Probabilistic Circuits for Irregular Multivariate Time Series Forecasting

TL;DR

This paper introduces CircuITS, a probabilistic circuit-based model for irregular multivariate time series forecasting that ensures valid joint distributions and improves density estimation accuracy.

Contribution

The paper presents a novel probabilistic circuit architecture for IMTS forecasting that balances expressivity with reliable marginalization, outperforming existing methods.

Findings

CircuITS achieves superior joint and marginal density estimation.

The model effectively captures complex dependencies between time series channels.

Experiments on four datasets validate its improved performance.

Abstract

Joint probabilistic modeling is essential for forecasting irregular multivariate time series (IMTS) to accurately quantify uncertainty. Existing approaches often struggle to balance model expressivity with consistent marginalization, frequently leading to unreliable or contradictory forecasts. To address this, we propose CircuITS, a novel architecture for probabilistic IMTS forecasting based on probabilistic circuits. Our model is flexible in capturing intricate dependencies between time series channels while structurally guaranteeing valid joint distributions. Experiments on four real world datasets demonstrate that CircuITS achieves superior joint and marginal density estimation compared to state of the art baselines.