Beyond Hawkes: Neural Multi-event Forecasting on Spatio-temporal Point Processes
This work addresses the challenge of multi-event forecasting in spatio-temporal point processes, which is crucial for applications like disaster prediction and public health, representing a novel method for a known bottleneck rather than an incremental improvement.
The paper tackled the problem of predicting multiple future discrete events in time and space, such as earthquakes or disease outbreaks, by proposing a neural architecture that uses transformers, normalizing flows, and probabilistic layers to achieve state-of-the-art performance on benchmark datasets like South California Earthquakes and Covid-19.
Predicting discrete events in time and space has many scientific applications, such as predicting hazardous earthquakes and outbreaks of infectious diseases. History-dependent spatio-temporal Hawkes processes are often used to mathematically model these point events. However, previous approaches have faced numerous challenges, particularly when attempting to forecast one or multiple future events. In this work, we propose a new neural architecture for simultaneous multi-event forecasting of spatio-temporal point processes, utilizing transformers, augmented with normalizing flows and probabilistic layers. Our network makes batched predictions of complex history-dependent spatio-temporal distributions of future discrete events, achieving state-of-the-art performance on a variety of benchmark datasets including the South California Earthquakes, Citibike, Covid-19, and Hawkes synthetic pinwheel datasets. More generally, we illustrate how our network can be applied to any dataset of discrete events with associated markers, even when no underlying physics is known.