New method quantifies AI-generated stems in music mixtures

Researchers have developed a new method to quantify the proportion of AI-generated stems in hybrid music mixtures, moving beyond binary detection. By reformulating AI music detection as a regression problem, they trained a CNN model to estimate an "AI energy ratio" (alpha) between 0 and 1. This approach achieved a Mean Absolute Error of 0.076 and an R^2 of 0.85 on test mixtures, indicating promise for detecting AI contributions in realistic music production. AI

IMPACT This research could lead to more nuanced AI music detection tools, impacting copyright and authenticity in music production.

RANK_REASON Academic paper detailing a new methodology and model for AI music detection. [lever_c_demoted from research: ic=1 ai=1.0]

Read on arXiv cs.AI →

• alphaXiv
• arXiv
• CatalyzeX
• CNN
• DagsHub
• Fernando Garcia De La Cruz
• Hugging Face

• paper
• product

AI-generated summary · Google Gemini · from 1 sources. How we write summaries →