Cyclic Graph Attentive Match Encoder (CGAME): A Novel Neural Network For OD Estimation
This addresses OD estimation for intelligent transportation systems, offering a novel deep learning architecture for cross-space inference tasks.
The authors tackled the under-determined problem of Origin-Destination (OD) estimation in intelligent transportation by proposing CGAME, a neural network that uses a cyclic graph attentive match encoder to handle cross-space inference, achieving state-of-the-art results in experiments.
Origin-Destination Estimation plays an important role in the era of Intelligent Transportation. Nevertheless, as a under-determined problem, OD estimation confronts many challenges from cross-space inference to non-convex, non-linear optimization. As a powerful nonlinear approximator, deep learning is an ideal data-driven method to provide a novel perspective for OD estimation. However, viewing multi-interval traffic counts as spatial-temporal inputs and OD matrix as heterogeneous graph-structured output, the existing neural network architecture is not suitable for the cross-space inference problem thus a new deep learning architecture is needed. We propose CGAME, short for cyclic graph attentive matching encoder, including bi-directional encoder-decoder networks and a novel graph matcher in the hidden layer with double-layer attention mechanism. It realizes effective information exchange between the forward networks and backward networks and establishes coupling relations across underlying feature space. The proposed model achieves state-of-the-art compared with baselines in the designed experiments and offers a paradigm for inference tasks across representation space.