Hyper Hawkes Processes: Interpretable Models of Marked Temporal Point Processes

ICLR 2026 Conference Submission14345 Authors

Keywords: Marked temporal point processes, interpretability, probabilistic methods

TL;DR: We present an efficient SotA neural MTPP that allows key event-level interpretability to be directly extracted.

Abstract: Foundational marked temporal point process (MTPP) models, such as the Hawkes process, often use inexpressive model families in order to offer interpretable parameterizations of event data. On the other hand, neural MTPPs models forego this interpretability in favor of absolute predictive performance. In this work, we present a new family MTPP models: the _hyper Hawkes process_ (HHP), which aims to be as flexible and performant as neural MTPPs, while retaining interpretable aspects. To achieve this, the HHP extends the classical Hawkes process to increase its expressivity by first expanding the dimension of the process into a latent space, and then introducing a hypernetwork to allow time and data dynamics. These extensions define a highly performant MTPP family, achieving state-of-the-art performance across a range of benchmark tasks and metrics. Furthermore, by retaining the now-conditionally linear recurrence, the HHP also retains much of the structure of the original Hawkes process, which we exploit to create direct probes into _how_ the model creates predictions. HHP models therefore offer both state-of-the-art predictions, while also providing an opportunity to ``open the box'' and inspect how predictions were generated.

Primary Area: probabilistic methods (Bayesian methods, variational inference, sampling, UQ, etc.)

Submission Number: 14345