Deep-Ensemble-Based Uncertainty Quantification in Spatiotemporal Graph Neural Networks for Traffic Forecasting

TL;DR

This paper introduces a scalable deep ensemble method to quantify uncertainty in traffic forecasting models, specifically enhancing the DCRNN by capturing joint hyperparameter distributions for improved uncertainty estimates.

Contribution

The paper presents a novel scalable deep ensemble approach that leverages Bayesian optimization and generative modeling to improve uncertainty quantification in traffic forecasting neural networks.

Findings

Outperforms existing Bayesian and frequentist uncertainty estimation techniques.

Provides more accurate and reliable uncertainty estimates for traffic forecasting.

Demonstrates scalability and effectiveness on short-term traffic prediction tasks.

Abstract

Deep-learning-based data-driven forecasting methods have produced impressive results for traffic forecasting. A major limitation of these methods, however, is that they provide forecasts without estimates of uncertainty, which are critical for real-time deployments. We focus on a diffusion convolutional recurrent neural network (DCRNN), a state-of-the-art method for short-term traffic forecasting. We develop a scalable deep ensemble approach to quantify uncertainties for DCRNN. Our approach uses a scalable Bayesian optimization method to perform hyperparameter optimization, selects a set of high-performing configurations, fits a generative model to capture the joint distributions of the hyperparameter configurations, and trains an ensemble of models by sampling a new set of hyperparameter configurations from the generative model. We demonstrate the efficacy of the proposed methods by comparing them with other uncertainty estimation techniques. We show that our generic and scalable approach outperforms the current state-of-the-art Bayesian and a number of other commonly used frequentist techniques.