This is a page that curates AI-related papers published worldwide. All content here is summarized using Google Gemini and operated on a non-profit basis. Copyright for each paper belongs to the authors and their institutions; please make sure to credit the source when sharing.
Graph RAG as Human Choice Model: Building a Data-Driven Mobility Agent with Preference Chain
Created by
Haebom
Author
Kai Hu, Parfait Atchade-Adelomou, Carlo Adornetto, Adrian Mora-Carrero, Luis Alonso-Pastor, Ariel Noyman, Yubo Liu, Kent Larson
Outline
This paper presents "Preference Chain," a novel method utilizing a generative agent based on a large-scale language model (LLM) to address the challenges of collecting human behavior data in newly developed areas in urban science. Preference Chain integrates Graph Search Augmented Generation (RAG) with LLM to enhance contextual simulation of human behavior in transportation systems. Experimental results using the Replica dataset demonstrate that Preference Chain outperforms standard LLM in terms of consistency with real-world transportation mode choices. The development of the Mobility Agent demonstrates potential applications in emerging cities, including urban mobility modeling, personalized travel behavior analysis, and dynamic traffic prediction. Despite limitations such as slow inference speed and the risk of hallucination (T268095_____), it provides a promising framework for simulating complex human behavior in data-scarce environments.
Takeaways, Limitations
•
Takeaways:
◦
A novel methodology for simulating human behavior in data-poor environments.
◦
Improving Context-Aware Human Behavior Simulation through Integration of RAG and LLM
◦
It presents various potential applications, including urban mobility modeling in emerging cities, personalized travel behavior analysis, and dynamic traffic prediction.
◦
Demonstrating the effectiveness of the methodology through high consistency with actual transportation choices.
