Saved in:
| Main Authors: | , |
|---|---|
| Format: | Preprint |
| Published: |
2024
|
| Subjects: | |
| Online Access: | https://arxiv.org/abs/2406.20046 |
| Tags: |
Add Tag
No Tags, Be the first to tag this record!
|
| _version_ | 1866913407919194112 |
|---|---|
| author | Sikar, Daniel Garcez, Artur |
| author_facet | Sikar, Daniel Garcez, Artur |
| contents | We posit that data can only be safe to use up to a certain threshold of the data distribution shift, after which control must be relinquished by the autonomous system and operation halted or handed to a human operator. With the use of a computer vision toy example we demonstrate that network predictive accuracy is impacted by data distribution shifts and propose distance metrics between training and testing data to define safe operation limits within said shifts. We conclude that beyond an empirically obtained threshold of the data distribution shift, it is unreasonable to expect network predictive accuracy not to degrade |
| format | Preprint |
| id |
arxiv_https___arxiv_org_abs_2406_20046 |
| institution | arXiv |
| publishDate | 2024 |
| record_format | arxiv |
| spellingShingle | Evaluation of autonomous systems under data distribution shifts Sikar, Daniel Garcez, Artur Machine Learning I.4.m We posit that data can only be safe to use up to a certain threshold of the data distribution shift, after which control must be relinquished by the autonomous system and operation halted or handed to a human operator. With the use of a computer vision toy example we demonstrate that network predictive accuracy is impacted by data distribution shifts and propose distance metrics between training and testing data to define safe operation limits within said shifts. We conclude that beyond an empirically obtained threshold of the data distribution shift, it is unreasonable to expect network predictive accuracy not to degrade |
| title | Evaluation of autonomous systems under data distribution shifts |
| topic | Machine Learning I.4.m |
| url | https://arxiv.org/abs/2406.20046 |