{"ID":6537604,"CreatedAt":"2026-07-14T02:54:43.516908796Z","UpdatedAt":"2026-07-15T03:28:55.185153975Z","DeletedAt":null,"paper_url":"https://arxiv.org/abs/2607.11334","arxiv_id":"2607.11334","title":"Verifier-Guided Twelve-Tone Composition: A Generate-Verify-Repair Harness for Symbolic Music Generation","abstract":"Large language models can produce superficially legal twelve-tone scores that collapse into degenerate textures. We introduce a neuro-symbolic harness that wraps a language-model proposer in a generate-verify-repair-trace loop with symbolic verification. The complete pipeline improves event-local consistency without claiming whole-piece legality. Across 40 controlled tasks and four paired models, audited delivery yield rises from 13.3% under raw generation to 48.1% with the harness, which explicitly abstains otherwise. The pass rate of a narrower collision and serialisation-consistency check rises from 33.5% to 58.3%, while degeneracy remains near 0.05, including under exploratory adversarial prompting. A blinded evaluation by five experts also shows a descriptive aggregate preference for harness candidates over raw generation in adherence, perceived legality, coherence, and overall quality.","short_abstract":"Large language models can produce superficially legal twelve-tone scores that collapse into degenerate textures. We introduce a neuro-symbolic harness that wraps a language-model proposer in a generate-verify-repair-trace loop with symbolic verification. The complete pipeline improves event-local consistency without cl...","url_abs":"https://arxiv.org/abs/2607.11334","url_pdf":"https://arxiv.org/pdf/2607.11334v1","authors":"[\"Congren Dai\",\"Danni Zhao\",\"Enyang Liu\",\"Michael Ching Yam\",\"Zhancheng Guo\",\"Siyi Gu\",\"Wentao Yang\",\"Bo Dai\",\"Xiaobing Li\",\"Maosong Sun\"]","published":"2026-07-13T09:52:09Z","proceeding":"cs.AI","tasks":"[\"cs.AI\"]","methods":"[\"Language Model\",\"LoRA\"]","has_code":false}