Bayesian-Symbolic Integration for Uncertainty-Aware Parking Prediction

2026-03-28

AI Summary

This work addresses the challenges of data sparsity, noise interference, and abrupt environmental changes that introduce uncertainty in real-world parking availability prediction. To tackle these issues, the authors propose a loosely coupled neuro-symbolic framework that integrates Bayesian neural networks with symbolic reasoning. The framework employs two hybrid strategies: falling back to interpretable symbolic rules under low prediction confidence or leveraging symbolic constraints to refine neural predictions, thereby effectively combining uncertainty awareness with domain knowledge. Experimental results demonstrate that the proposed approach consistently outperforms LSTM and standalone Bayesian neural network baselines across full, sparse, and noisy data regimes, confirming its enhanced robustness and accuracy in complex, real-world scenarios.

Abstract

Accurate parking availability prediction is critical for intelligent transportation systems, but real-world deployments often face data sparsity, noise, and unpredictable changes. Addressing these challenges requires models that are not only accurate but also uncertainty-aware. In this work, we propose a loosely coupled neuro-symbolic framework that integrates Bayesian Neural Networks (BNNs) with symbolic reasoning to enhance robustness in uncertain environments. BNNs quantify predictive uncertainty, while symbolic knowledge extracted via decision trees and encoded using probabilistic logic programming is leveraged in two hybrid strategies: (1) using symbolic reasoning as a fallback when BNN confidence is low, and (2) refining output classes based on symbolic constraints before reapplying the BNN. We evaluate both strategies on real-world parking data under full, sparse, and noisy conditions. Results demonstrate that both hybrid methods outperform symbolic reasoning alone, and the context-refinement strategy consistently exceeds the performance of Long Short-Term Memory (LSTM) networks and BNN baselines across all prediction windows. Our findings highlight the potential of modular neuro-symbolic integration in real-world, uncertainty-prone prediction tasks.

Problem

• parking prediction
• uncertainty-aware
• data sparsity
• noise
• intelligent transportation systems

Innovation

• Bayesian Neural Networks
• Symbolic Reasoning
• Neuro-Symbolic Integration
• Uncertainty Quantification
• Probabilistic Logic Programming