Section AI (Artificial Intelligence)
INVITED PAPER:
Non-adversarial Robustness of Deep Learning Methods for Computer Vision
Gorana Gojić,
The Institute for Artificial Intelligence Research and Development of Serbia,
Novi Sad, Serbia
Abstract:
Non-adversarial robustness, also known as natural robustness, is a property of deep learning models that enables them to maintain performance even when faced with distribution shifts caused by natural variations in data. However, achieving this property is challenging because it is difficult to predict in advance the types of distribution shifts that may occur. To address this challenge, researchers have proposed various approaches, some of which anticipate potential distribution shifts, while others utilize knowledge about the shifts that have already occurred to enhance model generalisability. In this paper, we present a brief overview of the most recent techniques for improving the robustness of computer vision methods, as well as a summary of commonly used robustness benchmark datasets for evaluating the model’s performance under data distribution shifts. Finally, we examine the strengths and limitations of the approaches reviewed and identify general trends in deep learning robustness improvement for computer vision.
Short Bio:
|
|
Gorana Gojić received her B.Sc. with honours and M.Sc. degrees in electrical engineering and computing from the University of Novi Sad, Faculty of Technical Sciences, Novi Sad, Serbia, in 2015 and 2016, respectively. She is currently working as a researcher at The Institute for Artificial Intelligence Research and Development of Serbia as a researcher, while pursuing a Ph.D. degree in the field of machine learning at the Faculty of Technical Sciences, University of Novi Sad. Her professional interests include deep learning and computer vision applied in medical image processing. |
