Toward Creative Autonomy: A Dual-Model Framework for Assessing Originality in Generative Music Systems
DOI:
https://doi.org/10.61978/harmonia.v3i1.1112Keywords:
AI Music Generation, Creativity, Memorization, Originality Metrics, MUSHRA, Symbolic Evaluation, Anti-Memorization GuidanceAbstract
AI-generated music systems such as MusicGen and Stable Audio 2.0 are increasingly capable of producing stylistically coherent and musically rich compositions. However, questions remain about whether these outputs constitute genuine creativity or mere replication of training data. This study evaluates the memorization and creativity levels of these models using symbolic and audio-based metrics, alongside perceptual assessments.
A dual-model evaluation was conducted: symbolic outputs were assessed using chroma-based DTW, Smith–Waterman, melodic n-grams, and MGEval metrics, while audio outputs were analyzed for waveform similarity and listener ratings. Anti-Memorization Guidance (AMG) was introduced to reduce overfitting, with 50 outputs generated per model under both standard and AMG conditions.
Results showed significant memorization in standard outputs, particularly with high Replication Index scores and latent similarity clusters. AMG effectively lowered memorization and increased Novelty Scores and Harmonic Surprise. Subjective tests using MUSHRA and Likert-style ratings revealed that AMG-enhanced outputs were perceived as more creative but slightly less typical in genre. Correlations between objective and subjective metrics further validated the effectiveness of the hybrid evaluation framework.
The study concludes that AI music systems can be guided toward greater originality using anti-memorization strategies. While achieving historical creativity remains challenging, perceptually and statistically creative outputs are attainable. This framework offers a replicable approach for evaluating creativity and informs ethical, legal, and design considerations in AI music generation.
References
Anantrasirichai, N., & Bull, D. (2021). Artificial Intelligence in the Creative Industries: A Review. Artificial Intelligence Review, 55(1), 589–656. https://doi.org/10.1007/s10462-021-10039-7 DOI: https://doi.org/10.1007/s10462-021-10039-7
Batlle-Roca, R., Gómez, E., Liao, W., Serra, X., & Mitsufuji, Y. (2023). Transparency in Music-Generative AI: A Systematic Literature Review. https://doi.org/10.21203/rs.3.rs-3708077/v1 DOI: https://doi.org/10.21203/rs.3.rs-3708077/v1
Berardinis, J. d., Meroño-Peñuela, A., Poltronieri, A., & Presutti, V. (2023). The Harmonic Memory: A Knowledge Graph of Harmonic Patterns as a Trustworthy Framework for Computational Creativity. 3873–3882. https://doi.org/10.1145/3543507.3587428 DOI: https://doi.org/10.1145/3543507.3587428
Bhattacharya, J., Ζιώγα, Ι., & Lewis, R. J. (2017). Novel or Consistent Music? An Electrophysiological Study Investigating Music Use in Advertising. Journal of Neuroscience Psychology and Economics, 10(4), 137–152. https://doi.org/10.1037/npe0000080 DOI: https://doi.org/10.1037/npe0000080
Biady, Y., Lee, T., Pham, L., Patanwala, A. E., Poon, S., Ritchie, A., Burke, R., & Penm, J. (2024). Factors Influencing Health Care Professionals’ Perceptions of Frequent Drug–Drug Interaction Alerts. Aci Open, 08(01), e25–e32. https://doi.org/10.1055/s-0044-1782534 DOI: https://doi.org/10.1055/s-0044-1782534
Briot, J.-P., Hadjeres, G., & Pachet, F.-D. (2017). Deep Learning Techniques for Music Generation—A Survey. https://doi.org/10.48550/arxiv.1709.01620
Cádiz, R. F., Macaya, A., Cartagena, M., & Parra, D. (2021). Creativity in Generative Musical Networks: Evidence From Two Case Studies. Frontiers in Robotics and Ai, 8. https://doi.org/10.3389/frobt.2021.680586 DOI: https://doi.org/10.3389/frobt.2021.680586
Carnovalini, F., & Rodà, A. (2020). Computational Creativity and Music Generation Systems: An Introduction to the State of the Art. Frontiers in Artificial Intelligence, 3. https://doi.org/10.3389/frai.2020.00014 DOI: https://doi.org/10.3389/frai.2020.00014
Creely, E., & Blannin, J. (2023). The Implications of Generative AI for Creative Composition in Higher Education and Initial Teacher Education. Ascilite Publications, 357–361. https://doi.org/10.14742/apubs.2023.618 DOI: https://doi.org/10.14742/apubs.2023.618
Dong, H., Chen, K., Dubnov, S., McAuley, J., & Berg-Kirkpatrick, T. (2023). Multitrack Music Transformer. https://doi.org/10.1109/icassp49357.2023.10094628 DOI: https://doi.org/10.1109/ICASSP49357.2023.10094628
Dong, H. K., Zhou, C., Berg-Kirkpatrick, T., & McAuley, J. (2022). Deep Performer: Score-to-Audio Music Performance Synthesis. 951–955. https://doi.org/10.1109/icassp43922.2022.9747217 DOI: https://doi.org/10.1109/ICASSP43922.2022.9747217
Eerola, T., & Peltola, H. (2016). Memorable Experiences With Sad Music—Reasons, Reactions and Mechanisms of Three Types of Experiences. Plos One, 11(6), e0157444. https://doi.org/10.1371/journal.pone.0157444 DOI: https://doi.org/10.1371/journal.pone.0157444
Esposti, M. D., Lagioia, F., & Sartor, G. (2019). The Use of Copyrighted Works by AI Systems: Art Works in the Data Mill. European Journal of Risk Regulation, 11(1), 51–69. https://doi.org/10.1017/err.2019.56 DOI: https://doi.org/10.1017/err.2019.56
Gabbolini, G., Hennequin, R., & Epure, E. V. (2022). Data-Efficient Playlist Captioning With Musical and Linguistic Knowledge. 11401–11415. https://doi.org/10.18653/v1/2022.emnlp-main.784 DOI: https://doi.org/10.18653/v1/2022.emnlp-main.784
Gee, C. S., Ramly, Z., & Zulkhairi, M. (2021). Value-Added Tax and Economic Efficiency: Role of Country Governance. Panoeconomicus, 68(3), 325–358. https://doi.org/10.2298/pan180201020c DOI: https://doi.org/10.2298/PAN180201020C
Gifford, T., Knotts, S., McCormack, J., Kalonaris, S., Yee-King, M., & d’Inverno, M. (2018). Computational Systems for Music Improvisation. Digital Creativity, 29(1), 19–36. https://doi.org/10.1080/14626268.2018.1426613 DOI: https://doi.org/10.1080/14626268.2018.1426613
Gordon, S., Mahari, R., Mishra, M., & Epstein, Z. (2022). Co-Creation and Ownership for AI Radio. https://doi.org/10.48550/arxiv.2206.00485
Hernandez-Olivan, C., Puyuelo, J. A., & Beltran, J. R. (2022). Subjective Evaluation of Deep Learning Models for Symbolic Music Composition. https://doi.org/10.48550/arxiv.2203.14641
Hong, J.-W., Peng, Q., & Williams, D. (2020). Are You Ready for Artificial Mozart and Skrillex? An Experiment Testing Expectancy Violation Theory and AI Music. New Media & Society, 23(7), 1920–1935. https://doi.org/10.1177/1461444820925798 DOI: https://doi.org/10.1177/1461444820925798
Jiang, F., Zhang, L., Wang, K., Deng, X., & Yang, W. (2022). BoYaTCN: Research on Music Generation of Traditional Chinese Pentatonic Scale Based on Bidirectional Octave Your Attention Temporal Convolutional Network. Applied Sciences, 12(18), 9309. https://doi.org/10.3390/app12189309 DOI: https://doi.org/10.3390/app12189309
Jo, S.-Y., & Jeong, J.-W. (2020). Prediction of Visual Memorability With EEG Signals: A Comparative Study. Sensors, 20(9), 2694. https://doi.org/10.3390/s20092694 DOI: https://doi.org/10.3390/s20092694
Jordanous, A. (2016). Four PPPPerspectives on Computational Creativity in Theory and in Practice. Connection Science, 28(2), 194–216. https://doi.org/10.1080/09540091.2016.1151860 DOI: https://doi.org/10.1080/09540091.2016.1151860
Kaşif, A., & Sevgen, S. (2024). Classical Turkish Music Composition With LSTM Self-Attention. Journal of Innovative Science and Engineering (Jise). https://doi.org/10.38088/jise.1406162 DOI: https://doi.org/10.38088/jise.1406162
Kharlashkin, L. (2024). Enhancing Multi-Dimensional Music Generation by an LLM-based Data Augmentation Technique. https://doi.org/10.31237/osf.io/9exyu DOI: https://doi.org/10.31237/osf.io/9exyu
Kovalkov, A., Paasen, B., Segal, A., Pinkwart, N., & Gal, K. (2021). Automatic Creativity Measurement in Scratch Programs Across Modalities. Ieee Transactions on Learning Technologies, 14(6), 740–753. https://doi.org/10.1109/tlt.2022.3144442 DOI: https://doi.org/10.1109/TLT.2022.3144442
Lam, J. T. (2024). Analyzing Principles and Applications of Machine Learning in Music: Emotion Music Generation, and Style Modeling. Applied and Computational Engineering, 68(1), 177–182. https://doi.org/10.54254/2755-2721/68/20241432 DOI: https://doi.org/10.54254/2755-2721/68/20241432
Mazzone, M., & Elgammal, A. (2019). Art, Creativity, and the Potential of Artificial Intelligence. Arts, 8(1), 26. https://doi.org/10.3390/arts8010026 DOI: https://doi.org/10.3390/arts8010026
Moruzzi, C. (2020). Measuring Creativity: An Account of Natural and Artificial Creativity. European Journal for Philosophy of Science, 11(1). https://doi.org/10.1007/s13194-020-00313-w DOI: https://doi.org/10.1007/s13194-020-00313-w
Rosalina, R., & Sahuri, G. (2024). MIDI-based Generative Neural Networks With Variational Autoencoders for Innovative Music Creation. International Journal of Advances in Applied Sciences, 13(2), 360. https://doi.org/10.11591/ijaas.v13.i2.pp360-370 DOI: https://doi.org/10.11591/ijaas.v13.i2.pp360-370
Sturm, B. L., Déguernel, K., Huang, R. S., Holzapfel, A., Bown, O., Collins, N., Sterne, J., Vila, L. C., Casini, L., Cabrera, D. A., Drott, E., & Ben‐Tal, O. (2024). MusAIcology: AI Music and the Need for a New Kind of Music Studies. https://doi.org/10.31235/osf.io/9pz4x DOI: https://doi.org/10.31235/osf.io/9pz4x
Xu, L. (2024). A Study on the Fair Use Principles of Artificial Intelligence Generated Music. Lecture Notes in Education Psychology and Public Media, 34(1), 228–235. https://doi.org/10.54254/2753-7048/34/20231932 DOI: https://doi.org/10.54254/2753-7048/34/20231932
Yin, Z., Reuben, F., Stepney, S., & Collins, T. (2022). Measuring When a Music Generation Algorithm Copies Too Much: The Originality Report, Cardinality Score, and Symbolic Fingerprinting by Geometric Hashing. Sn Computer Science, 3(5). https://doi.org/10.1007/s42979-022-01220-y DOI: https://doi.org/10.1007/s42979-022-01220-y
Zacharakis, A., Καλιακάτσος-Παπακώστας, Μ., Kalaitzidou, S., & Cambouropoulos, E. (2021). Evaluating Human-Computer Co-Creative Processes in Music: A Case Study on the CHAMELEON Melodic Harmonizer. Frontiers in Psychology, 12. https://doi.org/10.3389/fpsyg.2021.603752 DOI: https://doi.org/10.3389/fpsyg.2021.603752
Zhang, J. D., Schubert, E., & McPherson, G. E. (2020). Aspects of Music Performance That Are Most Highly Related to Musical Sophistication. Psychomusicology Music Mind and Brain, 30(2), 64–71. https://doi.org/10.1037/pmu0000252 DOI: https://doi.org/10.1037/pmu0000252
Zukowski, Z., & Carr, C. (2018). Generating Black Metal and Math Rock: Beyond Bach, Beethoven, and Beatles. https://doi.org/10.48550/arxiv.1811.06639
