Filters








412 Hits in 5.0 sec

Missing Data Imputation with Adversarially-trained Graph Convolutional Networks [article]

Indro Spinelli, Simone Scardapane, Aurelio Uncini
2019 arXiv   pre-print
Missing data imputation (MDI) is a fundamental problem in many scientific disciplines.  ...  In order to speed-up training and improve the performance, we use a combination of multiple losses, including an adversarial loss implemented with the Wasserstein metric and a gradient penalty.  ...  Then we introduced the graph and the graph convolutional layer in our imputer (GINN) followed by the addition of the critic and the adversarial training (A-GINN), the skip connection (A-GINN skip) and  ... 
arXiv:1905.01907v1 fatcat:bbf2i3ubobcnldzmls5tsd4uoy

Generative Adversarial Networks for Spatio-temporal Data: A Survey [article]

Nan Gao, Hao Xue, Wei Shao, Sichen Zhao, Kyle Kai Qin, Arian Prabowo, Mohammad Saiedur Rahaman, Flora D. Salim
2021 arXiv   pre-print
Generative Adversarial Networks (GANs) have shown remarkable success in producing realistic-looking images in the computer vision area.  ...  Recently, GAN-based techniques are shown to be promising for spatio-temporal-based applications such as trajectory prediction, events generation and time-series data imputation.  ...  AE Graph Convolutional Network (GCN).  ... 
arXiv:2008.08903v3 fatcat:pbhxbfgw65bodksjdmwazwo4dq

Dynamic Spatiotemporal Graph Convolutional Neural Networks for Traffic Data Imputation with Complex Missing Patterns [article]

Yuebing Liang, Zhan Zhao, Lijun Sun
2021 arXiv   pre-print
To fill these research gaps, we propose a novel deep learning framework called Dynamic Spatiotemporal Graph Convolutional Neural Networks (DSTGCN) to impute missing traffic data.  ...  The model combines the recurrent architecture with graph-based convolutions to model the spatiotemporal dependencies.  ...  GWNET [17] : a graph neural network which captures spatial dependencies with generalized diffusion graph convolution layers and temporal dependencies with dilated convolution layers.  ... 
arXiv:2109.08357v1 fatcat:b6lht4xmkrbgnjflkrqjmlwkbi

GCRINT: Network Traffic Imputation Using Graph Convolutional Recurrent Neural Network

Van An Le Et Al.
2021 Zenodo  
In this paper, we propose Graph Convolutional Recurrent Neural Network for Imputing Network Traffic (GCRINT), a combination between Recurrent Neural Network (RNN) and Graph Convolutional Neural Network  ...  , for filling the missing values of network traffic data.  ...  More recently, Generative Adversarial Network (GAN) has been used to impute missing values.  ... 
doi:10.5281/zenodo.5105956 fatcat:fvzzrjg6rzhzra6qurcpdg4sye

Filling the G_ap_s: Multivariate Time Series Imputation by Graph Neural Networks [article]

Andrea Cini, Ivan Marisca, Cesare Alippi
2022 arXiv   pre-print
Conversely, graph neural networks have recently surged in popularity as both expressive and scalable tools for processing sequential data with relational inductive biases.  ...  In particular, we introduce a novel graph neural network architecture, named GRIN, which aims at reconstructing missing data in the different channels of a multivariate time series by learning spatio-temporal  ...  Acknowledgements This research is funded by the Swiss National Science Foundation project 200021_172671: "ALPSFORT: A Learning graPh-baSed framework FOr cybeR-physical sysTems."  ... 
arXiv:2108.00298v3 fatcat:ph65jwmufzb7ni5ykoeovb6fke

Networked Time Series Prediction with Incomplete Data [article]

Yichen Zhu, Mengtian Zhang, Bo Jiang, Haiming Jin, Jianqiang Huang, Xinbing Wang
2021 arXiv   pre-print
We propose NETS-ImpGAN, a novel deep learning framework that can be trained on incomplete data with missing values in both history and future.  ...  In this paper, we study the problem of NETS prediction with incomplete data.  ...  RIHGCN first uses Graph Convolutional Network (GCN) [18] and Long Short-Term Memory (LSTM) [14] network to impute missing data in the history part, and then uses the completed history to predict the  ... 
arXiv:2110.02271v1 fatcat:4pu5temxyjavzhrmbetnq3na5u

A Higher-Order Motif-Based Spatiotemporal Graph Imputation Approach for Transportation Networks

Difeng Zhu, Guojiang Shen, Jingjing Chen, Wenfeng Zhou, Xiangjie Kong, Yan Huang
2022 Wireless Communications and Mobile Computing  
It utilized graph convolution network (GCN) to polymerize the correlated segment attributes of the missing data segments.  ...  In this paper, by leveraging motif-based graph aggregation, we propose a spatiotemporal imputation approach to address the issue of traffic data missing.  ...  Generative adversarial network (GAN) provides a class of generative models for adversarial training, and it applies actual data/parallel data to generate the true data distribution, so that the imputation  ... 
doi:10.1155/2022/1702170 fatcat:l3d72kmgjfeqzksbgmiqx5mitm

Partial Convolutional LSTM for Spatiotemporal Prediction of Incomplete Data

Hyesook Son, Yun Jang.
2020 IEEE Access  
Spatial-temporal graph neural network (STGNN) is trained with spatiotemporal data in a graph structure, but most STGNNs predict only graph nodes [9] .  ...  Therefore, they first impute the dataset, and then they train the classifier with the imputed data to evaluate the imputation performance with the classification accuracy.  ... 
doi:10.1109/access.2020.3022774 fatcat:cy4ragkcjve4zfis3o26wlza7u

Graph Convolutional Networks for Graphs Containing Missing Features [article]

Hibiki Taguchi, Xin Liu, Tsuyoshi Murata
2020 arXiv   pre-print
Graph Convolutional Network (GCN) has experienced great success in graph analysis tasks. It works by smoothing the node features across the graph.  ...  However, real-world graph data are often incomplete and containing missing features.  ...  Among various kinds of GNNs, graph convolutional network (GCN) [37] , a simplified version of spectral graph convolutional networks [18, 62] , has attracted a large amount of attention.  ... 
arXiv:2007.04583v1 fatcat:kk43wrze65ewvoigiwnb624zia

Joint Modeling of Local and Global Temporal Dynamics for Multivariate Time Series Forecasting with Missing Values

Xianfeng Tang, Huaxiu Yao, Yiwei Sun, Charu Aggarwal, Prasenjit Mitra, Suhang Wang
2020 PROCEEDINGS OF THE THIRTIETH AAAI CONFERENCE ON ARTIFICIAL INTELLIGENCE AND THE TWENTY-EIGHTH INNOVATIVE APPLICATIONS OF ARTIFICIAL INTELLIGENCE CONFERENCE  
We further introduce adversarial training to enhance the modeling of global temporal distribution.  ...  and recurrent neural networks.  ...  Even if MTS imputation methods, such as multivariate imputation by chained equations (Azur et al. 2011) and generative adversarial network Luo et al., can be applied to fill in missing values first, training  ... 
doi:10.1609/aaai.v34i04.6056 fatcat:roupph2qevch3jbcsn5jmvlwda

Siamese Attribute-missing Graph Auto-encoder [article]

Wenxuan Tu, Sihang Zhou, Yue Liu, Xinwang Liu
2021 arXiv   pre-print
Graph representation learning (GRL) on attribute-missing graphs, which is a common yet challenging problem, has recently attracted considerable attention.  ...  First, we entangle the attribute embedding and structure embedding by introducing a siamese network structure to share the parameters learned by both processes, which allows the network training to benefit  ...  data imputation with adversarially-trained graph convolu- Hu, L.; Jian, S.; Cao, L.; Gu, Z.; Chen, Q.; and Amirbekyan, tional networks. Neural Networks 129: 249–260. A. 2019.  ... 
arXiv:2112.04842v1 fatcat:2hcdamwg5zf2diopgvp2whjdhi

Graph Markov Network for Traffic Forecasting with Missing Data [article]

Zhiyong Cui, Longfei Lin, Ziyuan Pu, Yinhai Wang
2019 arXiv   pre-print
Although missing values can be imputed, existing data imputation methods normally need long-term historical traffic state data.  ...  By incorporating the spectral graph convolution operation, we also propose a spectral graph Markov network (SGMN).  ...  This work was supported by the Connected Cities with Smart Transportation (C2SMART) Tier 1 University Transportation Center with the USDOT Award No.: 69A3551747124.  ... 
arXiv:1912.05457v1 fatcat:6i47zxumafhphjdy7bx7537xb4

Deconvolutional Networks on Graph Data [article]

Jia Li, Jiajin Li, Yang Liu, Jianwei Yu, Yueting Li, Hong Cheng
2021 arXiv   pre-print
In this paper, we consider an inverse problem in graph learning domain -- "given the graph representations smoothed by Graph Convolutional Network (GCN), how can we reconstruct the input graph signal?"  ...  We demonstrate the effectiveness of the proposed method on several tasks including graph feature imputation and graph structure generation.  ...  The model is trained on an incomplete version of feature matrix (training data) and further used to infer the potential missing features (test data).  ... 
arXiv:2110.15528v1 fatcat:udi5k5qd25hgxalh6nnt6mvfl4

Incomplete Graph Representation and Learning via Partial Graph Neural Networks [article]

Bo Jiang, Ziyan Zhang
2021 arXiv   pre-print
Graph Neural Networks (GNNs) are gaining increasing attention on graph data learning tasks in recent years.  ...  Existing GNNs are generally designed on complete graphs which can not deal with attribute-incomplete graph data directly.  ...  [10] propose Generative Adversarial Imputation Nets (GAIN) to complete missing data by adapting a GAN [11] framework. Chen et al.  ... 
arXiv:2003.10130v2 fatcat:2whdcqkyvrg7dprqlchn5rhmnq

Emerging Artificial Intelligence Applications in Spatial Transcriptomics Analysis [article]

Yijun Li, Stefan Stanojevic, Lana X. Garmire
2022 arXiv   pre-print
Such advancement comes with the urgent need for novel computational methods to handle the unique challenges of ST data analysis.  ...  utilizes GAN (Generative Adversarial Network) loss.  ...  Finally, DSTG feeds the link graph and concatenation of the pseudo-ST dataset and the real ST dataset into a graph convolutional neural network with multiple convolution layers, effectively learning a  ... 
arXiv:2203.09664v1 fatcat:om6cen2vsrai3jgrkrvolxpg3q
« Previous Showing results 1 — 15 out of 412 results