Open Access
Review
Table 4
Theoretical hypotheses about adversarial examples
| Hypothesis | Year | Venue | Description |
|---|---|---|---|
| Low-probability hypothesis[3] | 2014 | ICLR | Adversarial examples lie in the low-probability spaces |
| Linearity hypothesis[4] | 2015 | arXiv | The highly linear behaviors of DNNs are responsible for adversarial examples |
| Off-manifold hypothesis[80] | 2017 | SIGSAC | Adversarial examples lie off the clean data manifold and obey a different distribution |
| Manifold geometry hypothesis[91] | 2018 | arXiv | The intricate geometric structure of the data manifold makes the model rather sensitive to adversarial examples |
Current usage metrics show cumulative count of Article Views (full-text article views including HTML views, PDF and ePub downloads, according to the available data) and Abstracts Views on Vision4Press platform.
Data correspond to usage on the plateform after 2015. The current usage metrics is available 48-96 hours after online publication and is updated daily on week days.
Initial download of the metrics may take a while.