Deep neural networks for computer vision are deployed in increasingly safety-critical and socially-impactful applications, motivating the need to close the gap in model performance …
Test-time adaptation provides an effective means of tackling the potential distribution shift between model training and inference, by dynamically updating the model at test time. This …
A model must adapt itself to generalize to new and different data during testing. In this setting of fully test-time adaptation the model has only the test data and its own parameters …
Abstract Convolutional Neural Networks (CNNs) have advanced existing medical systems for automatic disease diagnosis. However, there are still concerns about the reliability of …
J Gao, J Zhang, X Liu, T Darrell… - Proceedings of the …, 2023 - openaccess.thecvf.com
Test-time adaptation harnesses test inputs to improve the accuracy of a model trained on source data when tested on shifted target data. Most methods update the source model by …
OF Kar, T Yeo, A Atanov… - Proceedings of the IEEE …, 2022 - openaccess.thecvf.com
We introduce a set of image transformations that can be used as corruptions to evaluate the robustness of models as well as data augmentation mechanisms for training neural …
X Mao, G Qi, Y Chen, X Li, R Duan… - Proceedings of the …, 2022 - openaccess.thecvf.com
Abstract Recent advances on Vision Transformer (ViT) and its improved variants have shown that self-attention-based networks surpass traditional Convolutional Neural Networks …
Data augmentation is a simple yet effective way to improve the robustness of deep neural networks (DNNs). Diversity and hardness are two complementary dimensions of data …
Since out-of-distribution generalization is a generally ill-posed problem, various proxy targets (eg, calibration, adversarial robustness, algorithmic corruptions, invariance across …