AWhat changed
Connected joint music-text representations to hierarchical audio-token generation, allowing text/melody conditioning with mostly audio-only model training and releasing the MusicCaps evaluation resource.
Scientific scope
Connected joint music-text representations to hierarchical audio-token generation, allowing text/melody conditioning with mostly audio-only model training and releasing the MusicCaps evaluation resource. Inherits AudioLM hierarchy and MuLan representation learning rather than inventing either. Original method/resource bundle is one contribution; subjective quality and multi-minute coherence are source-reported and not guarantees.
Evidence
The linked original sources describe the stated contribution and its evidence limits. WRS has not independently rerun the experiments or certified a proof. Inherits AudioLM hierarchy and MuLan representation learning rather than inventing either. Original method/resource bundle is one contribution; subjective quality and multi-minute coherence are source-reported and not guarantees.
Contributors
- Source-listed authors and technical contributors · SCIENCE: Source-listed authors or explicitly credited technical contributors for the cited versions. Family entries retain source-specific attribution; a combined list is not one original-paper byline or equal individual work. Names alone do not establish resolved identities.
- MusicLM: Generating Music From Text
- Agostinelli et al.: MusicLM text-conditioned hierarchical music generation — institutional/author source
Provisional WRS assessment, edition 2026-10-16.1. Independent expert calibration is pending. This grade does not certify correctness or assign ownership.
Full scientific record · Browse contribution grades · Request a correction