Add and Thin: Diffusion for Temporal Point Processes
This addresses forecasting limitations in TPPs for applications like event sequence modeling, though it is incremental as it adapts diffusion models to a specific domain.
The authors tackled the problem of long-term forecasting in temporal point processes (TPPs) by developing ADD-THIN, a denoising diffusion model that operates on entire event sequences, which matches state-of-the-art models in density estimation and strongly outperforms them in forecasting.
Autoregressive neural networks within the temporal point process (TPP) framework have become the standard for modeling continuous-time event data. Even though these models can expressively capture event sequences in a one-step-ahead fashion, they are inherently limited for long-term forecasting applications due to the accumulation of errors caused by their sequential nature. To overcome these limitations, we derive ADD-THIN, a principled probabilistic denoising diffusion model for TPPs that operates on entire event sequences. Unlike existing diffusion approaches, ADD-THIN naturally handles data with discrete and continuous components. In experiments on synthetic and real-world datasets, our model matches the state-of-the-art TPP models in density estimation and strongly outperforms them in forecasting.