[{“box”:0,”content”:”[if 992 equals=”Open Access”]
Open Access
n
[/if 992]n
n
n
n
n
- n t
n
Ashwini Garole, Aditya Asabe, Mohini Jadhav, Shreepad Chavan, Shifa Gadiwale
[/foreach]
n
n
n[if 2099 not_equal=”Yes”]n
- [foreach 286] [if 1175 not_equal=””]n t
- Student, Student, Student, Student, Student, Department of CSE (AI & ML), Vishwaniketan’s IMEET, Khalapur Mumbai University, Mumbai, Department of CSE (AI & ML), Vishwaniketan’s IMEET, Khalapur Mumbai University, Mumbai, Department of CSE (AI & ML), Vishwaniketan’s IMEET, Khalapur Mumbai University, Mumbai, Department of CSE (AI & ML), Vishwaniketan’s IMEET, Khalapur Mumbai University, Mumbai, Department of CSE (AI & ML), Vishwaniketan’s IMEET, Khalapur Mumbai University, Mumbai, Maharashtra, Maharashtra, Maharashtra, Maharashtra, Maharashtra, India, India, India, India, India
n[/if 1175][/foreach]
[/if 2099][if 2099 equals=”Yes”][/if 2099]nn
Abstract
nMental health is a crucial aspect of overall well-being, yet access to professional therapy remains a significant challenge for many individuals due to various barriers, including cost, availability, and stigma. This research aims to develop an accessible and effective mental health therapy chatbot, named Mindwell Psychology, leveraging the power of large language models (LLMs) and state-of-the-art natural language processing techniques. The primary objective of this study is to create a conversational AI system capable of providing personalized, empathetic, and evidence-based psychological support to users. By harnessing the capabilities of Google’s GEMMA-2B, a cutting-edge LLM, and employing the LLaMA-Factory framework, we fine-tuned the model on a curated dataset of therapeutic conversations and mental health resources. The proposed approach involves curating a large dataset of mental health-related conversations and resources, preprocessing and formatting the data for model training, and fine-tuning the GEMMA-2B model using the LLaMA-Factory framework. The fine-tuned model is then integrated into a user-friendly web application, enabling seamless interaction with the Mindwell Psychology chatbot. Evaluation of the chatbot’s performance was conducted using a combination of quantitative metrics, such as perplexity and BLEU scores, and qualitative analysis of sample conversations. The results demonstrate the chatbot’s ability to provide empathetic and relevant responses, offering psychological support and coping strategies tailored to the user’s specific needs. The creation of the Mindwell Psychology chatbot marks a notable advancement in improving access to mental health assistance and fostering overall wellness. By leveraging cutting-edge language models and natural language processing techniques, this research contributes to the field of conversational AI and its application in the mental health domain.
n
Keywords: Mental Health Therapy, Chatbot, Artificial intelligence, Machine learning, Power, Large Language Models
n[if 424 equals=”Regular Issue”][This article belongs to Journal of Artificial Intelligence Research & Advances(joaira)]
n
n
n
n
n
n
n[if 992 equals=”Open Access”] Full Text PDF Download[/if 992] nn
n[if 379 not_equal=””]nBrowse Figures
n
n[/if 379]n
References
n[if 1104 equals=””]n
- World Health Organization. (2022). Mental health. https://www.who.int/health-topics/mental-health
- Kocaballi, A. B., Quiroz, J. C., Rezazadegan, D., Berkovsky, S., Magrabi, F., Coiera, E., & Laranjo, L. (2020). Conversational agents for health and wellness: Theory and reflection. Journal of Biomedical Informatics, 111, 103577.
- Rashkin, H., Sap, M., Lakshmi, N., & Swayamdipta, S. (2022). Towards empathetic and ethical chatbots: Bridging the gap between AI and human intelligence. arXiv preprint arXiv:2210.08427.
- Raffel, C., Shazeer, N., Roberts, A., Lee, K., Narang, S., Matena, M., … & Liu, P. J. (2020). Exploring the limits of transfer learning with a unified text-to-text transformer. Journal of Machine Learning Research, 21(140), 1-67.
- LLaMA-Factory (n.d.). GitHub repository. https://github.com/LLaMA-Factory/LLaMA-Factory
- Hugging Face. (n.d.). Transformers: State-of-the-art machine learning for language models. https://huggingface.co/transformers/
- Papineni, K., Roukos, S., Ward, T., & Zhu, W. J. (2002). BLEU: a method for automatic evaluation of machine translation. In Proceedings of the 40th annual meeting of the Association for Computational Linguistics (pp. 311-318).
- Health. Powered by Ada. Ada. Ada; 2024 . Available from: https://ada.com/
- Wysa – Everyday Mental Health . Wysa.com. 2024. Available from: https://www.wysa.com/ [10] https://elomia.com/
- Woebot Health. Woebot Health. 2024 . Available from: https://woebothealth.com/
- Nuna – Science and Research. Nuna: Mental Health Companion. 2020 . Available from: https://www.nuna.ai/research
- Youper: Artificial Intelligence For Mental Health Care. Youper.ai. 2023 . Available from: https://www.youper.ai/
nn[/if 1104][if 1104 not_equal=””]n
- [foreach 1102]n t
- [if 1106 equals=””], [/if 1106][if 1106 not_equal=””],[/if 1106]
n[/foreach]
n[/if 1104]
nn
nn[if 1114 equals=”Yes”]n
n[/if 1114]
n
n
n
Journal of Artificial Intelligence Research & Advances
n
n
n
n
n
n
| Volume | 11 | |
| [if 424 equals=”Regular Issue”]Issue[/if 424][if 424 equals=”Special Issue”]Special Issue[/if 424] [if 424 equals=”Conference”][/if 424] | 01 | |
| Received | March 28, 2024 | |
| Accepted | April 7, 2024 | |
| Published | April 22, 2024 |
n
n
n
n
n
nn function myFunction2() {n var x = document.getElementById(“browsefigure”);n if (x.style.display === “block”) {n x.style.display = “none”;n }n else { x.style.display = “Block”; }n }n document.querySelector(“.prevBtn”).addEventListener(“click”, () => {n changeSlides(-1);n });n document.querySelector(“.nextBtn”).addEventListener(“click”, () => {n changeSlides(1);n });n var slideIndex = 1;n showSlides(slideIndex);n function changeSlides(n) {n showSlides((slideIndex += n));n }n function currentSlide(n) {n showSlides((slideIndex = n));n }n function showSlides(n) {n var i;n var slides = document.getElementsByClassName(“Slide”);n var dots = document.getElementsByClassName(“Navdot”);n if (n > slides.length) { slideIndex = 1; }n if (n (item.style.display = “none”));n Array.from(dots).forEach(n item => (item.className = item.className.replace(” selected”, “”))n );n slides[slideIndex – 1].style.display = “block”;n dots[slideIndex – 1].className += ” selected”;n }n”}]