Artificial Intelligence in Mental Health Care: The T5 Chatbot Project

Document Type

Conference Proceeding

Source of Publication

2024 11th International Conference on Social Networks Analysis, Management and Security (SNAMS)

Publication Date

12-11-2024

Abstract

A unique mental health chatbot using the T5 (Text-to-Text Transfer Transformer) model is the focus of this study. This project aims to meet a crucial mental health need by providing quick, sympathetic, and effective conversational support to mental health patients. A comprehensive and representative dialogue model was created by carefully collecting and curating datasets like mental health counseling conversations and others. The chatbot was trained on meticulously cleaned and processed datasets. T5 was chosen for this application because it understands and generates human like text well. PEFT and LoRA were used to overcome the model's computing demands. These methods trained the T5 model efficiently on restricted hardware without affecting performance. Due to limited dataset quantity and training period, the training procedure was difficult. The training loss measures showed that the model improved gradually over three epochs despite these challenges.

ISBN

979-8-3315-1834-9

Publisher

IEEE

Volume

00

First Page

212

Last Page

218

Disciplines

Computer Sciences | Medicine and Health Sciences

Keywords

AI, mental health, chatbot, T5 model, conversational support

Indexed in Scopus

no

Open Access

no

Link to Full Text

Share

COinS