In the hustle and bustle of the city, everyone longs for a citywalk, walking through the streets and alleys, exploring historical sites, and immersing themselves in the local culture. However, planning a satisfying city walk is not easy. It requires the collection of travel information, the selection of points of interest (POI), route planning, and personalized customization. Now, a project led by XiangHKUThe new technology was developed by researchers from the Chinese Academy of Sciences, Shanghai Jiao Tong University, Massachusetts Institute of Technology, and the Hong Kong Polytechnic University.ItiNera, making this process simple and efficient.
ItiNera is an Open Domain Urban Itinerary Planning (OUIP) system that can generate personalized itineraries directly based on user requests described in natural language. Compared with traditional itinerary planning, ItiNera is not restricted to a specific field and can meet the personalized needs of users in a more detailed manner.
The core of the ItiNera system is that it combines spatial optimization with large language models (LLMs) to provide customized services. The system extracts and updates POI features through an LLM-based pipeline to create a personalized POI database owned by the user. For each user request, ItiNera uses LLM in conjunction with an embedding-based module to retrieve candidate POIs from the user's POI database, then uses the spatial optimization module to sort these POIs, and finally LLM carefully creates a personalized, spatially coherent itinerary.
The researchers conducted extensive experiments on ItiNera on multiple datasets and conducted online subjective evaluations. The results show that ItiNera is able to provide more responsive and spatially coherent itineraries compared to existing LLM-based solutions. In addition, after being deployed on the TuTu online travel service, ItiNera has attracted thousands of users to use its city travel planning service.
ItiNera's application not only provides an efficient planning tool for city walks, but also provides travelers with a new way to explore the city. It can generate personalized and detailed travel experiences based on the user's specific needs and lifestyle, greatly enriching the traveler's urban exploration experience.
The research results of the ItiNera system have been made public, and the research team hopes that this system can inspire more research on using LLMs to solve complex urban problems. By opening up the system, more researchers and developers can participate in the optimization and innovation of this technology.
Paper address: https://arxiv.org/pdf/2402.07204