ITPP: Learning Disentangled Event Dynamics in Marked Temporal Point Processes

Fuente: arXiv
Saved in:
Bibliographic Details
Main Authors: Zhou, Wang-Tao, Kang, Zhao, Yan, Ke, Tian, Ling
Format: Preprint
Published: 2025
Subjects:
Online Access:
Tags: Add Tag
No Tags, Be the first to tag this record!
_version_ 1866911255585882112
author Zhou, Wang-Tao
Kang, Zhao
Yan, Ke
Tian, Ling
author_facet Zhou, Wang-Tao
Kang, Zhao
Yan, Ke
Tian, Ling
contents Marked Temporal Point Processes (MTPPs) provide a principled framework for modeling asynchronous event sequences by conditioning on the history of past events. However, most existing MTPP models rely on channel-mixing strategies that encode information from different event types into a single, fixed-size latent representation. This entanglement can obscure type-specific dynamics, leading to performance degradation and increased risk of overfitting. In this work, we introduce ITPP, a novel channel-independent architecture for MTPP modeling that decouples event type information using an encoder-decoder framework with an ODE-based backbone. Central to ITPP is a type-aware inverted self-attention mechanism, designed to explicitly model inter-channel correlations among heterogeneous event types. This architecture enhances effectiveness and robustness while reducing overfitting. Comprehensive experiments on multiple real-world and synthetic datasets demonstrate that ITPP consistently outperforms state-of-the-art MTPP models in both predictive accuracy and generalization.
format Preprint
id arxiv_https___arxiv_org_abs_2511_06032
institution arXiv
publishDate 2025
record_format arxiv
spellingShingle ITPP: Learning Disentangled Event Dynamics in Marked Temporal Point Processes
Zhou, Wang-Tao
Kang, Zhao
Yan, Ke
Tian, Ling
Machine Learning
Artificial Intelligence
Marked Temporal Point Processes (MTPPs) provide a principled framework for modeling asynchronous event sequences by conditioning on the history of past events. However, most existing MTPP models rely on channel-mixing strategies that encode information from different event types into a single, fixed-size latent representation. This entanglement can obscure type-specific dynamics, leading to performance degradation and increased risk of overfitting. In this work, we introduce ITPP, a novel channel-independent architecture for MTPP modeling that decouples event type information using an encoder-decoder framework with an ODE-based backbone. Central to ITPP is a type-aware inverted self-attention mechanism, designed to explicitly model inter-channel correlations among heterogeneous event types. This architecture enhances effectiveness and robustness while reducing overfitting. Comprehensive experiments on multiple real-world and synthetic datasets demonstrate that ITPP consistently outperforms state-of-the-art MTPP models in both predictive accuracy and generalization.
title ITPP: Learning Disentangled Event Dynamics in Marked Temporal Point Processes
topic Machine Learning
Artificial Intelligence
url https://arxiv.org/abs/2511.06032