RODEO: Robust Outlier Detection via Exposing Adaptive Out-of-Distribution Samples

Hossein Mirzaei, Mohammad Jafari, Hamid Reza Dehbashi, Ali Ansari, Sepehr Ghobadi, Masoud Hadi, Arshia Soltani Moakhar, Mohammad Azizmalayeri, Mahdieh Soleymani Baghshah, Mohammad Hossein Rohban
Proceedings of the 41st International Conference on Machine Learning, PMLR 235:35744-35778, 2024.

Abstract

In recent years, there have been significant improvements in various forms of image outlier detection. However, outlier detection performance under adversarial settings lags far behind that in standard settings. This is due to the lack of effective exposure to adversarial scenarios during training, especially on unseen outliers, leading detection models failing to learn robust features. To bridge this gap, we introduce RODEO, a data-centric approach that generates effective outliers for robust outlier detection. More specifically, we show that incorporating outlier exposure (OE) and adversarial training could be an effective strategy for this purpose, as long as the exposed training outliers meet certain characteristics, including diversity, and both conceptual differentiability and analogy to the inlier samples. We leverage a text-to-image model to achieve this goal. We demonstrate both quantitatively and qualitatively that our adaptive OE method effectively generates ”diverse” and ”near-distribution” outliers, leveraging information from both text and image domains. Moreover, our experimental results show that utilizing our synthesized outliers significantly enhances the performance of the outlier detector, particularly in adversarial settings.

Cite this Paper

BibTeX
@InProceedings{pmlr-v235-mirzaei24a, title = {{RODEO}: Robust Outlier Detection via Exposing Adaptive Out-of-Distribution Samples}, author = {Mirzaei, Hossein and Jafari, Mohammad and Dehbashi, Hamid Reza and Ansari, Ali and Ghobadi, Sepehr and Hadi, Masoud and Soltani Moakhar, Arshia and Azizmalayeri, Mohammad and Baghshah, Mahdieh Soleymani and Rohban, Mohammad Hossein}, booktitle = {Proceedings of the 41st International Conference on Machine Learning}, pages = {35744--35778}, year = {2024}, editor = {Salakhutdinov, Ruslan and Kolter, Zico and Heller, Katherine and Weller, Adrian and Oliver, Nuria and Scarlett, Jonathan and Berkenkamp, Felix}, volume = {235}, series = {Proceedings of Machine Learning Research}, month = {21--27 Jul}, publisher = {PMLR}, pdf = {https://raw.githubusercontent.com/mlresearch/v235/main/assets/mirzaei24a/mirzaei24a.pdf}, url = {https://proceedings.mlr.press/v235/mirzaei24a.html}, abstract = {In recent years, there have been significant improvements in various forms of image outlier detection. However, outlier detection performance under adversarial settings lags far behind that in standard settings. This is due to the lack of effective exposure to adversarial scenarios during training, especially on unseen outliers, leading detection models failing to learn robust features. To bridge this gap, we introduce RODEO, a data-centric approach that generates effective outliers for robust outlier detection. More specifically, we show that incorporating outlier exposure (OE) and adversarial training could be an effective strategy for this purpose, as long as the exposed training outliers meet certain characteristics, including diversity, and both conceptual differentiability and analogy to the inlier samples. We leverage a text-to-image model to achieve this goal. We demonstrate both quantitatively and qualitatively that our adaptive OE method effectively generates ”diverse” and ”near-distribution” outliers, leveraging information from both text and image domains. Moreover, our experimental results show that utilizing our synthesized outliers significantly enhances the performance of the outlier detector, particularly in adversarial settings.} }
Endnote
%0 Conference Paper %T RODEO: Robust Outlier Detection via Exposing Adaptive Out-of-Distribution Samples %A Hossein Mirzaei %A Mohammad Jafari %A Hamid Reza Dehbashi %A Ali Ansari %A Sepehr Ghobadi %A Masoud Hadi %A Arshia Soltani Moakhar %A Mohammad Azizmalayeri %A Mahdieh Soleymani Baghshah %A Mohammad Hossein Rohban %B Proceedings of the 41st International Conference on Machine Learning %C Proceedings of Machine Learning Research %D 2024 %E Ruslan Salakhutdinov %E Zico Kolter %E Katherine Heller %E Adrian Weller %E Nuria Oliver %E Jonathan Scarlett %E Felix Berkenkamp %F pmlr-v235-mirzaei24a %I PMLR %P 35744--35778 %U https://proceedings.mlr.press/v235/mirzaei24a.html %V 235 %X In recent years, there have been significant improvements in various forms of image outlier detection. However, outlier detection performance under adversarial settings lags far behind that in standard settings. This is due to the lack of effective exposure to adversarial scenarios during training, especially on unseen outliers, leading detection models failing to learn robust features. To bridge this gap, we introduce RODEO, a data-centric approach that generates effective outliers for robust outlier detection. More specifically, we show that incorporating outlier exposure (OE) and adversarial training could be an effective strategy for this purpose, as long as the exposed training outliers meet certain characteristics, including diversity, and both conceptual differentiability and analogy to the inlier samples. We leverage a text-to-image model to achieve this goal. We demonstrate both quantitatively and qualitatively that our adaptive OE method effectively generates ”diverse” and ”near-distribution” outliers, leveraging information from both text and image domains. Moreover, our experimental results show that utilizing our synthesized outliers significantly enhances the performance of the outlier detector, particularly in adversarial settings.
APA
Mirzaei, H., Jafari, M., Dehbashi, H.R., Ansari, A., Ghobadi, S., Hadi, M., Soltani Moakhar, A., Azizmalayeri, M., Baghshah, M.S. & Rohban, M.H.. (2024). RODEO: Robust Outlier Detection via Exposing Adaptive Out-of-Distribution Samples. Proceedings of the 41st International Conference on Machine Learning, in Proceedings of Machine Learning Research 235:35744-35778 Available from https://proceedings.mlr.press/v235/mirzaei24a.html.

Related Material