Abstract:We introduce a toolkit for uncovering spurious correlations between recording characteristics and target class in speech datasets. Spurious correlations may arise due to heterogeneous recording conditions, a common scenario for health-related datasets. When present both in the training and test data, these correlations result in an overestimation of the system performance -- a dangerous situation, specially in high-stakes application where systems are required to satisfy minimum performance requirements. Our toolkit implements a diagnostic method based on the detection of the target class using only the non-speech regions in the audio. Better than chance performance at this task indicates that information about the target class can be extracted from the non-speech regions, flagging the presence of spurious correlations. The toolkit is publicly available for research use.
| Subjects: | Sound (cs.SD); Artificial Intelligence (cs.AI); Databases (cs.DB) |
| Cite as: | arXiv:2604.26676 [cs.SD] |
| (or arXiv:2604.26676v1 [cs.SD] for this version) | |
| https://doi.org/10.48550/arXiv.2604.26676 arXiv-issued DOI via DataCite (pending registration) |
Submission history
From: Lara Gauder [view email]
[v1]
Wed, 29 Apr 2026 13:47:22 UTC (152 KB)