Saved in:
Bibliographic Details
Main Author: Shou, Xiao
Format: Preprint
Published: 2025
Subjects:
Online Access:https://arxiv.org/abs/2508.04843
Tags: Add Tag
No Tags, Be the first to tag this record!
_version_ 1866911095293214720
author Shou, Xiao
author_facet Shou, Xiao
contents Modeling long horizon marked event sequences is a fundamental challenge in many real-world applications, including healthcare, finance, and user behavior modeling. Existing neural temporal point process models are typically autoregressive, predicting the next event one step at a time, which limits their efficiency and leads to error accumulation in long-range forecasting. In this work, we propose a unified flow matching framework for marked temporal point processes that enables non-autoregressive, joint modeling of inter-event times and event types, via continuous and discrete flow matching. By learning continuous-time flows for both components, our method generates coherent long horizon event trajectories without sequential decoding. We evaluate our model on six real-world benchmarks and demonstrate significant improvements over autoregressive and diffusion-based baselines in both accuracy and generation efficiency.
format Preprint
id arxiv_https___arxiv_org_abs_2508_04843
institution arXiv
publishDate 2025
record_format arxiv
spellingShingle Unified Flow Matching for Long Horizon Event Forecasting
Shou, Xiao
Machine Learning
Modeling long horizon marked event sequences is a fundamental challenge in many real-world applications, including healthcare, finance, and user behavior modeling. Existing neural temporal point process models are typically autoregressive, predicting the next event one step at a time, which limits their efficiency and leads to error accumulation in long-range forecasting. In this work, we propose a unified flow matching framework for marked temporal point processes that enables non-autoregressive, joint modeling of inter-event times and event types, via continuous and discrete flow matching. By learning continuous-time flows for both components, our method generates coherent long horizon event trajectories without sequential decoding. We evaluate our model on six real-world benchmarks and demonstrate significant improvements over autoregressive and diffusion-based baselines in both accuracy and generation efficiency.
title Unified Flow Matching for Long Horizon Event Forecasting
topic Machine Learning
url https://arxiv.org/abs/2508.04843