Authors:

Jungeun Hwang

Date Created:

Not specified

Course Title:

Professor:

Not specified

About Paper:

Chatbots have recently gained traction due to advancements in artificial intelligence, natural language processing, and computational resources. Chatbots are being developed and integrated with messaging platforms for multiple applications, including education, customer service, healthcare, etc. However, their application in agriculture for helping farmers manage crops and improve yield is limited. Therefore, this study focuses on developing a chatbot to assist farmers in managing weeds to improve crop yield. The proposed method enables farmers to ask questions and receive relevant responses for easier and faster decision- making corresponding to their fields. The user interface for the chatbot was developed using React Native, enabling deployment on both iOS and Android platforms. Pinecone was used to store the agriculture-related extension articles and corresponding vectors. Natural language processing (NLP) was used to vectorize and interpret user queries matched against a vectorized pinecone dataset to find the relevant results. Additionally, a large language model (ChatGPT-turbo-3.5) and an intent classification model were utilized to provide decision support for weed management. The intent classification was implemented through Bidirectional Encoder Representations from Transformers (BERT). The chatbot also included the ability to provide a task summary service to help farmers track real-time weather data. Overall, the proposed chatbot will aid farmers in obtaining answers to queries corresponding to weed management.

Source:

Purdue University / 2023

Topics:

No topics listed

Co-authors:

Jungeun Hwang