{"corpus_id":257767087,"paper_sha":"252351936bd6fabf4b6cd2962fa0ee613772278d","doi":"10.1109/TKDE.2023.3333824","arxiv_id":"2303.14483","pmid":null,"pmcid":null,"mag_id":null,"dblp_id":"journals/corr/abs-2303-14483","acl_id":null,"title":"Spatio-Temporal Graph Neural Networks for Predictive Learning in Urban Computing: A Survey","year":2023,"publication_date":"2023-03-25","venue":"IEEE Transactions on Knowledge and Data Engineering","journal":{"name":"IEEE Transactions on Knowledge and Data Engineering","pages":"5388-5408","volume":"36"},"journal_issn":null,"journal_title":null,"publication_types":["JournalArticle","Review"],"pubmed_pub_types":null,"s2_fields_of_study":["Computer Science","Engineering","Environmental Science"],"reference_count":274,"citation_count":414,"influential_citation_count":12,"is_open_access":true,"arxiv_categories":["cs.LG"],"arxiv_license":"http://creativecommons.org/licenses/by/4.0/","arxiv_journal_ref":null,"mesh_headings":null,"chemicals":null,"comments_corrections":null,"source_flags":1,"s2_open_access_pdf_url":"https://arxiv.org/pdf/2303.14483","s2_open_access_landing_url":"https://www.semanticscholar.org/paper/252351936bd6fabf4b6cd2962fa0ee613772278d","s2_open_access_license":null,"s2_open_access_status":"GREEN","pmc_open_access_pdf_url":null,"pmc_open_access_landing_url":null,"pmc_open_access_license":null,"pmc_open_access_status":null,"unpaywall_open_access_pdf_url":null,"unpaywall_open_access_landing_url":null,"unpaywall_open_access_license":null,"unpaywall_open_access_status":null,"abstract":"With recent advances in sensing technologies, a myriad of spatio-temporal data has been generated and recorded in smart cities. Forecasting the evolution patterns of spatio-temporal data is an important yet demanding aspect of urban computing, which can enhance intelligent management decisions in various fields, including transportation, environment, climate, public safety, healthcare, and others. Traditional statistical and deep learning methods struggle to capture complex correlations in urban spatio-temporal data. To this end, Spatio-Temporal Graph Neural Networks (STGNN) have been proposed, achieving great promise in recent years. STGNNs enable the extraction of complex spatio-temporal dependencies by integrating graph neural networks (GNNs) and various temporal learning methods. In this manuscript, we provide a comprehensive survey on recent progress on STGNN technologies for predictive learning in urban computing. Firstly, we provide a brief introduction to the construction methods of spatio-temporal graph data and the prevalent deep-learning architectures used in STGNNs. We then sort out the primary application domains and specific predictive learning tasks based on existing literature. Afterward, we scrutinize the design of STGNNs and their combination with some advanced technologies in recent years. Finally, we conclude the limitations of existing research and suggest potential directions for future work.","claims":[{"public_id":"cl_561a71cf236754cf79f1e18a5b1559bc","status":"active","text":"Spatio-temporal graph neural networks are a promising approach for extracting complex spatio-temporal dependencies in urban predictive learning.","confidence":0.95,"contributors":[{"id":1,"public_id":"12632b8b5f","public_label":"Anonymous (12632b8b5f)","roles":["extraction"],"url":"https://sah.borca.ai/u/12632b8b5f"}],"url":"https://sah.borca.ai/claims/cl_561a71cf236754cf79f1e18a5b1559bc"},{"public_id":"cl_c4af8e7004f06083aa1d17aff90d4ae6","status":"active","text":"The survey examines STGNN design choices, their combination with advanced technologies, and limitations in existing research.","confidence":0.96,"contributors":[{"id":1,"public_id":"12632b8b5f","public_label":"Anonymous (12632b8b5f)","roles":["extraction"],"url":"https://sah.borca.ai/u/12632b8b5f"}],"url":"https://sah.borca.ai/claims/cl_c4af8e7004f06083aa1d17aff90d4ae6"},{"public_id":"cl_4158618bae706d968fc4e6ecaacf64a3","status":"active","text":"The survey organizes the literature by spatio-temporal graph data construction methods, prevalent deep-learning architectures, application domains, and predictive learning tasks.","confidence":0.97,"contributors":[{"id":1,"public_id":"12632b8b5f","public_label":"Anonymous (12632b8b5f)","roles":["extraction"],"url":"https://sah.borca.ai/u/12632b8b5f"}],"url":"https://sah.borca.ai/claims/cl_4158618bae706d968fc4e6ecaacf64a3"},{"public_id":"cl_4b2c57c09fb107154597ca539f3676fe","status":"active","text":"Traditional statistical and deep learning methods struggle to capture complex correlations in urban spatio-temporal data.","confidence":0.89,"contributors":[{"id":1,"public_id":"12632b8b5f","public_label":"Anonymous (12632b8b5f)","roles":["extraction"],"url":"https://sah.borca.ai/u/12632b8b5f"}],"url":"https://sah.borca.ai/claims/cl_4b2c57c09fb107154597ca539f3676fe"}],"concepts":[{"public_id":"co_01f6b81dc259ebcf0a6ec1a6b222a8ca","status":"active","name":"spatio-temporal graph data","description":"Graph-structured data representing entities and relations that evolve over time.","types":["data"],"aliases":[],"contributors":[{"id":1,"public_id":"12632b8b5f","public_label":"Anonymous (12632b8b5f)","roles":["extraction"],"url":"https://sah.borca.ai/u/12632b8b5f"}],"url":"https://sah.borca.ai/concepts/co_01f6b81dc259ebcf0a6ec1a6b222a8ca"},{"public_id":"co_29268ab55358b9a30ee7bc1b88a77ca2","status":"active","name":"predictive learning tasks","description":"Tasks that aim to forecast future values, states, or patterns from observed data.","types":["task"],"aliases":[],"contributors":[{"id":1,"public_id":"12632b8b5f","public_label":"Anonymous (12632b8b5f)","roles":["extraction"],"url":"https://sah.borca.ai/u/12632b8b5f"}],"url":"https://sah.borca.ai/concepts/co_29268ab55358b9a30ee7bc1b88a77ca2"},{"public_id":"co_422dc9f4211819d8cd0baaadd6f88ec4","status":"active","name":"advanced technologies","description":"Recent technical components combined with STGNNs in the reviewed literature.","types":["technology"],"aliases":[],"contributors":[{"id":1,"public_id":"12632b8b5f","public_label":"Anonymous (12632b8b5f)","roles":["extraction"],"url":"https://sah.borca.ai/u/12632b8b5f"}],"url":"https://sah.borca.ai/concepts/co_422dc9f4211819d8cd0baaadd6f88ec4"},{"public_id":"co_4f57fde893e02e065b355c0f41d7e19b","status":"active","name":"traditional statistical methods","description":"Conventional statistical approaches used as baselines or alternatives for modeling data.","types":["method"],"aliases":[],"contributors":[{"id":1,"public_id":"12632b8b5f","public_label":"Anonymous (12632b8b5f)","roles":["extraction"],"url":"https://sah.borca.ai/u/12632b8b5f"}],"url":"https://sah.borca.ai/concepts/co_4f57fde893e02e065b355c0f41d7e19b"},{"public_id":"co_66dd4449eb82a32737e438f7668fb2ea","status":"active","name":"predictive learning in urban computing","description":"Predictive modeling tasks applied to urban data and smart-city settings.","types":["research area"],"aliases":["urban predictive learning"],"contributors":[{"id":1,"public_id":"12632b8b5f","public_label":"Anonymous (12632b8b5f)","roles":["extraction"],"url":"https://sah.borca.ai/u/12632b8b5f"}],"url":"https://sah.borca.ai/concepts/co_66dd4449eb82a32737e438f7668fb2ea"},{"public_id":"co_7f12ae648c24e9f8a30776a33b72661f","status":"active","name":"spatio-temporal graph neural networks","description":"Graph neural network models that integrate spatial graph structure with temporal learning to model evolving data.","types":["method"],"aliases":["STGNN","STGNNs"],"contributors":[{"id":1,"public_id":"12632b8b5f","public_label":"Anonymous (12632b8b5f)","roles":["extraction"],"url":"https://sah.borca.ai/u/12632b8b5f"}],"url":"https://sah.borca.ai/concepts/co_7f12ae648c24e9f8a30776a33b72661f"},{"public_id":"co_c150c593fe04324f8dd027a61457be4d","status":"active","name":"urban spatio-temporal data","description":"Data collected in smart-city environments that vary across both space and time.","types":["data"],"aliases":[],"contributors":[{"id":1,"public_id":"12632b8b5f","public_label":"Anonymous (12632b8b5f)","roles":["extraction"],"url":"https://sah.borca.ai/u/12632b8b5f"}],"url":"https://sah.borca.ai/concepts/co_c150c593fe04324f8dd027a61457be4d"},{"public_id":"co_cc0451ece4f47e65efdf3f4d19726c7c","status":"active","name":"deep learning methods","description":"Neural network-based approaches for learning from data.","types":["method"],"aliases":[],"contributors":[{"id":1,"public_id":"12632b8b5f","public_label":"Anonymous (12632b8b5f)","roles":["extraction"],"url":"https://sah.borca.ai/u/12632b8b5f"}],"url":"https://sah.borca.ai/concepts/co_cc0451ece4f47e65efdf3f4d19726c7c"},{"public_id":"co_d217b56eb87931ecbb91fa9f57c5909e","status":"active","name":"spatio-temporal dependencies","description":"Relationships that link observations across spatial locations and time points.","types":["phenomenon"],"aliases":[],"contributors":[{"id":1,"public_id":"12632b8b5f","public_label":"Anonymous (12632b8b5f)","roles":["extraction"],"url":"https://sah.borca.ai/u/12632b8b5f"}],"url":"https://sah.borca.ai/concepts/co_d217b56eb87931ecbb91fa9f57c5909e"},{"public_id":"co_e6f1d691c37fac7ef0768644fe269cdc","status":"active","name":"application domains","description":"Areas in which the reviewed methods are applied, such as transportation or healthcare.","types":["category"],"aliases":[],"contributors":[{"id":1,"public_id":"12632b8b5f","public_label":"Anonymous (12632b8b5f)","roles":["extraction"],"url":"https://sah.borca.ai/u/12632b8b5f"}],"url":"https://sah.borca.ai/concepts/co_e6f1d691c37fac7ef0768644fe269cdc"},{"public_id":"co_f2ca21b992de1a885590423018e3437e","status":"active","name":"limitations of existing research","description":"Constraints or open problems identified in the reviewed STGNN literature.","types":["issue"],"aliases":[],"contributors":[{"id":1,"public_id":"12632b8b5f","public_label":"Anonymous (12632b8b5f)","roles":["extraction"],"url":"https://sah.borca.ai/u/12632b8b5f"}],"url":"https://sah.borca.ai/concepts/co_f2ca21b992de1a885590423018e3437e"},{"public_id":"co_f344aa79453206026793f7aa679371e8","status":"active","name":"deep-learning architectures","description":"Neural network architecture designs used to implement learning models.","types":["method"],"aliases":[],"contributors":[{"id":1,"public_id":"12632b8b5f","public_label":"Anonymous (12632b8b5f)","roles":["extraction"],"url":"https://sah.borca.ai/u/12632b8b5f"}],"url":"https://sah.borca.ai/concepts/co_f344aa79453206026793f7aa679371e8"}],"external_ids":{"DOI":"10.1109/TKDE.2023.3333824","ArXiv":"2303.14483","PubMed":null,"PubMedCentral":null,"MAG":null,"DBLP":"journals/corr/abs-2303-14483","ACL":null},"open_access":{"is_open_access":true,"pdf_url":"https://arxiv.org/pdf/2303.14483","landing_url":"https://www.semanticscholar.org/paper/252351936bd6fabf4b6cd2962fa0ee613772278d","source":"semantic_scholar","pdf_url_source":"semantic_scholar_open_access_pdf","license":null,"status":"GREEN","reason":null},"reference_availability":{"status":"available","references_indexed":true,"full_text_available":true,"full_text_source":"arxiv","count_basis":"semantic_scholar_metadata","extraction_status":"not_applicable","reason":null},"source":{"provider":"episteme2","base_corpus":"semantic_scholar_dump","freshness_mode":"unknown","basis":["semantic_scholar_metadata","postgres_metadata"],"limits":["paper metadata is based on indexed upstream scholarly datasets","claims and concepts are available only for extracted papers","absence of claims or concepts means no extracted graph data is available in this response"],"status":"available","degraded":false,"degraded_reasons":[],"diagnostics":{"status":"available","degraded":false,"degraded_reasons":[],"metadata_status":"available","graph_status":"available","abstract_status":"available"},"source_flags":1},"paper_id":631868,"paper_uid":"c6355220-e2e3-4259-a7e1-1f89c13bbd7b","canonical_identity":{"paper_id":631868,"paper_uid":"c6355220-e2e3-4259-a7e1-1f89c13bbd7b","identity_status":"available","lookup_basis":"semantic_scholar_external_id","compatibility_path":"corpus_id"},"url":"https://sah.borca.ai/papers/257767087"}