firstbacksecondback
2 Results
|
Poster
|
Thu 14:30 |
Efficient Augmentation via Data Subsampling Michael Kuchnik · Virginia Smith |
|
|
Poster
|
Tue 14:30 |
ImageNet-trained CNNs are biased towards texture; increasing shape bias improves accuracy and robustness Robert Geirhos · Patricia Rubisch · Claudio Michaelis · Matthias Bethge · Felix Wichmann · Wieland Brendel |