Unified Flow Matching for Long Horizon Event Forecasting
This addresses the problem of error accumulation and inefficiency in long-range event forecasting for applications like healthcare, finance, and user behavior modeling, representing a novel approach rather than an incremental improvement.
The paper tackles the challenge of modeling long horizon marked event sequences by proposing a unified flow matching framework that enables non-autoregressive, joint modeling of inter-event times and event types. The method demonstrates significant improvements in accuracy and generation efficiency over autoregressive and diffusion-based baselines on six real-world benchmarks.
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.