Key Feature Replacement of In-Distribution Samples for Out-of-Distribution Detection
Jaeyoung Kim, Seo Taek Kong, Dongbin Na, Kyu-Hwan Jung
Abstract
Out-of-distribution (OOD) detection can be used in deep learning-based applications to reject outlier samples from being unreliably classified by deep neural networks. Learning to classify between OOD and in-distribution samples is difficult because data comprising the former is extremely diverse. It has been observed that an auxiliary OOD dataset is most effective in training a ``rejection'' network when its samples are semantically similar to in-distribution images. We first deduce that OOD images are perceived by a deep neural network to be semantically similar to in-distribution samples when they share a common background, as deep networks are observed to incorrectly classify such images with high confidence. We then propose a simple yet effective Key In-distribution feature Replacement BY inpainting (KIRBY) procedure that constructs a surrogate OOD dataset by replacing class-discriminative features of in-distribution samples with marginal background features. The procedure can be implemented using off-the-shelf vision algorithms, where each step within the algorithm is shown to make the surrogate data increasingly similar to in-distribution data. Design choices in each step are studied extensively, and an exhaustive comparison with state-of-the-art algorithms demonstrates KIRBY's competitiveness on various benchmarks.
BibTeX
@article{Kim_Kong_Na_Jung_2023, title={Key Feature Replacement of In-Distribution Samples for Out-of-Distribution Detection}, volume={37}, url={https://ojs.aaai.org/index.php/AAAI/article/view/25995}, DOI={10.1609/aaai.v37i7.25995}, abstractNote={Out-of-distribution (OOD) detection can be used in deep learning-based applications to reject outlier samples from being unreliably classified by deep neural networks. Learning to classify between OOD and in-distribution samples is difficult because data comprising the former is extremely diverse. It has been observed that an auxiliary OOD dataset is most effective in training a ``rejection’’ network when its samples are semantically similar to in-distribution images. We first deduce that OOD images are perceived by a deep neural network to be semantically similar to in-distribution samples when they share a common background, as deep networks are observed to incorrectly classify such images with high confidence. We then propose a simple yet effective Key In-distribution feature Replacement BY inpainting (KIRBY) procedure that constructs a surrogate OOD dataset by replacing class-discriminative features of in-distribution samples with marginal background features. The procedure can be implemented using off-the-shelf vision algorithms, where each step within the algorithm is shown to make the surrogate data increasingly similar to in-distribution data. Design choices in each step are studied extensively, and an exhaustive comparison with state-of-the-art algorithms demonstrates KIRBY’s competitiveness on various benchmarks.}, number={7}, journal={Proceedings of the AAAI Conference on Artificial Intelligence}, author={Kim, Jaeyoung and Kong, Seo Taek and Na, Dongbin and Jung, Kyu-Hwan}, year={2023}, month={Jun.}, pages={8246-8254} }