Discriminatively trained neural classifiers can be trusted, only when the input data comes from the training distribution (in-distribution). Therefore, detecting out-of-distribution (OOD) samples is very important to avoid classification errors.
Apr 27, 2019
Apr 27, 2019 · Discriminatively trained neural classifiers can be trusted, only when the input data comes from the training distribution (in-distribution).
Apr 27, 2019 · Discriminatively trained neural classifiers can be trusted, only when the input data comes from the training distribution (in-distribution).
Aug 27, 2022 · Abstract: Discriminatively trained neural classifiers can be trusted, only when the input data comes from the training distribution (in-distribution).
In essence, our method jointly trains both classification and generative neural networks for out-of-distribution. We demonstrate its effectiveness on various ...
Missing: Analysis | Show results with:Analysis
Analysis of Confident-Classifiers for Out-of-Distribution Detection ; Sachin Vernekar · Ashish Gaurav · Taylor Denouden · Buu Phan · Vahdat Abdelzad · Rick Salay.
Code for ICLR 2019 SafeML workshop paper: Analysis of Confident-Classifiers for Out-of-distribution Detection. 5 stars 4 forks Branches Tags Activity.
The problem of detecting whether a test sample is from in-distribution (i.e., train- ing distribution by a classifier) or out-of-distribution sufficiently ...
Oct 19, 2022 · In this article, we'll show you a novel and simple adjustment to model predicted probabilities that can improve OOD detection with classifier models trained on ...
People also ask
What is out of distribution detection?
How do you judge whether the classifier is well suited for the data set?
This paper suggests training a classifier by adding an explicit "reject" class for OOD samples by minimizing the standard cross-entropy loss on ...