Autors: Pleshkova, S. G., Kostov K.
Title: Generative LSTM- and Transformer-Based Models Development for Guitar Melodies Creation
Keywords: Generative model, Guitar melody, LSTM, Music generation, Transformer

Abstract: This study proposes two generative models for creating guitar melodies. The proposed models are based on a previously developed approach that uses Markov chains. The Markov chain generative mechanism is replaced with LSTMand Transformer-based neural-network architectures. The proposed models' predictive accuracy and generation efficiency are compared to those of the original Markov chain model, demonstrating their ability to produce coherent and stylistically consistent chord progressions

References

  1. L. Wang, Z. Zhao, H. Liu, J. Pang, Y. Qin, and Q. Wu, "A review of intelligent music generation systems," 2023, arXiv:2211.09124,. [Online]. Available: https://doi.org/10.48550/arXiv.2211.09124
  2. J. Wu, C. Hu, Y. Wang, X. Hu, and J. Zhu, "A hierarchical recurrent neural network for symbolic melody generation," 2018, arXiv:1712.05274, [Online]. Available: https://doi.org/10.48550/arXiv.1712.05274
  3. S. Islam, H. Elmekki, A. Elsebai, J. Bentahar, N. Drawel, G. Rjoub, W. Pedrycz, "A comprehensive survey on applications of transformers for deep learning tasks," 2023, arXiv:2306.07303, [Online]. Available: https://doi.org/10.48550/arXiv.2306.07303
  4. A. Vaswani, N. Shazeer, N. Parmar, J. Uszkoreit, L. Jones, A. N. Gomez, L. Kaiser, and I. Polosukhin, "Attention is all you need," 2023, arXiv:1706.03762, [Online]. Available:. https://doi.org/10.48550/arXiv.1706.03762
  5. S. G. Pleshkova, K. Kostov, "Generative Markov Chain model development for guitar melodies creation," 33rd National Conference with International Participation (TELECOM), pp. 1-4, 2025, DOI:10.1109/TELECOM66943.2025.11304051
  6. S. Kantarelis, K. Thomas, V. Lyberatos, E. Dervakos, and G. Stamou, "CHORDONOMICON: A dataset of 666,000 Songs and their chord progressions, 2024, arXiv:2410.22046, [Online]. Available:. https://doi.org/10.48550/arXiv.2410.22046
  7. F. Petersen, H. Kuehne, C. Borgelt, and O. Deussen, "Differentiable top-k classification learning," 2022, arXiv:2206.07290, [Online]. Available: https://doi.org/10.48550/arXiv.2206.07290
  8. J. Li, H. Liu, N. Yan, L. Wang, "Enhanced Memory Network: The novel network structure for Symbolic Music Generation," 2021, arXiv:2110.03392 [Online]. Available:https://doi.org/10.48550/arXiv.2110.03392
  9. W. Zheng, P. Zhao, K. Huang, G. Chen, "Understanding the Property of Long Term Memory for the LSTM with Attention Mechanism," Proceedings of the 30th ACM International Conference on Information & Knowledge Management, CIKM '21, pp. 2708-2717, 2021, DOI:10.1145/3459637.3482399
  10. S. Hager, K. Hablutzel, and K. M. Kinnaird, "Generating Music with Structure Using Self-Similarity as Attention," 2024, arXiv:2406.15647, [Online]. Available: https://doi.org/10.48550/arXiv.2406.15647
  11. G. Keerti, A. N. Vaishnavi, P. Mukherjee, A. S. Vidya, G. S. Sreenithya, D. Nayab, "Attentional networks for music generation," 2020, arXiv:2002.03854, [Online]. Available: https://doi.org/10.48550/arXiv.2002.03854
  12. N. Srivastava, G. Hinton, A. Krizhevsky, I. Sutskever, R. Salakhutdinov, "Dropout: A Simple Way to Prevent Neural Networks from Overfitting"
  13. I. Borovik, "PianoCoRe: Combined and Refined Piano MIDI Dataset," May 7 2026, [Online]. Available: https://doi.org/10.5334/tismir.333

Issue

2026 61st International Scientific Conference on Information, Communication and Energy Systems and Technologies, ICEST 2026 - Proceedings of Papers, pp. 483-489, 2026, Albania, https://doi.org/10.1109/ICEST71230.2026.11623552

Вид: публикация в международен форум, публикация в реферирано издание, индексирана в Scopus