Implementasi Model Bahasa Openai GPT-3 Speech Recognition untuk Aplikasi Pendukung Pembelajaran Produktif TKJ di SMK Telkom Purwokerto
Abstract
This research explores the integration of the OpenAI GPT-3 language model with speech recognition technology to enhance network computer engineering education at SMK Telkom Purwokerto. Four classes of students were involved in this initiative. The developed application successfully captured voice input, converted it to text, and used GPT-3 to generate responses, enabling sophisticated voice-based interaction. The research findings highlight the potential of this integration in developing interactive applications that understand and respond in natural language, bringing significant changes in education. Student perceptions from four classes (A, B, C, D) showed the majority found the material very helpful, with a 70.27% increase in computer and network engineering skills. This is in line with the objectives of Merdeka Belajar Curriculum in Indonesia. Recommendations include improving voice recognition accuracy, optimizing the GPT-3 API, handling privacy, developing contextualized learning content, and training for teachers and students, as well as periodic evaluation and user feedback for continuous improvement.
Keywords: OpenAI GPT-3; Speech Recognition; Kurikulum Merdeka Belajar; Voice-Based Interaction; Technology Integration
Abstrak
Penelitian ini mengeksplorasi integrasi model bahasa OpenAI GPT-3 dengan teknologi pengenalan suara untuk meningkatkan pendidikan teknik komputer jaringan di SMK Telkom Purwokerto. Empat kelas siswa terlibat dalam inisiatif ini. Aplikasi yang dikembangkan berhasil menangkap input suara, mengonversinya menjadi teks, dan menggunakan GPT-3 untuk menghasilkan respons, memungkinkan interaksi berbasis suara yang canggih. Temuan penelitian menyoroti potensi integrasi ini dalam mengembangkan aplikasi interaktif yang memahami dan merespons dalam bahasa alami, membawa perubahan signifikan dalam pendidikan. Persepsi siswa dari empat kelas (A, B, C, D) menunjukkan mayoritas merasa materi sangat membantu, dengan peningkatan 70,27% dalam kemampuan teknik komputer dan jaringan. Ini sejalan dengan tujuan Kurikulum Merdeka Belajar di Indonesia. Rekomendasi mencakup peningkatan akurasi pengenalan suara, optimalisasi API GPT-3, penanganan privasi, pengembangan konten pembelajaran yang kontekstual, dan pelatihan bagi guru dan siswa, serta evaluasi berkala dan umpan balik pengguna untuk perbaikan berkelanjutan.
Keywords
References
N. Amalia, "Implementasi Kurikulum Merdeka," 2023.
N. A. Ismawati, "Pengaruh Perbedaan Kemampuan Siswa dalam Kurikulum Merdeka," 2022.
N. A. Ismawati, "Tantangan Pembelajaran Terdiferensiasi dalam Kurikulum Merdeka," 2022.
Bendriyanti, "Efektivitas Pembelajaran Terdiferensiasi," 2022.
Shahriar, "Penggunaan AI dan NLP dalam Kurikulum Merdeka Belajar," 2023.
R. Sharma, "Keuntungan Penggunaan Speech Recognition dalam Pembelajaran," 2018.
N. Amalia, "Adaptasi Kurikulum dengan Perkembangan Zaman," 2023.
N. A. Ismawati, "Peran Pengajar dalam Kurikulum Merdeka," 2022.
R. Southwell, "Challenges and Feasibility of Automatic Speech Recognition for Modeling Student Collaborative Discourse in Classrooms," 2022.
Z. Meng, "Domain Adaptation via Teacher-Student Learning For End-To-End Speech Recognition," 2019.
R. Zubaedah, "Edukasi Belajar Membaca Menggunakan Teknologi Speech Recognition," 2023.
K. Nugroho, "Implementasi Sistem Speech To Text Berbasis Android Menggunakan App Inventor Speech Recognizer," 2019.
Z. Min, "Exploring the Integration of Large Language Models into Automatic Speech Recognition Systems: An Empirical Study," 2019.
R. Southwell, "Automatic Speech Recognition Tuned For Child Speech In The Classroom," 2022.
P. Gupta, "Hate Speech Detection using OpenAI and GPT-3," 2022.
F. Sandy, "Impelentasi Penggunaan Kecerdasan Buatan Dalam Pendidikan Tinggi," 2023.
I. Tejawiani, "Peran Artificial Intelligence Terhadap Peningkatan Kreativitas Siswa Dengan Menerapkan Proyek Penguatan Profil Pelajar Pancasila," 2023.
How To Cite This :
Refbacks
- There are currently no refbacks.
