Salvato in:
Dettagli Bibliografici
Autori principali: Mukherjee, Amartya, Deng, Ruizhi, Zhao, He, Mao, Yuzhen, Sigal, Leonid, Tung, Frederick
Natura: Preprint
Pubblicazione: 2025
Soggetti:
Accesso online:https://arxiv.org/abs/2504.20411
Tags: Aggiungi Tag
Nessun Tag, puoi essere il primo ad aggiungerne!!
_version_ 1866915265656127488
author Mukherjee, Amartya
Deng, Ruizhi
Zhao, He
Mao, Yuzhen
Sigal, Leonid
Tung, Frederick
author_facet Mukherjee, Amartya
Deng, Ruizhi
Zhao, He
Mao, Yuzhen
Sigal, Leonid
Tung, Frederick
contents This work introduces a novel approach to modeling temporal point processes using diffusion models with an asynchronous noise schedule. At each step of the diffusion process, the noise schedule injects noise of varying scales into different parts of the data. With a careful design of the noise schedules, earlier events are generated faster than later ones, thus providing stronger conditioning for forecasting the more distant future. We derive an objective to effectively train these models for a general family of noise schedules based on conditional flow matching. Our method models the joint distribution of the latent representations of events in a sequence and achieves state-of-the-art results in predicting both the next inter-event time and event type on benchmark datasets. Additionally, it flexibly accommodates varying lengths of observation and prediction windows in different forecasting settings by adjusting the starting and ending points of the generation process. Finally, our method shows superior performance in long-horizon prediction tasks, outperforming existing baseline methods.
format Preprint
id arxiv_https___arxiv_org_abs_2504_20411
institution arXiv
publishDate 2025
record_format arxiv
spellingShingle ADiff4TPP: Asynchronous Diffusion Models for Temporal Point Processes
Mukherjee, Amartya
Deng, Ruizhi
Zhao, He
Mao, Yuzhen
Sigal, Leonid
Tung, Frederick
Machine Learning
This work introduces a novel approach to modeling temporal point processes using diffusion models with an asynchronous noise schedule. At each step of the diffusion process, the noise schedule injects noise of varying scales into different parts of the data. With a careful design of the noise schedules, earlier events are generated faster than later ones, thus providing stronger conditioning for forecasting the more distant future. We derive an objective to effectively train these models for a general family of noise schedules based on conditional flow matching. Our method models the joint distribution of the latent representations of events in a sequence and achieves state-of-the-art results in predicting both the next inter-event time and event type on benchmark datasets. Additionally, it flexibly accommodates varying lengths of observation and prediction windows in different forecasting settings by adjusting the starting and ending points of the generation process. Finally, our method shows superior performance in long-horizon prediction tasks, outperforming existing baseline methods.
title ADiff4TPP: Asynchronous Diffusion Models for Temporal Point Processes
topic Machine Learning
url https://arxiv.org/abs/2504.20411