AnalysisGNN: Unified Music Analysis with Graph Neural Networks
Fuente:
arXiv
Saved in:
| Main Authors: | , , , |
|---|---|
| Format: | Preprint |
| Published: |
2025
|
| Subjects: | |
| Online Access: | |
| Tags: |
Add Tag
No Tags, Be the first to tag this record!
|
| _version_ | 1866912576440369152 |
|---|---|
| author | Karystinaios, Emmanouil Hentschel, Johannes Neuwirth, Markus Widmer, Gerhard |
| author_facet | Karystinaios, Emmanouil Hentschel, Johannes Neuwirth, Markus Widmer, Gerhard |
| contents | Recent years have seen a boom in computational approaches to music analysis, yet each one is typically tailored to a specific analytical domain. In this work, we introduce AnalysisGNN, a novel graph neural network framework that leverages a data-shuffling strategy with a custom weighted multi-task loss and logit fusion between task-specific classifiers to integrate heterogeneously annotated symbolic datasets for comprehensive score analysis. We further integrate a Non-Chord-Tone prediction module, which identifies and excludes passing and non-functional notes from all tasks, thereby improving the consistency of label signals. Experimental evaluations demonstrate that AnalysisGNN achieves performance comparable to traditional static-dataset approaches, while showing increased resilience to domain shifts and annotation inconsistencies across multiple heterogeneous corpora. |
| format | Preprint |
| id |
arxiv_https___arxiv_org_abs_2509_06654 |
| institution | arXiv |
| publishDate | 2025 |
| record_format | arxiv |
| spellingShingle | AnalysisGNN: Unified Music Analysis with Graph Neural Networks Karystinaios, Emmanouil Hentschel, Johannes Neuwirth, Markus Widmer, Gerhard Sound Artificial Intelligence Recent years have seen a boom in computational approaches to music analysis, yet each one is typically tailored to a specific analytical domain. In this work, we introduce AnalysisGNN, a novel graph neural network framework that leverages a data-shuffling strategy with a custom weighted multi-task loss and logit fusion between task-specific classifiers to integrate heterogeneously annotated symbolic datasets for comprehensive score analysis. We further integrate a Non-Chord-Tone prediction module, which identifies and excludes passing and non-functional notes from all tasks, thereby improving the consistency of label signals. Experimental evaluations demonstrate that AnalysisGNN achieves performance comparable to traditional static-dataset approaches, while showing increased resilience to domain shifts and annotation inconsistencies across multiple heterogeneous corpora. |
| title | AnalysisGNN: Unified Music Analysis with Graph Neural Networks |
| topic | Sound Artificial Intelligence |
| url | https://arxiv.org/abs/2509.06654 |