ItiNera: Integrating Spatial Optimization with Large Language Models for Open-domain Urban Itinerary Planning

2024-02-11

Conference on Empirical Methods in Natural Language Processing

AI Summary

Personalized itinerary planning for urban walking—a novel travel paradigm—remains challenging due to the difficulty of interpreting fine-grained, open-domain natural language user preferences and satisfying geographic constraints simultaneously. Method: This paper formally defines the open-domain natural language–driven urban itinerary planning task and proposes a collaborative framework integrating large language models (LLMs) and spatial optimization. Specifically, an LLM parses nuanced user intents; a graph neural network learns joint semantic-geographic representations of points of interest (POIs); and a clustering-aware spatial path optimization algorithm ensures both personalization and geographic coherence. The method employs a multi-stage itinerary generation pipeline. Contribution/Results: Evaluated on a real-world dataset, our approach significantly outperforms state-of-the-art baselines. The system is open-sourced and deployed in practice, empirically validating the effectiveness of jointly modeling semantic understanding and geometric constraints for urban itinerary planning.

Technology Category

• Planning, Routing, and Scheduling: Planning with Language Models
• Search and Optimization: Learning to Search
• Constraint Satisfaction and Optimization: Search

Application Category

• User Modeling, Personalization and Recommendation: Large Language Models (LLM) for user modeling and recommendation
• Graph Algorithms and Modeling for the Web: Foundation models and LLMs for Web-related graphs
• Semantics and Knowledge: Methods to enhance, augment, integrate or synergize semantic models such as knowledge graphs and LLMs

Abstract

Citywalk, a recently popular form of urban travel, requires genuine personalization and understanding of fine-grained requests compared to traditional itinerary planning. In this paper, we introduce the novel task of Open-domain Urban Itinerary Planning (OUIP), which generates personalized urban itineraries from user requests in natural language. We then present ItiNera, an OUIP system that integrates spatial optimization with large language models to provide customized urban itineraries based on user needs. This involves decomposing user requests, selecting candidate points of interest (POIs), ordering the POIs based on cluster-aware spatial optimization, and generating the itinerary. Experiments on real-world datasets and the performance of the deployed system demonstrate our system’s capacity to deliver personalized and spatially coherent itineraries compared to current solutions. Source codes of ItiNera are available at https://github.com/YihongT/ITINERA.

Problem

• Personalized Travel Planning
• Urban Wanderlust
• Customized Itinerary

Innovation

• Spatial Optimization
• Natural Language Processing
• Personalized Urban Itinerary Planning