Synergetic Learning of Heterogeneous Temporal Sequences for Multi-Horizon Probabilistic Forecasting

Authors

  • Longyuan Li MoE Key Lab of Artificial Intelligence, AI Institute, Shanghai Jiao Tong University State Key Lab of Advanced Optical Communication System and Network, Shanghai Jiao Tong University
  • Jihai Zhang Department of Computer Science and Engineering, Shanghai Jiao Tong University
  • Junchi Yan MoE Key Lab of Artificial Intelligence, AI Institute, Shanghai Jiao Tong University Department of Computer Science and Engineering, Shanghai Jiao Tong University
  • Yaohui Jin MoE Key Lab of Artificial Intelligence, AI Institute, Shanghai Jiao Tong University State Key Lab of Advanced Optical Communication System and Network, Shanghai Jiao Tong University
  • Yunhao Zhang Department of Computer Science and Engineering, Shanghai Jiao Tong University
  • Yanjie Duan Huawei Noah’s Ark Lab
  • Guangjian Tian Huawei Noah’s Ark Lab

Keywords:

Time-Series/Data Streams, Neural Generative Models & Autoencoders, Mining of Spatial, Temporal or Spatio-Temporal Da

Abstract

Time-series is ubiquitous across applications, such as transportation, finance and healthcare. Time-series is often influenced by external factors, especially in the form of asynchronous events, making forecasting difficult. However, existing models are mainly designated for either synchronous time-series or asynchronous event sequence, and can hardly provide a synthetic way to capture the relation between them. We propose Variational Synergetic Multi-Horizon Network (VSMHN), a novel deep conditional generative model. To learn complex correlations across heterogeneous sequences, a tailored encoder is devised to combine the advances in deep point processes models and variational recurrent neural networks. In addition, an aligned time coding and an auxiliary transition scheme are carefully devised for batched training on unaligned sequences. Our model can be trained effectively using stochastic variational inference and generates probabilistic predictions with Monte-Carlo simulation. Furthermore, our model produces accurate, sharp and more realistic probabilistic forecasts. We also show that modeling asynchronous event sequences is crucial for multi-horizon time-series forecasting.

Downloads

Published

2021-05-18

How to Cite

Li, L., Zhang, J., Yan, J., Jin, Y., Zhang, Y., Duan, Y., & Tian, G. (2021). Synergetic Learning of Heterogeneous Temporal Sequences for Multi-Horizon Probabilistic Forecasting. Proceedings of the AAAI Conference on Artificial Intelligence, 35(10), 8420-8428. Retrieved from https://ojs.aaai.org/index.php/AAAI/article/view/17023

Issue

Section

AAAI Technical Track on Machine Learning III