Representation-Based Data Quality Audits for Audio
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_ | 1866918151649755136 |
|---|---|
| author | Gonzalez-Jimenez, Alvaro Gröger, Fabian Wermelinger, Linda Bürli, Andrin Kastanis, Iason Lionetti, Simone Pouly, Marc |
| author_facet | Gonzalez-Jimenez, Alvaro Gröger, Fabian Wermelinger, Linda Bürli, Andrin Kastanis, Iason Lionetti, Simone Pouly, Marc |
| contents | Data quality issues such as off-topic samples, near duplicates, and label errors often limit the performance of audio-based systems. This paper addresses these issues by adapting SelfClean, a representation-to-rank data auditing framework, from the image to the audio domain. This approach leverages self-supervised audio representations to identify common data quality issues, creating ranked review lists that surface distinct issues within a single, unified process. The method is benchmarked on the ESC-50, GTZAN, and a proprietary industrial dataset, using both synthetic and naturally occurring corruptions. The results demonstrate that this framework achieves state-of-the-art ranking performance, often outperforming issue-specific baselines and enabling significant annotation savings by efficiently guiding human review. |
| format | Preprint |
| id |
arxiv_https___arxiv_org_abs_2509_26291 |
| institution | arXiv |
| publishDate | 2025 |
| record_format | arxiv |
| spellingShingle | Representation-Based Data Quality Audits for Audio Gonzalez-Jimenez, Alvaro Gröger, Fabian Wermelinger, Linda Bürli, Andrin Kastanis, Iason Lionetti, Simone Pouly, Marc Sound Artificial Intelligence Machine Learning Data quality issues such as off-topic samples, near duplicates, and label errors often limit the performance of audio-based systems. This paper addresses these issues by adapting SelfClean, a representation-to-rank data auditing framework, from the image to the audio domain. This approach leverages self-supervised audio representations to identify common data quality issues, creating ranked review lists that surface distinct issues within a single, unified process. The method is benchmarked on the ESC-50, GTZAN, and a proprietary industrial dataset, using both synthetic and naturally occurring corruptions. The results demonstrate that this framework achieves state-of-the-art ranking performance, often outperforming issue-specific baselines and enabling significant annotation savings by efficiently guiding human review. |
| title | Representation-Based Data Quality Audits for Audio |
| topic | Sound Artificial Intelligence Machine Learning |
| url | https://arxiv.org/abs/2509.26291 |