Audio and music attribution.

Melodic similarity, generator identification, and catalog-scale monitoring for music labels, publishers, and rights organizations.

• MelodySim
  Melodic-contour similarity
• Audio MIA framework
  STFT / MFCC / chroma-based membership inference
• Audio generator ID
  Distinguishes Suno, Udio, MusicLM outputs
• Synthetic provenance
  Model-collapse detection in audio
• Cross-modal MIA
  Text-prompt → audio-output linkage

Major label catalog monitoring across music generators.
CMO per-work attribution at scale (ASCAP / BMI / GEMA mechanics).
Producer detecting unauthorized sampling.
Publisher pursuing melodic-similarity claims.