EMILIA: Emotional Mental-Health Interactive Listener and Intelligent Assistant

Hartanto, Malvin Leonardo and Pramudya, Rafli Raihan (2025) EMILIA: Emotional Mental-Health Interactive Listener and Intelligent Assistant. Project Report. [s.n.], [s.l.]. (Unpublished)

Text 5025221033_5025221266-Project_Report.pdf - Accepted Version Restricted to Repository staff only Download (22MB) | Request a copy

Abstract

Permasalahan kesehatan mental memberikan pengaruh yang signifikan terhadap individu di berbagai kelompok usia, terutama remaja, dengan dampak yang dapat mengganggu aspek kognitif, emosional, dan perilaku. Penelitian ini berfokus pada pengembangan Emotional Mental-health Interactive Listener and Intelligent Assistant (EMILIA), sebuah asisten virtual empatik berbasis Large Language Model (LLM) yang dirancang untuk memberikan dukungan emosional secara interaktif. Sistem ini mengintegrasikan avatar antropomorfik bergaya anime dengan palet warna lembut untuk membangun keterikatan emosional dan meningkatkan rasa percaya pengguna. Proses implementasi meliputi pengembangan antarmuka interaktif menggunakan Unity, perancangan avatar melalui Live2D, serta integrasi sistem kecerdasan buatan dengan model Qwen-3 yang diadaptasi melalui teknik Low-Rank Adaptation (LoRA). Aplikasi ini dilengkapi dengan fitur utama seperti percakapan teks dan suara, pengelolaan jurnal pengguna, serta kemampuan agen AI untuk merespons konteks emosional dari isi jurnal tersebut. Hasil pengujian terhadap enam kebutuhan fungsional utama menunjukkan bahwa seluruh fitur sistem, meliputi autentikasi pengguna, interaksi percakapan, pengelolaan jurnal, komunikasi dua arah melalui teks dan suara, serta pemanfaatan AI agent, telah berjalan dengan baik tanpa ditemukan kesalahan fungsional yang berarti. Secara keseluruhan, EMILIA menunjukkan potensi yang kuat dalam meningkatkan keterlibatan dan dukungan emosional bagi remaja dalam konteks layanan kesehatan mental.
=================================================================================================================================
Mental health problems have a significant impact on individuals across different age groups, particularly adolescents, as they can disrupt cognitive, emotional, and behavioral aspects. This study focuses on the development of the Emotional Mental-health Interactive Listener and Intelligent Assistant (EMILIA), an empathetic virtual assistant based on a Large Language Model (LLM) designed to provide interactive emotional support. The system integrates an anthropomorphic anime-style avatar with a soft color palette to foster emotional attachment and build user trust. The implementation process includes developing an interactive interface using Unity, designing the avatar through Live2D, and integrating an artificial intelligence system utilizing the Qwen-3 model adapted through the Low-Rank Adaptation (LoRA) technique. The application is equipped with core features such as text and voice-based conversations, user journal management, and an AI agent capable of responding to the emotional context of the user’s journal entries. Testing conducted on six main functional requirements indicates that all system features, including user authentication, conversational interaction, journal management, two-way communication via text and voice, and AI agent functionality, performed well without significant functional errors. Overall, EMILIA demonstrates strong potential to enhance emotional engagement and provide supportive interaction for adolescents within the context of mental health services.

Item Type:	Monograph (Project Report)
Uncontrolled Keywords:	mental health, virtual assistant, artificial intelligence, Large Language Model, empathetic interaction, Unity, kesehatan mental, asisten virtual, kecerdasan buatan, Large Language Model, interaksi empatik, Unity
Subjects:	B Philosophy. Psychology. Religion > BF Psychology T Technology > T Technology (General)
Divisions:	Faculty of Intelligent Electrical and Informatics Technology (ELECTICS) > Informatics Engineering > 55201-(S1) Undergraduate Thesis
Depositing User:	Malvin Leonardo Hartanto
Date Deposited:	20 Oct 2025 06:21
Last Modified:	20 Oct 2025 06:21
URI:	http://repository.its.ac.id/id/eprint/128621

Actions (login required)

View Item