Efficient Temporal Point Processes via Monotone Alternating Splines
For researchers working on temporal point processes, this work addresses fundamental limitations in existing CCIF parameterizations, offering a more flexible and efficient modeling framework.
The paper identifies structural deadlocks in Monotone Neural Networks for modeling cumulative conditional intensity functions in temporal point processes and proposes Monotone Alternating Splines (MAS) to overcome them. MAS achieves superior performance on synthetic and real-world datasets, reducing approximation gaps.
Temporal point processes (TPPs) have widespread applications across various domains. Compared to modeling the conditional intensity of a TPP, modeling its cumulative conditional intensity function (CCIF) improves computational efficiency and eliminates numerical approximation errors. However, current CCIF parameterizations uniformly rely on Monotone Neural Networks (MNNs), which we identify as suffering from three structural deadlocks--convexity restrictions, saturation limits, and violations of CCIF modeling requirements--that fundamentally restrict their representational capacity for complex temporal dynamics. To resolve these bottlenecks, this paper proposes a novel framework called Monotone Alternating Splines (MAS). By leveraging distinct interpolation and extrapolation components, MAS provides a flexible and efficient framework for modeling CCIFs. Theoretically, MAS's interpolation provides strong fitting accuracy, while its extrapolation supports robust generalization, reducing the irreducible approximation gaps of MNNs. Extensive experiments show that MAS achieves superior performance on both synthetic and real-world datasets.