A Contrastive Learning Approach to Mitigate Bias in Speech Models
Fuente:
arXiv
Saved in:
| Main Authors: | , , , |
|---|---|
| Format: | Preprint |
| Published: |
2024
|
| Subjects: | |
| Online Access: | |
| Tags: |
Add Tag
No Tags, Be the first to tag this record!
|
| _version_ | 1866913499914960896 |
|---|---|
| author | Koudounas, Alkis Giobergia, Flavio Pastor, Eliana Baralis, Elena |
| author_facet | Koudounas, Alkis Giobergia, Flavio Pastor, Eliana Baralis, Elena |
| contents | Speech models may be affected by performance imbalance in different population subgroups, raising concerns about fair treatment across these groups. Prior attempts to mitigate unfairness either focus on user-defined subgroups, potentially overlooking other affected subgroups, or do not explicitly improve the internal representation at the subgroup level. This paper proposes the first adoption of contrastive learning to mitigate speech model bias in underperforming subgroups. We employ a three-level learning technique that guides the model in focusing on different scopes for the contrastive loss, i.e., task, subgroup, and the errors within subgroups. The experiments on two spoken language understanding datasets and two languages demonstrate that our approach improves internal subgroup representations, thus reducing model bias and enhancing performance. |
| format | Preprint |
| id |
arxiv_https___arxiv_org_abs_2406_14686 |
| institution | arXiv |
| publishDate | 2024 |
| record_format | arxiv |
| spellingShingle | A Contrastive Learning Approach to Mitigate Bias in Speech Models Koudounas, Alkis Giobergia, Flavio Pastor, Eliana Baralis, Elena Computation and Language Machine Learning Audio and Speech Processing Speech models may be affected by performance imbalance in different population subgroups, raising concerns about fair treatment across these groups. Prior attempts to mitigate unfairness either focus on user-defined subgroups, potentially overlooking other affected subgroups, or do not explicitly improve the internal representation at the subgroup level. This paper proposes the first adoption of contrastive learning to mitigate speech model bias in underperforming subgroups. We employ a three-level learning technique that guides the model in focusing on different scopes for the contrastive loss, i.e., task, subgroup, and the errors within subgroups. The experiments on two spoken language understanding datasets and two languages demonstrate that our approach improves internal subgroup representations, thus reducing model bias and enhancing performance. |
| title | A Contrastive Learning Approach to Mitigate Bias in Speech Models |
| topic | Computation and Language Machine Learning Audio and Speech Processing |
| url | https://arxiv.org/abs/2406.14686 |