Cumulative Difference Learning VAE for Time-Series with Temporally Correlated Inflow-Outflow
Tianchun Li, Chengxiang Wu, Pengyi Shi, Xiaoqian Wang
Abstract
Time-series generation has crucial practical significance for decision-making under uncertainty. Existing methods have various limitations like accumulating errors over time, significantly impacting downstream tasks. We develop a novel generation method, DT-VAE, that incorporates generalizable domain knowledge, is mathematically justified, and significantly outperforms existing methods by mitigating error accumulation through a cumulative difference learning mechanism. We evaluate the performance of DT-VAE on several downstream tasks using both semi-synthetic and real time-series datasets, including benchmark datasets and our newly curated COVID-19 hospitalization datasets. The COVID-19 datasets enrich existing resources for time-series analysis. Additionally, we introduce Diverse Trend Preserving (DTP), a time-series clustering-based evaluation for direct and interpretable assessments of generated samples, serving as a valuable tool for evaluating time-series generative models.
BibTeX
@article{Li_Wu_Shi_Wang_2024, title={Cumulative Difference Learning VAE for Time-Series with Temporally Correlated Inflow-Outflow}, volume={38}, url={https://ojs.aaai.org/index.php/AAAI/article/view/29266}, DOI={10.1609/aaai.v38i12.29266}, abstractNote={Time-series generation has crucial practical significance for decision-making under uncertainty. Existing methods have various limitations like accumulating errors over time, significantly impacting downstream tasks. We develop a novel generation method, DT-VAE, that incorporates generalizable domain knowledge, is mathematically justified, and significantly outperforms existing methods by mitigating error accumulation through a cumulative difference learning mechanism. We evaluate the performance of DT-VAE on several downstream tasks using both semi-synthetic and real time-series datasets, including benchmark datasets and our newly curated COVID-19 hospitalization datasets. The COVID-19 datasets enrich existing resources for time-series analysis. Additionally, we introduce Diverse Trend Preserving (DTP), a time-series clustering-based evaluation for direct and interpretable assessments of generated samples, serving as a valuable tool for evaluating time-series generative models.}, number={12}, journal={Proceedings of the AAAI Conference on Artificial Intelligence}, author={Li, Tianchun and Wu, Chengxiang and Shi, Pengyi and Wang, Xiaoqian}, year={2024}, month={Mar.}, pages={13619-13627} }