Open AkiraTOSEI opened 4 years ago
Compared to models trained normally, models trained with adversarial noise performed better in transfer learning. The results of the visualization showed that they were classifying in a more human-like sense, which may have influenced the results.
https://arxiv.org/abs/2007.05869
2020/07/11
Francisco Utrera, Evan Kravitz, N. Benjamin Erichson, Rajiv Khanna, Michael W. Mahoney
TL;DR
Compared to models trained normally, models trained with adversarial noise performed better in transfer learning. The results of the visualization showed that they were classifying in a more human-like sense, which may have influenced the results.
Why it matters:
Paper URL
https://arxiv.org/abs/2007.05869
Submission Dates(yyyy/mm/dd)
2020/07/11
Authors and institutions
Francisco Utrera, Evan Kravitz, N. Benjamin Erichson, Rajiv Khanna, Michael W. Mahoney
Methods
Results
Comments