Research map / Large language models / Agents, code and math
Is Symbolic Music a Specific Language? Exploring Inspiration-to-Structure Machine Composition via LLMs
Zhejing Hu, Yan Liu, Zhi Zhang, Aiwei Zhang, Sheng-hua Zhong, Bruce X.B. Yu, Gong Chen
Read the paperCheck your idea against 137,217 papers
AI-extracted. The measured effects below were read by AI from the paper's full text (no summary yet). Numbers and claims can be wrong or missing context, so check the paper before you cite or build on it.
Measured effects reported in the paper
| What was changed | Where | Measured | Effect |
|---|---|---|---|
| add Structured Triplet Data (STD) fine-tuning vs plain SFT (ablation) | Llama3.2-3B · Ablation Table 3 | PD / DD / CD: 0.60 -> 0.64 PD, 0.61 -> 0.75 DD, 0.61 -> 0.76 CD | Improved |
| add intra-song contrastive (DiSCO) to STD (ablation) | Llama3.2-3B · Ablation Table 3 | PD / DD / CD: 0.64 -> 0.68 PD, 0.75 -> 0.79 DD, 0.76 -> 0.79 CD | Improved |
| add inter-song contrastive (DiSCO) to STD (ablation) | Llama3.2-3B · Ablation Table 3 | PD / DD / CD: 0.64 -> 0.68 PD, 0.75 -> 0.80 DD, 0.76 -> 0.80 CD | Improved |
| add both intra- and inter-song contrastive losses (ablation) | Llama3.2-3B · Ablation Table 3 | PD / DD / CD: 0.68 -> 0.74 PD, 0.80 -> 0.83 DD, 0.80 -> 0.87 CD | Improved |
| add IoS (STD + DiSCO) instead of SFT | Llama3.2-3B · Human evaluation | overall rating: 3.00 -> 3.66 | Improved |
| add IoS instead of SFT | Qwen2.5-3B · Human evaluation | overall rating: 2.89 -> 3.57 | Improved |
| add IoS instead of SFT | Llama3.2-3B · Human evaluation | structure rating: 2.96 -> 3.93 | Improved |
| add IoS instead of SFT | Qwen2.5-0.5B · POP909 objective | CD: 0.67 -> 0.87 | Improved |
| add IoS instead of SFT | Llama3.2-1B · POP909 objective | CD: 0.63 -> 0.82 | Improved |
| add IoS to small Llama3.2-3B vs larger GPT-4o | Llama3.2-3B +IoS · Human evaluation | overall rating: 3.09 -> 3.66 | Improved |
| add IoS training | LLMs · multi-turn interaction | PD / DD | Improved |
| use ABC notation instead of MIDI tokens | music representation · token length | average tokens: 5,803 -> 3,229 | Improved |
Closest papers by meaning
- Responding to the Call: Exploring Automatic Music Composition Using a Knowledge-Enhanced Model (AAAI 2024 · similarity 0.89)
- MuPT: A Generative Symbolic Music Pretrained Transformer (ICLR 2025 · similarity 0.89)
- Whole-Song Hierarchical Generation of Symbolic Music Using Cascaded Diffusion Models (ICLR 2024 · similarity 0.88)
- LLark: A Multimodal Instruction-Following Language Model for Music (ICML 2024 · similarity 0.87)
- Efficient Fine-Grained Guidance for Diffusion Model Based Symbolic Music Generation (ICML 2025 · similarity 0.87)
- From Isolated feature to Orbits:\\Discovering Music Concepts via Multi-SAE Alignment (NeurIPS 2026 · similarity 0.86)
Writing a paper in this area? scime draft is a free LaTeX editor with AI included, real-time co-authors and a check of your references and venue rules before you submit. It also imports OpenAI Prism projects.
