Enregistré dans:
| Auteurs principaux: | , , , , |
|---|---|
| Format: | Preprint |
| Publié: |
2026
|
| Sujets: | |
| Accès en ligne: | https://arxiv.org/abs/2603.27877 |
| Tags: |
Ajouter un tag
Pas de tags, Soyez le premier à ajouter un tag!
|
| _version_ | 1866912987758985216 |
|---|---|
| author | Weck, Benno Puentes, Pablo Poltronieri, Andrea Prabhu, Satyajeet Bogdanov, Dmitry |
| author_facet | Weck, Benno Puentes, Pablo Poltronieri, Andrea Prabhu, Satyajeet Bogdanov, Dmitry |
| contents | The evaluation of music understanding in Large Audio-Language Models (LALMs) requires a rigorously defined benchmark that truly tests whether models can perceive and interpret music, a standard that current data methodologies frequently fail to meet. This paper introduces a meticulously structured approach to music evaluation, proposing a new dataset of 320 hand-written questions curated and validated by experts with musical training, arguing that such focused, manual curation is superior for probing complex audio comprehension. To demonstrate the use of the dataset, we benchmark six state-of-the-art LALMs and additionally test their robustness to uni-modal shortcuts. |
| format | Preprint |
| id |
arxiv_https___arxiv_org_abs_2603_27877 |
| institution | arXiv |
| publishDate | 2026 |
| record_format | arxiv |
| spellingShingle | HumMusQA: A Human-written Music Understanding QA Benchmark Dataset Weck, Benno Puentes, Pablo Poltronieri, Andrea Prabhu, Satyajeet Bogdanov, Dmitry Computation and Language Sound The evaluation of music understanding in Large Audio-Language Models (LALMs) requires a rigorously defined benchmark that truly tests whether models can perceive and interpret music, a standard that current data methodologies frequently fail to meet. This paper introduces a meticulously structured approach to music evaluation, proposing a new dataset of 320 hand-written questions curated and validated by experts with musical training, arguing that such focused, manual curation is superior for probing complex audio comprehension. To demonstrate the use of the dataset, we benchmark six state-of-the-art LALMs and additionally test their robustness to uni-modal shortcuts. |
| title | HumMusQA: A Human-written Music Understanding QA Benchmark Dataset |
| topic | Computation and Language Sound |
| url | https://arxiv.org/abs/2603.27877 |