---
title: "PROB-EMOE: A Probabilistic Ensemble Mixture-of-Experts Framework for Metro Network Expansion Forecasting"
authors:
  - "Fangyi Ding"
  - "Zhan Zhao"
  - "Zhi Li"
  - "Xudong Guo"
  - "Ning Zhang"
  - "Yamin Wang"
  - "Yihong Tang"
doi: "10.24963/ijcai.2026/792"
url: "https://ijcai.org/proceedings/2026/792"
pdf: "https://ijcai.org/proceedings/2026/0792.pdf"
venue: "Proceedings of the Thirty-Fifth International Joint Conference on Artificial Intelligence"
year: 2026
pages: "7120-7128"
track: "AI and Social Good"
keywords:
  - "Uncertainty in AI: Uncertainty in AI"
  - "Multidisciplinary Topics and Applications: Multidisciplinary Topics and Applications"
  - "Data Mining: Data Mining"
---

# PROB-EMOE: A Probabilistic Ensemble Mixture-of-Experts Framework for Metro Network Expansion Forecasting

Fangyi Ding, Zhan Zhao, Zhi Li, Xudong Guo, Ning Zhang, Yamin Wang, Yihong Tang

*Proceedings of the Thirty-Fifth International Joint Conference on Artificial Intelligence*, pages 7120-7128. DOI: [10.24963/ijcai.2026/792](https://doi.org/10.24963/ijcai.2026/792)

## Abstract

Forecasting Origin-Destination (OD) demand for new metro lines is critical for sustainable infrastructure planning but faces spatiotemporal out-of-distribution challenges. Existing models often struggle to capture heterogeneous interaction patterns in changing topologies and overlook inherent uncertainty and over-dispersion issues. To bridge these gaps, we propose PROB-EMOE, a planning-oriented probabilistic framework tailored for network expansion. To ensure robust generalization, we design a Mixture-of-Experts (MoE) predictor that integrates diverse expert views to capture heterogeneous demand patterns across changing topologies. To quantify extrapolation uncertainty, our framework functions as a unified probabilistic system by synergizing Deep Ensembles with a probabilistic output, effectively quantifying both data and model uncertainty. Through a systematic investigation of likelihood families, we empirically demonstrate that the Negative Binomial distribution offers the optimal fit in this context. Extensive experiments on a multi-year Shenzhen metro dataset demonstrate that our approach achieves state-of-the-art predictive performance and provides the sharpest calibrated uncertainty intervals. The framework has been deployed in a metropolitan smart-data platform to support risk-aware investment decisions.