Raising the bar on the evaluation of out-of-distribution detection
In image classification, a lot of development has happened in detecting out-of-distribution (OoD) data. However, most OoD detection methods are evaluated on a standard set of datasets, arbitrarily different from training data. There is no clear definition of what forms a "good" OoD dataset...
Κύριοι συγγραφείς: | Mukhoti, J, Lin, T-Y, Chen, B-C, Shah, A, Torr, PHS, Dokania, PK, Lim, S-N |
---|---|
Μορφή: | Conference item |
Γλώσσα: | English |
Έκδοση: |
EEE
2023
|
Παρόμοια τεκμήρια
Παρόμοια τεκμήρια
-
Using mixup as a regularizer can surprisingly improve accuracy and out-of-distribution robustness
ανά: Pinto, F, κ.ά.
Έκδοση: (2023) -
Placing objects in context via inpainting for out-of-distribution segmentation
ανά: De Jorge, P, κ.ά.
Έκδοση: (2024) -
On using focal loss for neural network calibration
ανά: Mukhoti, J, κ.ά.
Έκδοση: (2020) -
Calibrating deep neural networks using focal loss
ανά: Mukhoti, J, κ.ά.
Έκδοση: (2020) -
GDumb: A simple approach that questions our progress in continual learning
ανά: Prabhu, A, κ.ά.
Έκδοση: (2020)