Identifying Contemporaneous and Lagged Dependence Structures by Promoting Sparsity in Continuous-time Neural Networks
Continuous-time dynamics models, e.g., neural ordinary differential equations, enable accurate modeling of underlying dynamics in time-series data. However, employing neural networks for parameterizing dynamics makes it challenging for humans to identify dependence structures, especially in the presence of delayed effects. In consequence, these models are not an attractive option when capturing dependence carries more importance than accurate modeling, e.g., in tsunami forecasting.
Authors
- Donsub Rim (ORCID: https://orcid.org/0000-0002-6721-2070)
- Kookjin Lee (ORCID: https://orcid.org/0000-0003-1557-5862)
- F.‐J. WU (ORCID: https://orcid.org/0009-0004-1131-2816)
- Noseong Park (ORCID: https://orcid.org/0000-0002-1268-840X)
- Sanghyun Hong (ORCID: https://orcid.org/0000-0003-4154-7611)
- Woojin Cho (ORCID: https://orcid.org/0000-0002-9413-7029)
- David Korotky
Institutions
- Oregon State University (US)
- Korea Advanced Institute of Science and Technology (KR)
- Yonsei University (KR)
- Washington University in St. Louis (US)
- Arizona State University (US)
Publication Details
- Published
- 2024-10-20
- DOI
- https://doi.org/10.1145/3627673.3679751
- Primary Topic
- Neural Networks and Applications
- Type
- article
- Field-Weighted Citation Impact
- 0.00