Intensity-Free Learning of Temporal Point Processes [article]

Oleksandr Shchur, Marin Biloš, Stephan Günnemann
2020 arXiv   pre-print
Temporal point processes are the dominant paradigm for modeling sequences of events happening at irregular intervals. The standard way of learning in such models is by estimating the conditional intensity function. However, parameterizing the intensity function usually incurs several trade-offs. We show how to overcome the limitations of intensity-based approaches by directly modeling the conditional distribution of inter-event times. We draw on the literature on normalizing flows to design
more » ... ls that are flexible and efficient. We additionally propose a simple mixture model that matches the flexibility of flow-based models, but also permits sampling and computing moments in closed form. The proposed models achieve state-of-the-art performance in standard prediction tasks and are suitable for novel applications, such as learning sequence embeddings and imputing missing data.
arXiv:1909.12127v2 fatcat:6fgz2rfrc5fabd3sszdyub6fiq