Transparent Trade-offs between Properties of Explanations
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_ | 1866909697712324608 |
|---|---|
| author | Tadesse, Hiwot Belay Hüyük, Alihan Yacoby, Yaniv Pan, Weiwei Doshi-Velez, Finale |
| author_facet | Tadesse, Hiwot Belay Hüyük, Alihan Yacoby, Yaniv Pan, Weiwei Doshi-Velez, Finale |
| contents | When explaining black-box machine learning models, it's often important for explanations to have certain desirable properties. Most existing methods `encourage' desirable properties in their construction of explanations. In this work, we demonstrate that these forms of encouragement do not consistently create explanations with the properties that are supposedly being targeted. Moreover, they do not allow for any control over which properties are prioritized when different properties are at odds with each other. We propose to directly optimize explanations for desired properties. Our direct approach not only produces explanations with optimal properties more consistently but also empowers users to control trade-offs between different properties, allowing them to create explanations with exactly what is needed for a particular task. |
| format | Preprint |
| id |
arxiv_https___arxiv_org_abs_2410_23880 |
| institution | arXiv |
| publishDate | 2024 |
| record_format | arxiv |
| spellingShingle | Transparent Trade-offs between Properties of Explanations Tadesse, Hiwot Belay Hüyük, Alihan Yacoby, Yaniv Pan, Weiwei Doshi-Velez, Finale Machine Learning When explaining black-box machine learning models, it's often important for explanations to have certain desirable properties. Most existing methods `encourage' desirable properties in their construction of explanations. In this work, we demonstrate that these forms of encouragement do not consistently create explanations with the properties that are supposedly being targeted. Moreover, they do not allow for any control over which properties are prioritized when different properties are at odds with each other. We propose to directly optimize explanations for desired properties. Our direct approach not only produces explanations with optimal properties more consistently but also empowers users to control trade-offs between different properties, allowing them to create explanations with exactly what is needed for a particular task. |
| title | Transparent Trade-offs between Properties of Explanations |
| topic | Machine Learning |
| url | https://arxiv.org/abs/2410.23880 |