Large-Scale Training Data Attribution for Music Generative Models via Unlearning

2025-06-23

AI Summary

This study addresses the copyright and attribution challenges in AI music generation arising from ambiguous training data provenance. We propose the first data attribution method tailored to large-scale music generation models. Methodologically, we introduce inverse learning (unlearning) to the music generation domain, integrating it with diffusion model architectures and cross-sample similarity analysis; attribution consistency is optimized via hyperparameter grid search. Experiments on real-world music generation models demonstrate feasibility: our approach yields significantly higher attribution consistency and interpretability compared to conventional similarity-based baselines, while enabling verifiable generation provenance tracing. This work establishes a practical technical foundation and methodological framework for artistic contribution assessment, copyright liability allocation, and ethical governance in AI-assisted music creation.

Technology Category

• Machine Learning: Adversarial Learning & Robustness
• Natural Language Processing: Generation
• Computer Vision: Diffusion Models for Vision

Application Category

• Web Mining and Content Analysis: Web data provenance, reliability, and authenticity
• Security and Privacy: Data transparency and provenance
• Economics, Online Markets and Human Computation: Trust and reliance of crowd workers and data experts on GenAI

Abstract

This paper explores the use of unlearning methods for training data attribution (TDA) in music generative models trained on large-scale datasets. TDA aims to identify which specific training data points contributed to the generation of a particular output from a specific model. This is crucial in the context of AI-generated music, where proper recognition and credit for original artists are generally overlooked. By enabling white-box attribution, our work supports a fairer system for acknowledging artistic contributions and addresses pressing concerns related to AI ethics and copyright. We apply unlearning-based attribution to a text-to-music diffusion model trained on a large-scale dataset and investigate its feasibility and behavior in this setting. To validate the method, we perform a grid search over different hyperparameter configurations and quantitatively evaluate the consistency of the unlearning approach. We then compare attribution patterns from unlearning with those from a similarity-based approach. Our findings suggest that unlearning-based approaches can be effectively adapted to music generative models, introducing large-scale TDA to this domain and paving the way for more ethical and accountable AI systems for music creation.

Problem

• Identify training data influencing AI-generated music outputs
• Enable fair attribution for original artists in AI music
• Adapt unlearning methods for large-scale music data attribution

Innovation

• Unlearning methods for training data attribution
• White-box attribution in music generative models
• Grid search for hyperparameter validation