var DATABASE_BIB = `
A COMPARISON BETWEEN RECURSIVE NEURAL NETWORKS AND GRAPH NEURAL NETWORKS
  AUTH : Vincenzo Di Massa et al.
  SRCE : IEEE International Conference on Neural Networks - Conference Proceedings
  LINK : https://doi.org/10.1109/IJCNN.2006.246763
  TYPE : article
  DATE : 2006-01
  TAGS : GNNs
  DONE : false

A NOTE ON THE NP-HARDNESS OF TWO MATCHING PROBLEMS IN INDUCED SUBGRIDS
  AUTH : Marc Demange et al.
  SRCE : Discrete Mathematics and Theoretical Computer Science
  LINK : https://inria.hal.science/hal-00980770
  TYPE : article
  DATE : 2013-09
  TAGS : Math
  NOTE : Graph Theory
  DONE : false

APPRENTISSAGE PROFOND DES GRAPHES ATTRIBUÉS POUR LA CARTOGRAPHIE DU RISQUE DE LEPTOSPIROSE
  AUTH : Rodrigue Govan
  SRCE : Université de la Nouvelle-Calédonie
  LINK : https://hal.science/tel-05268357
  TYPE : book
  DATE : 2025-08
  DONE : false

A SURVEY ON MACHINE LEARNING SOLUTIONS FOR GRAPH PATTERN EXTRACTION
  AUTH : Kai Siong Yow et al.
  SRCE : arXiv e-prints
  LINK : https://doi.org/10.48550/arXiv.2204.01057
  TYPE : article
  DATE : 2022-04
  DONE : false

ATTENTION IS ALL YOU NEED
  AUTH : Ashish Vaswani et al.
  SRCE : Advances in Neural Information Processing Systems
  LINK : https://proceedings.neurips.cc/paper_files/paper/2017/file/3f5ee243547dee91fbd053c1c4a845aa-Paper.pdf
  TYPE : article
  DATE : 2017
  TAGS : GNNs
  DONE : false

CROSS-SENTENCE N-ARY RELATION EXTRACTION WITH GRAPH LSTMS
  AUTH : Nanyun Peng et al.
  SRCE : arXiv e-prints
  LINK : https://doi.org/10.48550/arXiv.1708.03743
  TYPE : article
  DATE : 2017-08
  TAGS : GNNs
  DONE : false

EFFICIENT RECOGNITION OF EQUIMATCHABLE GRAPHS
  AUTH : Marc Demange et al.
  SRCE : Information Processing Letters
  LINK : https://www.sciencedirect.com/science/article/pii/S0020019013002184
  TYPE : article
  DATE : 2014
  NOTE : In this paper, we give a new characterization of equimatchable graphs that are graphs with all maximal matchings having the same size. This gives an O(n2m)-algorithm for deciding whether a general graph of order n and with m edges is equimatchable. An O(n4.5) recognition algorithm based on the Gallai–Edmonds Decomposition already follows from Lesk et al. (1984) [8]. Our characterization and algorithm use only some basic knowledge on matchings and can be formulated in a simplier way. Moreover it leads to a better time complexity.
  DONE : false
  WIDE : true

GATED GRAPH SEQUENCE NEURAL NETWORKS
  AUTH : Yujia Li et al.
  SRCE : arXiv e-prints
  LINK : https://doi.org/10.48550/arXiv.1511.05493
  TYPE : article
  DATE : 2015-11
  TAGS : GNNs
  DONE : true

GEOMETRIC DEEP LEARNING ON GRAPHS AND MANIFOLDS USING MIXTURE MODEL CNNS
  AUTH : Federico Monti et al.
  SRCE : arXiv e-prints
  LINK : https://doi.org/10.48550/arXiv.1611.08402
  TYPE : article
  DATE : 2016-11
  TAGS : GNNs
  DONE : true

GNNEXPLAINER: GENERATING EXPLANATIONS FOR GRAPH NEURAL NETWORKS
  AUTH : Rex Ying et al.
  SRCE : arXiv e-prints
  LINK : https://doi.org/10.48550/arXiv.1903.03894
  TYPE : article
  DATE : 2019-03
  TAGS : GNNs
  DONE : true

GRAPH ATTENTION NETWORKS
  AUTH : Petar Veli\\vckovic et al.
  SRCE : arXiv e-prints
  LINK : https://doi.org/10.48550/arXiv.1710.10903
  TYPE : article
  DATE : 2017-10
  TAGS : GNNs
  DONE : true

GRAPHICAL-BASED LEARNING ENVIRONMENTS FOR PATTERN RECOGNITION
  AUTH : Franco Scarselli et al.
  LINK : https://doi.org/10.1007/978-3-540-27868-9_4
  TYPE : article
  DATE : 2004-08
  TAGS : The GNN Model
  DONE : false

GRAPH NEURAL NETWORKS: A REVIEW OF METHODS AND APPLICATIONS
  AUTH : Jie Zhou et al.
  SRCE : AI Open
  LINK : https://www.sciencedirect.com/science/article/pii/S2666651021000012
  TYPE : article
  DATE : 2020
  TAGS : GNNs
  NOTE : Lots of learning tasks require dealing with graph data which contains rich relation information among elements. Modeling physics systems, learning molecular fingerprints, predicting protein interface, and classifying diseases demand a model to learn from graph inputs. In other domains such as learning from non-structural data like texts and images, reasoning on extracted structures (like the dependency trees of sentences and the scene graphs of images) is an important research topic which also needs graph reasoning models. Graph neural networks (GNNs) are neural models that capture the dependence of graphs via message passing between the nodes of graphs. In recent years, variants of GNNs such as graph convolutional network (GCN), graph attention network (GAT), graph recurrent network (GRN) have demonstrated ground-breaking performances on many deep learning tasks. In this survey, we propose a general design pipeline for GNN models and discuss the variants of each component, systematically categorize the applications, and propose four open problems for future research.
  DONE : false
  WIDE : true

GRAPH NEURAL NETWORKS FOR RANKING WEB PAGES
  AUTH : Franco Scarselli et al.
  LINK : https://doi.org/10.1109/WI.2005.67
  TYPE : article
  DATE : 2005-01
  TAGS : The GNN Model
  DONE : true

GRAPH NEURAL NETWORKS: FOUNDATIONS, FRONTIERS, AND APPLICATIONS
  AUTH : Lingfei Wu et al.
  SRCE : Springer Singapore
  TYPE : book
  DATE : 2022
  TAGS : GNNs
  DONE : false

HOW POWERFUL ARE GRAPH NEURAL NETWORKS?
  AUTH : Keyulu Xu et al.
  SRCE : arXiv e-prints
  LINK : https://doi.org/10.48550/arXiv.1810.00826
  TYPE : article
  DATE : 2018-10
  TAGS : GNNs
  DONE : true

IMPROVED SEMANTIC REPRESENTATIONS FROM TREE-STRUCTURED LONG SHORT-TERM MEMORY NETWORKS
  AUTH : Kai Sheng Tai et al.
  SRCE : arXiv e-prints
  LINK : https://doi.org/10.48550/arXiv.1503.00075
  TYPE : article
  DATE : 2015-02
  DONE : false

INDUCTIVE REPRESENTATION LEARNING ON LARGE GRAPHS
  AUTH : William L. Hamilton et al.
  SRCE : arXiv e-prints
  LINK : https://doi.org/10.48550/arXiv.1706.02216
  TYPE : article
  DATE : 2017-06
  TAGS : GNNs
  DONE : true

MINING EVOLUTIONS OF COMPLEX SPATIAL OBJECTS USING A SINGLE-ATTRIBUTED DIRECTED ACYCLIC GRAPH
  AUTH : Frédéric Flouvat et al.
  SRCE : Knowledge and Information Systems (KAIS)
  LINK : https://hal.science/hal-02909702
  TYPE : article
  DATE : 2020-06
  TAGS : Pattern Mining
  DONE : false

NEURAL RELATIONAL INFERENCE FOR INTERACTING SYSTEMS
  AUTH : Thomas Kipf et al.
  SRCE : arXiv e-prints
  LINK : https://doi.org/10.48550/arXiv.1802.04687
  TYPE : article
  DATE : 2018-02
  DONE : true

ON EXPLAINABILITY OF GRAPH NEURAL NETWORKS VIA SUBGRAPH EXPLORATIONS
  AUTH : Hao Yuan et al.
  SRCE : arXiv e-prints
  LINK : https://doi.org/10.48550/arXiv.2102.05152
  TYPE : article
  DATE : 2021-02
  TAGS : GNNs
  DONE : false

ON THE EXPRESSIVE POWER OF GNN DERIVATIVES
  AUTH : Yam Eitan et al.
  SRCE : arXiv e-prints
  LINK : https://doi.org/10.48550/arXiv.2510.02565
  TYPE : article
  DATE : 2025-10
  TAGS : GNNs
  DONE : true

RECHERCHE DE MOTIFS SPATIO-TEMPORELS FRÉQUENTS DANS UN GRAND GRAPHE VIA DES MÉTHODES D'APPRENTISSAGE PROFOND
  AUTH : Assaad Oussama Zeghina
  SRCE : Université de Strasbourg
  LINK : https://theses.hal.science/tel-04985098
  TYPE : book
  DATE : 2024-12
  DONE : false

SEMI-SUPERVISED USER GEOLOCATION VIA GRAPH CONVOLUTIONAL NETWORKS
  AUTH : Afshin Rahimi et al.
  SRCE : arXiv e-prints
  LINK : https://doi.org/10.48550/arXiv.1804.08049
  TYPE : article
  DATE : 2018-04
  TAGS : GNNs
  DONE : true

SPAPOOL: SOFT PARTITION ASSIGNMENT POOLING FOR__GRAPH NEURAL NETWORKS
  AUTH : Rodrigue Govan et al.
  SRCE : arXiv e-prints
  LINK : https://doi.org/10.48550/arXiv.2509.11675
  TYPE : article
  DATE : 2025-09
  DONE : false

STOCHASTIC TRAINING OF GRAPH CONVOLUTIONAL NETWORKS WITH VARIANCE REDUCTION
  AUTH : Jianfei Chen et al.
  SRCE : arXiv e-prints
  LINK : https://doi.org/10.48550/arXiv.1710.10568
  TYPE : article
  DATE : 2017-10
  TAGS : GNNs
  DONE : false

UNDERSTANDING SPECTRAL GRAPH NEURAL NETWORK
  AUTH : Xinye Chen
  SRCE : arXiv e-prints
  LINK : https://doi.org/10.48550/arXiv.2012.06660
  TYPE : article
  DATE : 2020-12
  TAGS : GNNs
  DONE : false

WEATHERGNN: EXPLOITING METEO- AND SPATIAL-DEPENDENCIES FOR LOCAL NUMERICAL WEATHER PREDICTION BIAS-CORRECTION
  AUTH : Binqing Wu et al.
  SRCE : Proceedings of the Thirty-Third International Joint Conference on Artificial Intelligence, IJCAI-24
  LINK : https://doi.org/10.24963/ijcai.2024/269
  TYPE : article
  DATE : 2024-08
  NOTE : Main Track
  DONE : true

WHEN DO GNNS WORK: UNDERSTANDING AND IMPROVING NEIGHBORHOOD AGGREGATION
  AUTH : Yiqing Xie et al.
  SRCE : Proceedings of the Twenty-Ninth International Joint Conference on Artificial Intelligence, IJCAI-20
  LINK : https://doi.org/10.24963/ijcai.2020/181
  TYPE : article
  DATE : 2020-07
  TAGS : GNNs
  NOTE : Main track
  DONE : false

SEMI-SUPERVISED CLASSIFICATION WITH GRAPH CONVOLUTIONAL NETWORKS
  AUTH : Thomas N. Kipf et al.
  SRCE : arXiv
  LINK : https://doi.org/10.48550/ARXIV.1609.02907
  TYPE : article
  DATE : 2016-09
  TAGS : GNNs
  NOTE : We present a scalable approach for semi-supervised learning on graph-structured data that is based on an efficient variant of convolutional neural networks which operate directly on graphs. We motivate the choice of our convolutional architecture via a localized first-order approximation of spectral graph convolutions. Our model scales linearly in the number of graph edges and learns hidden layer representations that encode both local graph structure and features of nodes. In a number of experiments on citation networks and on a knowledge graph dataset we demonstrate that our approach outperforms related methods by a significant margin.
  DONE : false
  WIDE : true

HETEROGENEOUS GRAPH ATTENTION NETWORK
  AUTH : Xiao Wang et al.
  SRCE : WWW 2019
  LINK : https://doi.org/10.48550/ARXIV.1903.07293
  TYPE : article
  DATE : 2019-03
  NOTE : Graph neural network, as a powerful graph representation technique based on deep learning, has shown superior performance and attracted considerable research interest. However, it has not been fully considered in graph neural network for heterogeneous graph which contains different types of nodes and links. The heterogeneity and rich semantic information bring great challenges for designing a graph neural network for heterogeneous graph. Recently, one of the most exciting advancements in deep learning is the attention mechanism, whose great potential has been well demonstrated in various areas. In this paper, we first propose a novel heterogeneous graph neural network based on the hierarchical attention, including node-level and semantic-level attentions. Specifically, the node-level attention aims to learn the importance between a node and its metapath based neighbors, while the semantic-level attention is able to learn the importance of different meta-paths. With the learned importance from both node-level and semantic-level attention, the importance of node and meta-path can be fully considered. Then the proposed model can generate node embedding by aggregating features from meta-path based neighbors in a hierarchical manner. Extensive experimental results on three real-world heterogeneous graphs not only show the superior performance of our proposed model over the state-of-the-arts, but also demonstrate its potentially good interpretability for graph analysis.
  DONE : false
  WIDE : true

THE GRAPH NEURAL NETWORK MODEL
  AUTH : Franco Scarselli et al.
  SRCE : IEEE Transactions on Neural Networks
  LINK : https://doi.org/10.1109/TNN.2008.2005605
  TYPE : article
  DATE : 2009-01
  TAGS : The GNN Model
  DONE : true

THE ANATOMY OF A LARGE-SCALE HYPERTEXTUAL WEB SEARCH ENGINE
  AUTH : Sergey Brin et al.
  SRCE : Computer Networks and ISDN Systems
  LINK : https://doi.org/10.1016/s0169-7552(98)00110-x
  TYPE : article
  DATE : 1998-04
  DONE : false

A GENERAL FRAMEWORK FOR ADAPTIVE PROCESSING OF DATA STRUCTURES
  AUTH : P. Frasconi et al.
  SRCE : IEEE Transactions on Neural Networks
  LINK : https://doi.org/10.1109/72.712151
  TYPE : article
  DATE : 1998
  TAGS : The GNN Model
  DONE : false

COMPUTATIONAL CAPABILITIES OF GRAPH NEURAL NETWORKS
  AUTH : F. Scarselli et al.
  SRCE : IEEE Transactions on Neural Networks
  LINK : https://doi.org/10.1109/tnn.2008.2005141
  TYPE : article
  DATE : 2009-01
  TAGS : The GNN Model
  DONE : false

A DIRECT ADAPTIVE METHOD FOR FASTER BACKPROPAGATION LEARNING: THE RPROP ALGORITHM
  AUTH : M. Riedmiller et al.
  SRCE : IEEE International Conference on Neural Networks
  LINK : https://doi.org/10.1109/icnn.1993.298623
  TYPE : article
  TAGS : Neural Networks
  DONE : false

NON-NEGATIVE MATRICES AND MARKOV CHAINS
  AUTH : E. Seneta
  SRCE : Springer Series in Statistics
  LINK : https://doi.org/10.1007/0-387-32792-4
  TYPE : book
  DATE : 1981
  TAGS : Math
  DONE : false

A NEW MODEL FOR LEARNING IN GRAPH DOMAINS
  AUTH : M. Gori et al.
  SRCE : Proceedings. 2005 IEEE International Joint Conference on Neural Networks, 2005.
  LINK : https://doi.org/10.1109/ijcnn.2005.1555942
  TYPE : article
  DATE : 2005
  TAGS : The GNN Model
  DONE : true

GENERALIZATION OF BACK-PROPAGATION TO RECURRENT NEURAL NETWORKS
  AUTH : Fernando J. Pineda
  SRCE : Physical Review Letters
  LINK : https://doi.org/10.1103/physrevlett.59.2229
  TYPE : article
  DATE : 1987-11
  TAGS : , Neural Networks
  DONE : false

THE PATTERN NEXT DOOR: TOWARDS SPATIO-SEQUENTIAL PATTERN DISCOVERY
  AUTH : Hugo Alatrista Salas et al.
  SRCE : Advances in Knowledge Discovery and Data Mining
  LINK : https://doi.org/10.1007/978-3-642-30220-6_14
  TYPE : book
  DATE : 2012
  TAGS : Pattern Mining
  DONE : true

A GRAPH-BASED APPROACH TO SPATIOTEMPORAL EVENT SEQUENCE MINING
  AUTH : Berkay Aydin et al.
  SRCE : 2016 IEEE 16th International Conference on Data Mining Workshops (ICDMW)
  LINK : https://doi.org/10.1109/icdmw.2016.0157
  TYPE : article
  DATE : 2016-12
  DONE : false

MIXED-DROVE SPATIOTEMPORAL CO-OCCURRENCE PATTERN MINING
  AUTH : M. Celik et al.
  SRCE : IEEE Transactions on Knowledge and Data Engineering
  LINK : https://doi.org/10.1109/tkde.2008.97
  TYPE : article
  DATE : 2008-10
  TAGS : Pattern Mining
  DONE : false

EFFICIENT MINING OF SPATIOTEMPORAL PATTERNS
  AUTH : IIias Tsoukatos et al.
  SRCE : Advances in Spatial and Temporal Databases
  LINK : https://doi.org/10.1007/3-540-47724-1_22
  TYPE : book
  DATE : 2001
  TAGS : Pattern Mining
  DONE : true

FLOWMINER: FINDING FLOW PATTERNS IN SPATIO-TEMPORAL DATABASES
  AUTH : J. Wang et al.
  SRCE : 16th IEEE International Conference on Tools with Artificial Intelligence
  LINK : https://doi.org/10.1109/ictai.2004.63
  TYPE : article
  DATE : 2004
  TAGS : Pattern Mining
  DONE : false

A GENERALIZED FRAMEWORK FOR MINING SPATIO-TEMPORAL PATTERNS IN SCIENTIFIC DATA
  AUTH : Hui Yang et al.
  SRCE : Proceedings of the eleventh ACM SIGKDD international conference on Knowledge discovery in data mining
  LINK : https://doi.org/10.1145/1081870.1081962
  TYPE : article
  DATE : 2005-08
  TAGS : Pattern Mining
  DONE : false

GSPAN: GRAPH-BASED SUBSTRUCTURE PATTERN MINING
  AUTH : Xifeng Yan et al.
  SRCE : 2002 IEEE International Conference on Data Mining, 2002. Proceedings.
  LINK : https://doi.org/10.1109/icdm.2002.1184038
  TYPE : article
  TAGS : Pattern Mining
  DONE : true

MINING COHESIVE PATTERNS FROM GRAPHS WITH FEATURE VECTORS
  AUTH : Flavia Moser et al.
  SRCE : Proceedings of the 2009 SIAM International Conference on Data Mining
  LINK : https://doi.org/10.1137/1.9781611972795.51
  TYPE : article
  DATE : 2009-04
  TAGS : Pattern Mining
  DONE : true

CASCADING SPATIO-TEMPORAL PATTERN DISCOVERY
  AUTH : Pradeep Mohan et al.
  SRCE : IEEE Transactions on Knowledge and Data Engineering
  LINK : https://doi.org/10.1109/tkde.2011.146
  TYPE : article
  DATE : 2012-11
  TAGS : Pattern Mining
  DONE : true

INTRODUCTION TO LATTICES AND ORDER
  AUTH : B. A. Davey et al.
  SRCE : Cambridge Univ. Press
  TYPE : book
  DATE : 2010
  NOTE : Literaturverz. S. [280] - 285
  DONE : false

ADVANCES IN SPATIAL AND TEMPORAL DATABASES
  AUTH : Christian S Jensen et al.
  SRCE : Springer-Verlag Berlin Heidelberg
  TYPE : book
  DATE : 2006
  NOTE : In: Springer-Online
  DONE : false

PREFIXSPAN,: MINING SEQUENTIAL PATTERNS EFFICIENTLY BY PREFIX-PROJECTED PATTERN GROWTH
  AUTH : Jian Pei et al.
  SRCE : Proceedings 17th International Conference on Data Engineering
  LINK : https://doi.org/10.1109/icde.2001.914830
  TYPE : article
  TAGS : Pattern Mining
  DONE : false

ATTENTION BASED SPATIAL-TEMPORAL GRAPH CONVOLUTIONAL NETWORKS FOR TRAFFIC FLOW FORECASTING
  AUTH : Shengnan Guo et al.
  SRCE : Proceedings of the AAAI Conference on Artificial Intelligence
  LINK : https://github.com/wanhuaiyu/ASTGCN
  TYPE : article
  DATE : 2019-07
  DONE : false

DIFFUSION CONVOLUTIONAL RECURRENT NEURAL NETWORK: DATA-DRIVEN TRAFFIC FORECASTING
  AUTH : Yaguang Li et al.
  SRCE : arXiv
  LINK : https://doi.org/10.48550/ARXIV.1707.01926
  TYPE : article
  DATE : 2017-07
  TAGS : Traffic
  NOTE : Spatiotemporal forecasting has various applications in neuroscience, climate and transportation domain. Traffic forecasting is one canonical example of such learning task. The task is challenging due to (1) complex spatial dependency on road networks, (2) non-linear temporal dynamics with changing road conditions and (3) inherent difficulty of long-term forecasting. To address these challenges, we propose to model the traffic flow as a diffusion process on a directed graph and introduce Diffusion Convolutional Recurrent Neural Network (DCRNN), a deep learning framework for traffic forecasting that incorporates both spatial and temporal dependency in the traffic flow. Specifically, DCRNN captures the spatial dependency using bidirectional random walks on the graph, and the temporal dependency using the encoder-decoder architecture with scheduled sampling. We evaluate the framework on two real-world large scale road network traffic datasets and observe consistent improvement of 12% - 15% over state-of-the-art baselines.
  DONE : true
  WIDE : true

AST-GCN: ATTRIBUTE-AUGMENTED SPATIOTEMPORAL GRAPH CONVOLUTIONAL NETWORK FOR TRAFFIC FORECASTING
  AUTH : Jiawei Zhu et al.
  SRCE : arXiv
  LINK : https://doi.org/10.48550/ARXIV.2011.11004
  TYPE : article
  DATE : 2020-11
  TAGS : Traffic
  NOTE : Traffic forecasting is a fundamental and challenging task in the field of intelligent transportation. Accurate forecasting not only depends on the historical traffic flow information but also needs to consider the influence of a variety of external factors, such as weather conditions and surrounding POI distribution. Recently, spatiotemporal models integrating graph convolutional networks and recurrent neural networks have become traffic forecasting research hotspots and have made significant progress. However, few works integrate external factors. Therefore, based on the assumption that introducing external factors can enhance the spatiotemporal accuracy in predicting traffic and improving interpretability, we propose an attribute-augmented spatiotemporal graph convolutional network (AST-GCN). We model the external factors as dynamic attributes and static attributes and design an attribute-augmented unit to encode and integrate those factors into the spatiotemporal graph convolution model. Experiments on real datasets show the effectiveness of considering external information on traffic forecasting tasks when compared to traditional traffic prediction methods. Moreover, under different attribute-augmented schemes and prediction horizon settings, the forecasting accuracy of the AST-GCN is higher than that of the baselines.
  DONE : false
  WIDE : true

SPATIO-TEMPORAL GRAPH CONVOLUTIONAL NETWORKS: A DEEP LEARNING FRAMEWORK FOR TRAFFIC FORECASTING
  AUTH : Bing Yu et al.
  SRCE : Proceedings of the Twenty-Seventh International Joint Conference on Artificial Intelligence
  LINK : https://doi.org/10.24963/ijcai.2018/505
  TYPE : article
  DATE : 2017-07
  TAGS : Traffic
  NOTE : Timely accurate traffic forecast is crucial for urban traffic control and guidance. Due to the high nonlinearity and complexity of traffic flow, traditional methods cannot satisfy the requirements of mid-and-long term prediction tasks and often neglect spatial and temporal dependencies. In this paper, we propose a novel deep learning framework, Spatio-Temporal Graph Convolutional Networks (STGCN), to tackle the time series prediction problem in traffic domain. Instead of applying regular convolutional and recurrent units, we formulate the problem on graphs and build the model with complete convolutional structures, which enable much faster training speed with fewer parameters. Experiments show that our model STGCN effectively captures comprehensive spatio-temporal correlations through modeling multi-scale traffic networks and consistently outperforms state-of-the-art baselines on various real-world traffic datasets.
  DONE : true
  WIDE : true

CONNECTING THE DOTS: MULTIVARIATE TIME SERIES FORECASTING WITH GRAPH NEURAL NETWORKS
  AUTH : Zonghan Wu et al.
  SRCE : arXiv
  LINK : https://doi.org/10.48550/ARXIV.2005.11650
  TYPE : article
  DATE : 2020-05
  TAGS : Traffic
  NOTE : Modeling multivariate time series has long been a subject that has attracted researchers from a diverse range of fields including economics, finance, and traffic. A basic assumption behind multivariate time series forecasting is that its variables depend on one another but, upon looking closely, it is fair to say that existing methods fail to fully exploit latent spatial dependencies between pairs of variables. In recent years, meanwhile, graph neural networks (GNNs) have shown high capability in handling relational dependencies. GNNs require well-defined graph structures for information propagation which means they cannot be applied directly for multivariate time series where the dependencies are not known in advance. In this paper, we propose a general graph neural network framework designed specifically for multivariate time series data. Our approach automatically extracts the uni-directed relations among variables through a graph learning module, into which external knowledge like variable attributes can be easily integrated. A novel mix-hop propagation layer and a dilated inception layer are further proposed to capture the spatial and temporal dependencies within the time series. The graph learning, graph convolution, and temporal convolution modules are jointly learned in an end-to-end framework. Experimental results show that our proposed model outperforms the state-of-the-art baseline methods on 3 of 4 benchmark datasets and achieves on-par performance with other approaches on two traffic datasets which provide extra structural information.
  DONE : false
  WIDE : true

ATTENTION-BASED SPATIAL-TEMPORAL GRAPH CONVOLUTIONAL RECURRENT NETWORKS FOR TRAFFIC FORECASTING
  AUTH : Haiyang Liu et al.
  SRCE : arXiv
  LINK : https://doi.org/10.48550/ARXIV.2302.12973
  TYPE : article
  DATE : 2023-02
  TAGS : Traffic
  NOTE : Traffic forecasting is one of the most fundamental problems in transportation science and artificial intelligence. The key challenge is to effectively model complex spatial-temporal dependencies and correlations in modern traffic data. Existing methods, however, cannot accurately model both long-term and short-term temporal correlations simultaneously, limiting their expressive power on complex spatial-temporal patterns. In this paper, we propose a novel spatial-temporal neural network framework: Attention-based Spatial-Temporal Graph Convolutional Recurrent Network (ASTGCRN), which consists of a graph convolutional recurrent module (GCRN) and a global attention module. In particular, GCRN integrates gated recurrent units and adaptive graph convolutional networks for dynamically learning graph structures and capturing spatial dependencies and local temporal relationships. To effectively extract global temporal dependencies, we design a temporal attention layer and implement it as three independent modules based on multi-head self-attention, transformer, and informer respectively. Extensive experiments on five real traffic datasets have demonstrated the excellent predictive performance of all our three models with all their average MAE, RMSE and MAPE across the test datasets lower than the baseline methods.
  DONE : false
  WIDE : true

`