Revealing hidden dynamics from time-series data by ODENet

Hu, Pipi; Yang, Wuyue; Zhu, Yan; Liu, Hong

doi:10.48550/arxiv.2005.04849

Cited by 2 publications

(2 citation statements)

References 12 publications

(17 reference statements)

Supporting

Mentioning

Contrasting

Order By: Relevance

“…For example, many climate models based on mathematical equations that describe the physical processes have been built to predict climate variables while the prediction abilities may vary dramatically across other regions and time [4]. Although physics-informed ML has emerged to build hybrid models for robust predictions recently [18]- [22], these methods neither consider the homogeneity and heterogeneity simultaneously nor the data limitation.…”

Section: Introductionmentioning

confidence: 99%

Physics-Coupled Spatio-Temporal Active Learning for Dynamical Systems

et al. 2022

View full text Add to dashboard Cite

Spatio-temporal forecasting is of great importance in a wide range of dynamical systems applications, such as earth science, transport planning, etc. These applications rely on accurate predictions of spatio-temporal structured data reflecting real-world phenomena. A stunning characteristic is that the dynamical system is not only driven by some physics laws but also impacted by the localized factor in spatial and temporal regions. One of the major challenges is to infer the underlying causes, which generate the perceived data stream and propagate the involved causal dynamics through the distributed observing units. Another challenge is that the success of machine learning based predictive models requires massive annotated data for model training. However, the acquisition of high-quality annotated data is objectively manual and tedious as it needs a considerable amount of human intervention, making it infeasible in fields that require high levels of expertise. To tackle these challenges, we advocate a spatio-temporal physicscoupled neural networks (ST-PCNN) model to learn the underlying physics of the dynamical system and further couple the learned physics to assist the learning of the recurring dynamics. To deal with data-acquisition constraints, an active learning mechanism with Kriging for actively acquiring the most informative data is proposed for ST-PCNN training in a partially observable environment. Our experiments on both synthetic and real-world datasets exhibit that the proposed ST-PCNN with active learning converges to near optimal accuracy with substantially fewer instances.

show abstract

Section: Introductionmentioning

confidence: 99%

Physics-Coupled Spatio-Temporal Active Learning for Dynamical Systems

et al. 2022

View full text Add to dashboard Cite

show abstract

“…For example, many climate models based on mathematical equations that describe the physical processes have been built to predict climate variables while the prediction abilities may vary dramatically across other regions and time [21]. Although physics-informed ML has emerged to build hybrid models for robust predictions recently [5,11,22,31,41], these methods neither consider the homogeneity and heterogeneity simultaneously nor the data limitation.…”

Section: Introductionmentioning

confidence: 99%

Physics-Coupled Spatio-Temporal Active Learning for Dynamical Systems

Huang¹,

Tang²,

Zhu³

et al. 2021

Preprint

View full text Add to dashboard Cite

Spatio-temporal forecasting is of great importance in a wide range of dynamical systems applications from atmospheric science, to recent COVID-19 spread modeling. These applications rely on accurate predictions of spatio-temporal structured data reflecting realworld phenomena. A stunning characteristic is that the dynamical system is not only driven by some physics laws but also impacted by the localized factor in spatial and temporal regions. One of the major challenges is to infer the underlying causes, which generate the perceived data stream and propagate the involved causal dynamics through the distributed observing units. Another challenge is that the success of machine learning based predictive models requires massive annotated data for model training. However, the acquisition of high-quality annotated data is objectively manual and tedious as it needs a considerable amount of human intervention, making it infeasible in fields that require high levels of expertise. To tackle these challenges, we advocate a spatio-temporal physicscoupled neural networks (ST-PCNN) model to learn the underlying physics of the dynamical system and further couple the learned physics to assist the learning of the recurring dynamics. To deal with data-acquisition constraints, an active learning mechanism with Kriging for actively acquiring the most informative data is proposed for ST-PCNN training in a partially observable environment. Our experiments on both synthetic and real-world datasets exhibit that the proposed ST-PCNN with active learning converges to near optimal accuracy with substantially fewer instances. CCS CONCEPTS• Applied computing → Environmental sciences; • Theory of computation → Active learning.

show abstract

Revealing hidden dynamics from time-series data by ODENet

Cited by 2 publications

References 12 publications

Physics-Coupled Spatio-Temporal Active Learning for Dynamical Systems

Physics-Coupled Spatio-Temporal Active Learning for Dynamical Systems

Physics-Coupled Spatio-Temporal Active Learning for Dynamical Systems

Contact Info

Product

Resources

About