Publicação
Chatbot for Lisbon Tourism: A Proof of Concept Using LLMs and Retrieval-Augmented Generation to Assist Travelers in Choosing Airbnb Accommodations
| Resumo: | This work focuses on exploring chatbot technologies, proceed by developing, evaluating and deploying a prototype solution of a chatbot to help tourists in Lisbon, with the initial focus on helping tourists choosing an Airbnb to stay in given a set number of properties provided by our chosen dataset. The two main technologies we will investigate to power our chatbot with a LLM are finetuning and Retrieval Augmented Generation (RAG) in which we explore their uses and practical scenarios, but the version we developed will be using RAG with the help of LangFlow and Vector databases to deploy the chatbot. The chatbot was tested through a structured user survey, and feedback was collected to evaluate its usability, accuracy, and overall utility. Results indicate that users appreciated the ability to ask natural language questions and receive context-aware answers, although limitations in understanding complex or multi-step prompts were noted. This work contributes to the continuous growth of the AI and Chatbots world, and studies its potential uses in Tourism, demonstrating a practical case scenario of the use of RAG to implement a chatbot, while also providing insight into differences and benefits of using RAG or Fine-tuning in each scenario for LLM deployment in other works. |
|---|---|
| Autores principais: | Massapina, Vasco Lhansol Souto |
| Assunto: | Chatbot Tourism Lisbon Retrieval Augmented Generation LangFlow Finetuning Large Language Models Proof of Concept SDG 8 - Decent work and economic growth SDG 9 - Industry, innovation and infrastructure SDG 11 - Sustainable cities and communities |
| Ano: | 2025 |
| País: | Portugal |
| Tipo de documento: | dissertação de mestrado |
| Tipo de acesso: | acesso aberto |
| Instituição associada: | Universidade Nova de Lisboa |
| Idioma: | inglês |
| Origem: | Repositório Institucional da UNL |
| Resumo: | This work focuses on exploring chatbot technologies, proceed by developing, evaluating and deploying a prototype solution of a chatbot to help tourists in Lisbon, with the initial focus on helping tourists choosing an Airbnb to stay in given a set number of properties provided by our chosen dataset. The two main technologies we will investigate to power our chatbot with a LLM are finetuning and Retrieval Augmented Generation (RAG) in which we explore their uses and practical scenarios, but the version we developed will be using RAG with the help of LangFlow and Vector databases to deploy the chatbot. The chatbot was tested through a structured user survey, and feedback was collected to evaluate its usability, accuracy, and overall utility. Results indicate that users appreciated the ability to ask natural language questions and receive context-aware answers, although limitations in understanding complex or multi-step prompts were noted. This work contributes to the continuous growth of the AI and Chatbots world, and studies its potential uses in Tourism, demonstrating a practical case scenario of the use of RAG to implement a chatbot, while also providing insight into differences and benefits of using RAG or Fine-tuning in each scenario for LLM deployment in other works. |
|---|
Atividades financiadas
Não foram encontrados projetos financiados.