Data-Efficient and Robust Trajectory Generation through Pathlet Dictionary Learning

TL;DR

This paper introduces a novel deep generative model using pathlet dictionary learning for efficient, robust, and interpretable trajectory generation, especially effective with noisy data and applicable to downstream mobility tasks.

Contribution

It proposes a probabilistic graphical model combining VAE and linear decoding to learn trajectory patterns and generate customized trajectories with improved robustness and efficiency.

Findings

Achieves 35.4% and 26.3% improvements over baselines on real datasets.

Effectively denoises data and supports downstream tasks like prediction.

Reduces training time by 64.8% and GPU memory usage by 56.5%.

Abstract

Trajectory generation has recently drawn growing interest in privacy-preserving urban mobility studies and location-based service applications. Although many studies have used deep learning or generative AI methods to model trajectories and have achieved promising results, the robustness and interpretability of such models are largely unexplored. This limits the application of trajectory generation algorithms on noisy real-world data and their trustworthiness in downstream tasks. To address this issue, we exploit the regular structure in urban trajectories and propose a deep generative model based on the pathlet representation, which encode trajectories with binary vectors associated with a learned dictionary of trajectory segments. Specifically, we introduce a probabilistic graphical model to describe the trajectory generation process, which includes a Variational Autoencoder (VAE) component and a linear decoder component. During training, the model can simultaneously learn the latent embedding of pathlet representations and the pathlet dictionary that captures mobility patterns in the trajectory dataset. The conditional version of our model can also be used to generate customized trajectories based on temporal and spatial constraints. Our model can effectively learn data distribution even using noisy data, achieving relative improvements of $35.4\%$ and $26.3\%$ over strong baselines on two real-world trajectory datasets. Moreover, the generated trajectories can be conveniently utilized for multiple downstream tasks, including trajectory prediction and data denoising. Lastly, the framework design offers a significant efficiency advantage, saving $64.8\%$ of the time and $56.5\%$ of GPU memory compared to previous approaches.