Publicação
Malus Chatbot: A Chatbot for Apple Tree Cultivation in Portugal
| Resumo: | This thesis presents the development of the Malus Chatbot, a Retrieval-Augmented Generation conversational agent designed to support apple cultivation in Portugal. Apple growers and stakeholders frequently face challenges accessing timely and reliable information due to the fragmented and unstructured nature of agricultural knowledge sources. While advances in Large Language Models have enabled significant improvements in natural language understanding, most existing agricultural chatbots remain limited by static or rulebased approaches, resulting in outdated or generic responses. The Malus Chatbot addresses this gap by combining a curated, domain-specific knowledge base with advanced retrieval and generation techniques to provide contextually relevant, evidence-based answers. Uniquely, the system can extract and present visual information like the figures and tables from source documents and offers direct links to these sources alongside each answer, supporting transparency and user verification. The system employs state-of-the-art embeddings, a hybrid retrieval strategy, and the GPT-4.1-mini language model to generate accurate and informative responses. Evaluation of the chatbot was carried out using RAGAS metrics and an expert provided question-answer dataset. Results demonstrate that the RAG-based approach substantially enhances answer quality, reliability, and source traceability compared to traditional methods. This work highlights the potential of retrieval-augmented conversational AI for advancing specialized knowledge access in agriculture and lays the foundation for future research in domain-adapted chatbot systems. |
|---|---|
| Autores principais: | Chen, Zenan |
| Assunto: | Retrieval-Augmented Generation Chatbot Generative AI Large Language Models Natural Language Processing Agricultural Informatics Apple Cultivation SDG 8 - Decent work and economic growth SDG 9 - Industry, innovation and infrastructure SDG 12 - Responsible production and consumption SDG 15 - Life on land |
| 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 thesis presents the development of the Malus Chatbot, a Retrieval-Augmented Generation conversational agent designed to support apple cultivation in Portugal. Apple growers and stakeholders frequently face challenges accessing timely and reliable information due to the fragmented and unstructured nature of agricultural knowledge sources. While advances in Large Language Models have enabled significant improvements in natural language understanding, most existing agricultural chatbots remain limited by static or rulebased approaches, resulting in outdated or generic responses. The Malus Chatbot addresses this gap by combining a curated, domain-specific knowledge base with advanced retrieval and generation techniques to provide contextually relevant, evidence-based answers. Uniquely, the system can extract and present visual information like the figures and tables from source documents and offers direct links to these sources alongside each answer, supporting transparency and user verification. The system employs state-of-the-art embeddings, a hybrid retrieval strategy, and the GPT-4.1-mini language model to generate accurate and informative responses. Evaluation of the chatbot was carried out using RAGAS metrics and an expert provided question-answer dataset. Results demonstrate that the RAG-based approach substantially enhances answer quality, reliability, and source traceability compared to traditional methods. This work highlights the potential of retrieval-augmented conversational AI for advancing specialized knowledge access in agriculture and lays the foundation for future research in domain-adapted chatbot systems. |
|---|