Sao, A.; Gottschalk, S.; Tempelmeier, N.; Demidova, E.: MetaCitta: Deep Meta-Learning for Spatio-Temporal Prediction Across Cities and Tasks. In: Kashima, Hisashi; Ide, Tsuyoshi; Peng, Wen-Chih (Eds.): Advances in knowledge discovery and data mining : Part 4. Berlin ; Heidelberg : Springer, 2023 (Lecture Notes in Computer Science (LNCS) ; 13938), S. 70-82. DOI: https://doi.org/10.1007/978-3-031-33383-5_6
Abstract: | |
Accurate spatio-temporal prediction is essential for capturing city dynamics and planning mobility services. State-of-the-art deep spatio-temporal predictive models depend on rich and representative training data for target regions and tasks. However, the availability of such data is typically limited. Furthermore, existing predictive models fail to utilize cross-correlations across tasks and cities. In this paper, we propose MetaCitta, a novel deep meta-learning approach that addresses the critical challenges of data scarcity and model generalization. MetaCitta adopts the data from different cities and tasks in a generalizable spatio-temporal deep neural network. We propose a novel meta-learning algorithm that minimizes the discrepancy between spatio-temporal representations across tasks and cities. Our experiments with real-world data demonstrate that the proposed MetaCitta approach outperforms state-of-the-art prediction methods for zero-shot learning and pre-training plus fine-tuning. Furthermore, MetaCitta is computationally more efficient than the existing meta-learning approaches. | |
License of this version: | CC BY 4.0 Unported |
Document Type: | BookPart |
Publishing status: | publishedVersion |
Issue Date: | 2023 |
Appears in Collections: | Forschungszentren |
pos. | country | downloads | ||
---|---|---|---|---|
total | perc. | |||
1 | United States | 2 | 28.57% | |
2 | Germany | 2 | 28.57% | |
3 | Europe | 1 | 14.29% | |
4 | China | 1 | 14.29% | |
5 | Canada | 1 | 14.29% |
Hinweis
Zur Erhebung der Downloadstatistiken kommen entsprechend dem „COUNTER Code of Practice for e-Resources“ international anerkannte Regeln und Normen zur Anwendung. COUNTER ist eine internationale Non-Profit-Organisation, in der Bibliotheksverbände, Datenbankanbieter und Verlage gemeinsam an Standards zur Erhebung, Speicherung und Verarbeitung von Nutzungsdaten elektronischer Ressourcen arbeiten, welche so Objektivität und Vergleichbarkeit gewährleisten sollen. Es werden hierbei ausschließlich Zugriffe auf die entsprechenden Volltexte ausgewertet, keine Aufrufe der Website an sich.