Aligning LLM with human travel choices: a persona-based embedding learning approach

TL;DR

This paper presents a new framework that aligns large language models with human travel choices using persona-based embeddings, improving prediction accuracy and interpretability in travel demand modeling.

Contribution

It introduces a novel persona inference and loading method that enhances LLM alignment with human travel behavior using behavioral embeddings and empirical data.

Findings

Outperforms baseline models in predicting mode choice shares.

Accurately predicts individual travel choices.

Provides interpretable insights into population behavior.

Abstract

The advent of large language models (LLMs) presents new opportunities for travel demand modeling. However, behavioral misalignment between LLMs and humans presents obstacles for the usage of LLMs, and existing alignment methods are frequently inefficient or impractical given the constraints of typical travel demand data. This paper introduces a novel framework for aligning LLMs with human travel choice behavior, tailored to the current travel demand data sources. Our framework uses a persona inference and loading process to condition LLMs with suitable prompts to enhance alignment. The inference step establishes a set of base personas from empirical data, and a learned persona loading function driven by behavioral embeddings guides the loading process. We validate our framework on the Swissmetro mode choice dataset, and the results show that our proposed approach significantly outperformed baseline choice models and LLM-based simulation models in predicting both aggregate mode choice shares and individual choice outcomes. Furthermore, we showcase that our framework can generate insights on population behavior through interpretable parameters. Overall, our research offers a more adaptable, interpretable, and resource-efficient pathway to robust LLM-based travel behavior simulation, paving the way to integrate LLMs into travel demand modeling practice in the future.