeksha Wajage, Sakshi Verma, Yogesh Tawde, Parth Prasad, Jayashree MIT ADT University Computer Science and Engineering Pune India

ISBN: (纸本)9798350304268

With multiple applications and use cases of Digital Elevation Models (DEMs), focusing significantly on forming 3D models. It is used in various research areas such as flood modeling, agriculture, satellite navigation, farming, and forestry. This research paper presents a methodology for extracting Digital Elevation Models (DEMs) from satellite images using U-Net. The U-NET is a Deep Learning Model architecture that is commonly used for image segmentation tasks. It is a fully convolutional neural network that uses an encoder-decoder architecture to extract features from the input image and generate a pixel wise output. The U-Net model is ideal for DEM extraction from satellite images as it can effectively distinguish between the different features present in the images and accurately segment the ground elevation ranges. The creation of the dataset was a crucial step in the project as it ensured that the U - N et model was trained on a diverse set of images. The SWISSIMAGE dataset provided high-resolution RGB imagery, while the SRTM dataset provided the corresponding mask images with ground elevation ranges. The results obtained from the project indicate that the model can perform reasonably well even with limited data and without any significant pre-processing of the input dataset. However, there are still some challenges and limitations corresponding to the use of U-Net for DEM extraction. In conclusion, this study provides a valuable contribution to the field of remote sensing and computer vision, highlighting the potential of U-Net for DEM extraction from satellite images. In addition, the latest improvements in both software and hardware have made it possible to carry the processing with remarkable speed, even on moderately powerful hardware. We anticipate that our research will act as a catalyst for further advancements and will facilitate the integration of this research in generating, updating, and analysis of DEMs. However, further research is needed to addre

关键词： Deep learning Digital Elevation Models encoder-decoder Image segmentation Remote Sensing U-Net

来源：评论

学校读者我要写书评

暂无评论

Investigation of Transfer Learning for End-to-End Russian Speech Recognition 24th

Investigation of Transfer Learning for End-to-End Russian Sp...

引用

24th International Conference on Speech and Computer (SPECOM)

作者： Kipyatkova, Irina Russian Acad Sci SPC RAS St Petersburg Fed Res Ctr St Petersburg Russia

ISBN: (纸本)9783031209796;9783031209802

End-to-end speech recognition systems reduce the speech decoding time and required amount of memory comparing to standard systems. However they need much more data for training, which complicates creation of such systems for low-resourced languages. One way to improve performance of end-to-end low-resourced speech recognition system is model's pre-training by transfer learning, that is training the model on the non-target data and then transferring the trained parameters to the target model. The aim of the current research was to investigate application of transfer learning to the training of the end-to-end Russian speech recognition system in low-resourced conditions. We used several speech corpora of different languages for pre-training. Then end-to-end model was fine-tuned on a small Russian speech corpus of 60 h. We conducted experiments on application of transfer learning in different parts of the model (feature extraction block, encoder, and attention mechanism) as well as on freezing of the lower layers. We have achieved 24.53% relative word error rate reduction comparing to the baseline system trained without transfer learning.

关键词： End-to-end speech recognition Transfer learning encoder-decoder Russian speech

来源：评论

学校读者我要写书评

暂无评论

BART-Reader: Predicting Relations Between Entities via Reading Their Document-Level Context Information 11th

BART-Reader: Predicting Relations Between Entities via Readi...

引用

11th CCF International Conference on Natural Language Processing and Chinese Computing (NLPCC)

作者： Yan, Hang Sun, Yu Dai, Junqi Hu, Xiangkun Guo, Qipeng Qiu, Xipeng Huang, Xuanjing Fudan Univ Shanghai Key Lab Intelligent Informat Proc Shanghai Peoples R China Fudan Univ Sch Comp Sci Shanghai Peoples R China Amazon AWS AI Shanghai Peoples R China

ISBN: (纸本)9783031171208;9783031171192

Document-level relation extraction (Doc-RE) aims to classify relations between entities spread over multiple sentences. When one entity is paired with separate entities, the importance of its mentions varies, which means the entity representation should be different. However, most of the previous RE models failed to make the relation classification entity-pair aware effectively. To that end, we propose a novel adaptation to simultaneously utilize the encoder and decoder of the sequence-to-sequence (Seq2Seq) pre-trained model BART in a nongenerative manner to tackle the Doc-RE task. The encoder encodes the document to get the entity-aware contextualized mention representation. The decoder uses a non-causal self-attention mechanism and masked cross-attention to model the interactions between the entity pair under consideration explicitly. By doing so, we can fully take advantage of the pre-trained model in the encoder and decoder sides. And experiments in three Doc-RE datasets show that our model can not only take more advantage of BART, but surpass various BERT and RoBERTa based models.

关键词： Document-level relation extraction encoder-decoder

来源：评论

学校读者我要写书评

暂无评论

Attention-based efficient robot grasp detection network

引用

Frontiers of Information Technology & Electronic Engineering 2023年第10期24卷 1430-1444页

作者： Xiaofei QIN Wenkai HU Chen XIAO Changxiang HE Songwen PEI Xuedian ZHANG School of Optical-Electrical and Computer Engineering University of Shanghai for Science and TechnologyShanghai 200093China College of Science University of Shanghai for Science and TechnologyShanghai 200093China Shanghai Key Laboratory of Modern Optical System Shanghai 200093China Key Laboratory of Biomedical Optical Technology and Devices of Ministry of Education Shanghai 200093China Shanghai Institute of Intelligent Science and Technology Tongji UniversityShanghai 201210China

To balance the inference speed and detection accuracy of a grasp detection algorithm,which are both important for robot grasping tasks,we propose an encoder–decoder structured pixel-level grasp detection neural network named the attention-based efficient robot grasp detection network(AE-GDN).Three spatial attention modules are introduced in the encoder stages to enhance the detailed information,and three channel attention modules are introduced in the decoder stages to extract more semantic *** lightweight and efficient DenseBlocks are used to connect the encoder and decoder paths to improve the feature modeling capability of AE-GDN.A high intersection over union(IoU)value between the predicted grasp rectangle and the ground truth does not necessarily mean a high-quality grasp configuration,but might cause a *** is because traditional IoU loss calculation methods treat the center part of the predicted rectangle as having the same importance as the area around the *** design a new IoU loss calculation method based on an hourglass box matching mechanism,which will create good correspondence between high IoUs and high-quality grasp *** achieves the accuracy of 98.9%and 96.6%on the Cornell and Jacquard datasets,*** inference speed reaches 43.5 frames per second with only about 1.2×10^(6)*** proposed AE-GDN has also been deployed on a practical robotic arm grasping system and performs grasping *** are available at https://***/robvincen/robot_gradet.

关键词： Robot grasp detection Attention mechanism encoder-decoder Neural network

来源：评论

学校读者我要写书评

暂无评论

TransGlow: Attention-augmented Transduction model based on Graph Neural Networks for Water Flow Forecasting 22

TransGlow: Attention-augmented Transduction model based on G...

引用

22nd IEEE International Conference on Machine Learning and Applications, ICMLA 2023

作者： Roudbari, Naghmeh Shafiee Poullis, Charalambos Patterson, Zachary Eicker, Ursula Gina Cody School of Engineering and Computer Science Concordia University MontrealQC Canada

ISBN: (纸本)9798350345346

The hydrometric prediction of water quantity is useful for a variety of applications, including water management, flood forecasting, and flood control. However, the task is difficult due to the dynamic nature and limited data of water systems. Highly interconnected water systems can significantly affect hydrometric forecasting. Consequently, it is crucial to develop models that represent the relationships between other system components. In recent years, numerous hydrological applications have been studied, including streamflow prediction, flood forecasting, and water quality prediction. Existing methods are unable to model the influence of adjacent regions between pairs of variables. In this paper, we propose a spatiotemporal forecasting model that augments the hidden state in Graph Convolution Recurrent Neural Network (GCRN) encoder-decoder using an efficient version of the attention mechanism. The attention layer allows the decoder to access different parts of the input sequence selectively. Since water systems are interconnected and the connectivity information between the stations is implicit, the proposed model leverages a graph learning module to extract a sparse graph adjacency matrix adaptively based on the data. Spatiotemporal forecasting relies on historical data. In some regions, however, historical data may be limited or incomplete, making it difficult to accurately predict future water conditions. Further, we present a new benchmark dataset of water flow from a network of Canadian stations on rivers, streams, and lakes. Experimental results demonstrate that our proposed model TransGlow significantly outperforms baseline methods by a wide margin. © 2023 IEEE.

关键词： Attention encoder-decoder Graph Neural Net-works Spatiotemporal forecasting

来源：评论

学校读者我要写书评

暂无评论

Network traffic prediction with Attention-based Spatial-Temporal Graph Network

引用

COMPUTER NETWORKS 2024年 243卷

作者： Peng, Yufei Guo, Yingya Hao, Run Xu, Chengzhe Fuzhou Univ Coll Comp & Data Sci Fuzhou Peoples R China Fuzhou Univ Fujian Prov Key Lab Network Comp & Intelligent Inf Fuzhou Peoples R China Minist Educ Engn Res Ctr Big Data Intelligence Beijing Peoples R China

Network traffic prediction plays a significant role in network management. Previous network traffic prediction methods mainly focus on the temporal relationship between network traffic, and used time series models to predict network traffic, ignoring the spatial information contained in traffic data. Therefore, the prediction accuracy is limited, especially in long-term prediction. To improve the prediction accuracy of the dynamic network traffic in the long term, we propose an Attention -based Spatial-Temporal Graph Network (ASTGN) model for network traffic prediction to better capture both the temporal and spatial relations between the network traffic. Specifically, in ASTGN, we exploit an encoder-decoder architecture, where the encoder encodes the input network traffic and the decoder outputs the predicted network traffic sequences, integrating the temporal and spatial information of the network traffic data through the Spatio-Temporal Embedding module. The experimental results demonstrate the superiority of our proposed method ASTGN in long-term prediction.

关键词： Network traffic prediction Attention mechanism Graph neural network Temporal and spatial encoder-decoder

来源：评论

学校读者我要写书评

暂无评论

Dual temporal attention mechanism-based convolutional LSTM model for industrial dynamic soft sensor

引用

MEASUREMENT SCIENCE AND TECHNOLOGY 2024年第11期35卷 115103-115103页

作者： Cui, Jiarui Shi, Yuyu Huang, Jian Yang, Xu Gao, Jingjing Li, Qing Univ Sci & Technol Beijing Sch Automat & Elect Engn Beijing Peoples R China Univ Sci & Technol Beijing Shunde Innovat Sch Foshan Peoples R China

Deep learning is an appropriate methodology for modeling complex industrial data in the field of soft sensors, owing to its powerful feature representation capability. Given the nonlinear and dynamic nature of the process industry, the key challenge for soft sensor technology is to effectively mine dynamic information from long sequences and accurately extract features of relevance to quality. A dual temporal attention mechanism-based convolutional long short-term memory network (DTA-ConvLSTM) under an encoder-decoder framework is proposed as a soft sensor model to acquire quality-relevant dynamic features from serial data. Considering different influences of process variables for prediction at multiple time steps and various locations, ConvLSTM and temporal self-attention mechanism are utilized as the encoder to adaptively fuse spatiotemporal features and capture long-term dynamic properties of process in order to capture the trends of industrial variables. Furthermore, a quality-driven temporal attention mechanism is employed throughout the decoding process to dynamically select relevant features to more accurately track quality changes. The encoder-decoder model meticulously analyses the interactions between process and quality variables by incorporating dual-sequence dynamic information to improve the prediction performance. The validity and superiority of the DTA-ConvLSTM model was validated on two industrial case studies of the debutanizer column and sulfur recovery unit. Compared to the traditional LSTM model, the proposed model demonstrated a substantial improvement with the accuracy R2 up to 97.3% and 94.9% and the root mean square error reducing to 0.122 and 0.022.

关键词： dual temporal attention mechanism LSTM encoder-decoder convolutional layer soft sensor

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：