Filters








296 Hits in 4.2 sec

Using An Attention-based LSTM Encoder-Decoder Network for Near Real-Time Disturbance Detection

Yuan Yuan, Lei Lin, Lian-Zhi Huo, Yun-Long Kong, Zeng-Guang Zhou, Bin Wu, Yan Jia
2020 IEEE Journal of Selected Topics in Applied Earth Observations and Remote Sensing  
Accurate prediction of future observations based on past data is the key to near real-time disturbance detection using satellite image time series (SITS).  ...  Based on the proposed model, we develop a framework for near real-time disturbance detection and verify its effectiveness in the case of burned area mapping.  ...  Cristina Milesi for her linguistic assistance during the preparation of this manuscript.  ... 
doi:10.1109/jstars.2020.2988324 fatcat:uwvt4d74hjacbfdkbzfebouxd4

Attention-augmented Spatio-Temporal Segmentation for Land Cover Mapping [article]

Rahul Ghosh, Praveen Ravirathinam, Xiaowei Jia, Chenxi Lin, Zhenong Jin, Vipin Kumar
2021 arXiv   pre-print
The availability of massive earth observing satellite data provide huge opportunities for land use and land cover mapping.  ...  We also visualise the attention weights to study its effectiveness in mitigating noise and identifying discriminative time period.  ...  Attention based aggregation The spatio-temporal hidden representations are aggregated using an attention mechanism which calculates the relevance scores for each time-step based on their contribution to  ... 
arXiv:2105.02963v3 fatcat:z2hvrakncnggxffwdpstax6vva

An Exploratory Analysis on Visual Counterfeits using Conv-LSTM Hybrid Architecture

Mohammad Farukh Hashmi, B Kiran Kumar Ashish, Avinash G. Keskar, Neeraj Dhanraj Bokde, Jin Hee Yoon, Zong Woo Geem
2020 IEEE Access  
This temporal-detection pipeline compares very minute visual traces on the faces of real and fake frames using Convolutional Neural Network (CNN) and stores the abnormal features for training.  ...  The proposed algorithm and designed network set a new benchmark for detecting the visual counterfeits and show how this system can achieve competitive results on any fake generated video or image.  ...  LSTM is mainly used as an RNN language model to generate a sequence of words framewise based on visual features extracted framewise. The model deploys an encoder-decoder network for image captioning.  ... 
doi:10.1109/access.2020.2998330 fatcat:dnmk264igjfrvbfwex3awk55sa

Seismic Reflection Analysis of AETA Electromagnetic Signals

Zhenyu Bao, Shanshan Yong, Xin'an Wang, Chao Yang, Jinhan Xie, Chunjiu He
2021 Applied Sciences  
Therefore, the AETA electromagnetic disturbance signal can be used as an earthquake precursor and for further earthquake prediction.  ...  Generally, the electromagnetic signals of AETA stations near the epicenter have abnormal disturbances before an earthquake.  ...  [21] In an LAE, both the encoder and decoder are LSTM networks.  ... 
doi:10.3390/app11135869 fatcat:674o2yf4zbghlceytrfspzeywu

Wind Power Forecasting Using Attention-Based Recurrent Neural Networks: A Comparative Study

Bin Huang, Yuying Liang, Xiaolin Qiu
2021 IEEE Access  
To interpret how the three models under consideration-the long-and short-term time-series network (LSTNet), the temporal pattern attention-based long short-term memory (TPA-LSTM) and the dual-stage attention-based  ...  INDEX TERMS Wind power forecast, time-series forecast, recurrent neural network, attention, deep learning, DA-RNN, LSTNet, TPA-LSTM.  ...  To be elaborate 1) Given the exogenous input series X, an LSTM network is used to obtain encoder hidden states.  ... 
doi:10.1109/access.2021.3065502 fatcat:3eahff4pzjauvbqgpi56iufylm

Time Series Forecasting and Classification Models Based on Recurrent with Attention Mechanism and Generative Adversarial Networks

Kun Zhou, Wenyong Wang, Teng Hu, Kai Deng
2020 Sensors  
Then, long short-term memory with autoencoder and attention-based models, the temporal convolutional network and the generative adversarial model are proposed and applied to time series classification  ...  The unstable training process for generative adversarial network is circumvented by tuning hyperparameters and carefully choosing the appropriate optimizer of "Adam".  ...  The near-real-time disturbance detection method was based on least-squares spectral and cross-wavelet analyses and could be used as assessment of the results of time series [13] .  ... 
doi:10.3390/s20247211 pmid:33339314 pmcid:PMC7766176 fatcat:ksw4kioxmjhrxjlap6r4g3lpae

SoPhie: An Attentive GAN for Predicting Paths Compliant to Social and Physical Constraints

Amir Sadeghian, Vineet Kosaraju, Ali Sadeghian, Noriaki Hirose, Hamid Rezatofighi, Silvio Savarese
2019 2019 IEEE/CVF Conference on Computer Vision and Pattern Recognition (CVPR)  
We present SoPhie; an interpretable framework based on Generative Adversarial Network (GAN), which leverages two sources of information, the path history of all the agents in a scene, and the scene context  ...  information, using images of the scene.  ...  Similar to [8] , first an LSTM is used to capture the temporal dependency between all states of an agent i and encode them into a high dimensional feature representation for time t, i.e.  ... 
doi:10.1109/cvpr.2019.00144 dblp:conf/cvpr/SadeghianKSHRS19 fatcat:ldtxunvo4bbsvp4cpel7xnl47i

SoPhie: An Attentive GAN for Predicting Paths Compliant to Social and Physical Constraints [article]

Amir Sadeghian, Vineet Kosaraju, Ali Sadeghian, Noriaki Hirose, S. Hamid Rezatofighi, Silvio Savarese
2018 arXiv   pre-print
We present SoPhie; an interpretable framework based on Generative Adversarial Network (GAN), which leverages two sources of information, the path history of all the agents in a scene, and the scene context  ...  information, using images of the scene.  ...  Similar to [8] , first an LSTM is used to capture the temporal dependency between all states of an agent i and encode them into a high dimensional feature representation for time t, i.e.  ... 
arXiv:1806.01482v2 fatcat:gvk2i5zrnvflhbwh6nvinjmfeq

Online estimation of the heat flux during turning using long short-term memory based encoder-decoder

Jinghui Han, Long Xu, Kaiwei Cao, Tianxiang Li, Xianhua Tan, Zirong Tang, Tielin Shi, Guanglan Liao
2021 Case Studies in Thermal Engineering  
In this paper, we introduce a long short-term memory (LSTM) based encoder-decoder (ED) scheme in online estimation of the heat flux imposed at the tool-chip region during turning.  ...  acceptable time cost for online process.  ...  Nonlinear IHCP solving by LSTM-ED A long short-term memory recurrent neural network (RNN) based encoder-decoder called LSTM-ED is proposed to solve the nonlinear IHCP at real time.  ... 
doi:10.1016/j.csite.2021.101002 fatcat:drfura6jazcn5dndv57gsreuai

Deep learning-driven wireless communication for edge-cloud computing: opportunities and challenges

Huaming Wu, Xiangyi Li, Yingjun Deng
2020 Journal of Cloud Computing: Advances, Systems and Applications  
, encoding and decoding, and security and privacy.  ...  Autoencoder-based wireless communication models are drawing more and more attention [10] [11] [12] .  ...  Acknowledgements The authors thank the editor and anonymous reviewers for their helpful comments and valuable suggestions.  ... 
doi:10.1186/s13677-020-00168-9 fatcat:7n6r2pozgfb5rgfwyxoxpqxq3q

A Survey of Voice Translation Methodologies - Acoustic Dialect Decoder [article]

Hans Krupakar, Keerthika Rajvel, Bharathi B, Angel Deborah S, Vallidevi Krishnamurthy
2016 arXiv   pre-print
The recognition unit will use Hidden Markov Models (HMMs) Based Tool-Kit (HTK), hybrid RNN systems with gated memory cells, and the synthesis unit, HMM based speech synthesis system HTS.  ...  This system will initially be built as an English to Tamil translation device.  ...  To address this problem, a hybrid encoder decoder system with an attention mechanism has been used [23] .  ... 
arXiv:1610.03934v1 fatcat:3jgucadf6jcirnpygijtfew4ai

Identifying Cause-and-Effect Relationships of Manufacturing Errors using Sequence-to-Sequence Learning [article]

Jeff Reimer, Yandong Wang, Sofiane Laridi, Juergen Urdich, Sören Wilmsmeier, Gregory Palmer
2022 arXiv   pre-print
The timely completion of orders depends on the individual station-based operations concluding within their scheduled cycle times.  ...  Utilizing real-time information about conditions collected by a production data acquisition system, we propose a novel vehicle manufacturing analysis system, which uses deep learning to establish a link  ...  Acknowledgements The authors gratefully acknowledge, that the proposed research is a result of the research project "IIP-Ecosphere", granted by the German Federal Ministry for Economics and Climate Action  ... 
arXiv:2205.02827v1 fatcat:hl5hbj3jlzeqnmqyqa2dkvjrze

Learning Sparse Interaction Graphs of Partially Observed Pedestrians for Trajectory Prediction [article]

Zhe Huang, Ruohua Li, Kazuki Shin, Katherine Driggs-Campbell
2021 arXiv   pre-print
A Node Transformer Encoder and a Masked LSTM encode the pedestrian features with the sampled sparse graphs to predict trajectories.  ...  Thus, we propose Gumbel Social Transformer, in which an Edge Gumbel Selector samples a sparse interaction graph of partially observed pedestrians at each time step.  ...  that generates deterministic prediction; (4) Social GAN (SGAN) has a LSTM-based encoder-decoder architecture with a socially aware global pooling layer, and is trained using Generative Adversarial Networks  ... 
arXiv:2107.07056v2 fatcat:l55it5xqxfgh3k7reotldfqtsy

Bringing Emotion Recognition Out of the Lab into Real Life: Recent Advances in Sensors and Machine Learning

Stanisław Saganowski
2022 Electronics  
A survey on existing systems for recognizing emotions in real-life scenarios—their possibilities, limitations, and identified problems—is also provided.  ...  The review is concluded with a debate on what challenges need to be overcome in the domain in the near future.  ...  Moreover, the LSTM utilized as encoder-decoder is often extended with the attention mechanizm.  ... 
doi:10.3390/electronics11030496 fatcat:pagrnyshp5fq7nkcdcqd2gzdbm

Noise-Robust Wagon Text Extraction Based on Defect-Restore Generative Adversarial Network

Meng Lei, Yi Zhou, Li Zhou, Jiannan Zheng, Ming Li, Liang Zou
2019 IEEE Access  
The generator is made up of encoder-decoder-encoder sub-networks, enabling it to learn discriminative representations from the intermediate layer.  ...  The propose strategy yields an accuracy of 97.76% on 2682 real-world test sub-images, remarkably outperforms the prior arts.  ...  Then, the utilization of a sequential attention-based modeling mechanism enabled the network to select useful features.  ... 
doi:10.1109/access.2019.2954475 fatcat:bbxhcjdiungx7ckmyad2qdw5r4
« Previous Showing results 1 — 15 out of 296 results