English

Enhancing Asynchronous Time Series Forecasting with Contrastive Relational Inference

Machine Learning 2023-10-10 v2

Abstract

Asynchronous time series, also known as temporal event sequences, are the basis of many applications throughout different industries. Temporal point processes(TPPs) are the standard method for modeling such data. Existing TPP models have focused on parameterizing the conditional distribution of future events instead of explicitly modeling event interactions, imposing challenges for event predictions. In this paper, we propose a novel approach that leverages Neural Relational Inference (NRI) to learn a relation graph that infers interactions while simultaneously learning the dynamics patterns from observational data. Our approach, the Contrastive Relational Inference-based Hawkes Process (CRIHP), reasons about event interactions under a variational inference framework. It utilizes intensity-based learning to search for prototype paths to contrast relationship constraints. Extensive experiments on three real-world datasets demonstrate the effectiveness of our model in capturing event interactions for event sequence modeling tasks. Code will be integrated into the EasyTPP framework.

Keywords

Cite

@article{arxiv.2309.02868,
  title  = {Enhancing Asynchronous Time Series Forecasting with Contrastive Relational Inference},
  author = {Yan Wang and Zhixuan Chu and Tao Zhou and Caigao Jiang and Hongyan Hao and Minjie Zhu and Xindong Cai and Qing Cui and Longfei Li and James Y Zhang and Siqiao Xue and Jun Zhou},
  journal= {arXiv preprint arXiv:2309.02868},
  year   = {2023}
}

Comments

ICDM 2023 AI4TS Workshop

R2 v1 2026-06-28T12:14:04.934Z