Aria : A Diagnostic Framework for Music Training Data Attribution
Changheon Han, Ashkan Panahi, Kıvanç Tatar · 2026
Read Aria : A Diagnostic Framework for Music Training Data Attribution by Changheon Han, Ashkan Panahi, Kıvanç Tatar (2026) in the Tower of Babel, a free online library of public-domain works preserved in clean, searchable text. No sign-up. The full text is served as a plain-text file at a predictable URL, where text is present: https://antarctic-labs.com/tower-of-babel/arxiv-2605-16181.txt.
Description
Han, Panahi, and Tatar propose ARIA, a framework that attributes music-generation outputs to training songs along specific musical aspects, not just a single influence score. It decomposes attribution into five aspects for symbolic music and three for audio, paired with reliability diagnostics from the segment-level score matrix. On a symbolic model with counterfactual ground truth, the diagnostics rank four attribution methods identically to that truth. On audio models, ARIA flags score matrices whose retrieved tracks barely vary across queries.
Catalog record
| CREATOR | Changheon Han, Ashkan Panahi, Kıvanç Tatar |
|---|---|
| YEAR | 2026 |
| COLLECTION | PAPERS |
| FORMAT | TXT |
| FILE SIZE | 60084 |
| LICENSE | http://creativecommons.org/licenses/by/4.0/ |
| RIGHTS | CC BY |
| SOURCE | arXiv |
Source: https://arxiv.org/abs/2605.16181