arXiv:2607. 12857v1 Announce Type: cross Abstract: A generated rhythm-game chart need not reproduce one official note sequence: many note choices can fit the same song and difficulty.
Summary generated by The Flow from the publisher's feed. The full article lives at arXiv AI.
A generated rhythm-game chart need not reproduce one official note sequence: many note choices can fit the same song and difficulty. Reference-note agreement therefore measures reconstruction, not the full design problem.
arXiv:2607. 11334v1 Announce Type: new Abstract: Large language models can produce superficially legal twelve-tone scores that collapse into degenerate textures.
arXiv:2607. 13587v1 Announce Type: cross Abstract: Automatic symbolic music analysis has made substantial progress, yet existing systems are typically designed for a single mode of use, such as full-score prediction, and therefore do not match the broader range of operations that arise in analysis workflows, including partial completion, local correction, and iterative refinement.
arXiv:2606. 17006v1 Announce Type: cross Abstract: We introduce TuneJury, an open, instance-level pairwise reward model for text-to-music that predicts a music preference score from a text prompt and an audio clip.
arXiv:2604. 16009v2 Announce Type: replace Abstract: Most large language model benchmarks evaluate final-answer quality but reveal little about how models revise beliefs under disagreement or conflicting evidence.
arXiv:2608. 14916v1 Announce Type: cross Abstract: AI-generated music detectors are commonly evaluated against original songs, but real-world uploads are often remixed, re-encoded, pitch-shifted, or otherwise edited.