Abstract
Detecting out-of-distribution (OOD) inputs for deep learning models is a critical task when models are deployed in real-world environments. Recently, a large number of works have been dedicated to tackling the OOD detection problem. One of the most straightforward and effective ways is OOD training, which adds heterogeneous auxiliary data in the training stage. However, the extra auxiliary data cannot be involved arbitrarily. A high-quality and powerful auxiliary dataset must contain samples that belong to OOD but are close to in-distribution (ID), which can teach the model to learn more information about OOD samples, furthermore, distinguish OOD from ID. The key issue for this problem is how to simply acquire such distinctive OOD samples. In this article, we propose an enhanced Mixup-based OOD (MixOOD) detection strategy that can be attached to any threshold-based OOD detecting method. Different from the traditional Mixup designed for ID data augmentation, our proposed MixOOD generates augmented images with deliberately modified Mixup and then uses them as auxiliary OOD data to leverage the OOD detection. We test our method with classical OOD detecting approaches like Maximum Softmax Probability, Energy Score, and Out-of-distribution detector for Neural networks. Experiments show that models with MixOOD can better distinguish in- and out-of-distribution samples than the original version of each approach.
- [1] . 2012. Imagenet classification with deep convolutional neural networks. Adv. Neural Info. Process. Syst. 25 (2012).Google Scholar
- [2] . 1998. Gradient-based learning applied to document recognition. Proc. IEEE 86, 11 (1998), 2278–2324.Google Scholar
- [3] . 2021. Yolox: Exceeding yolo series in 2021. Retrieved from https://arXiv:2107.08430.Google Scholar
- [4] . 2019. Efficientnet: Rethinking model scaling for convolutional neural networks. In Proceedings of the International Conference on Machine Learning. PMLR, 6105–6114.Google Scholar
- [5] . 2020. Energy-based out-of-distribution detection. Adv Neural Info. Process. Syst. 33 (2020), 21464–21475.Google Scholar
- [6] . 2017. A baseline for detecting misclassified and out-of-distribution examples in neural networks. In Proceedings of the International Conference on Learning Representations.Google Scholar
- [7] . 2011. Reading digits in natural images with unsupervised feature learning. In Proceedings of the Conference on Neural Information Processing Workshops.Google Scholar
- [8] . 2017. Enhancing the reliability of out-of-distribution image detection in neural networks. Retrieved from https://arXiv:1706.02690.Google Scholar
- [9] . 2019. Deep anomaly detection with outlier exposure. In Proceedings of the International Conference on Learning Representations.Google Scholar
- [10] . 2017. Training confidence-calibrated classifiers for detecting out-of-distribution samples. Retrieved from https://arXiv:1711.09325.Google Scholar
- [11] 1996. Tiny Imagenet Visual Recognition Challenge. Website. Retrieved from https://tiny-imagenet.herokuapp.com.Google Scholar
- [12] . 2017. mixup: Beyond empirical risk minimization. Retrieved from https://arXiv:1710.09412.Google Scholar
- [13] . 2019. CutMix: Regularization strategy to train strong classifiers with localizable features. In Proceedings of IEEE International Conference on Computer Vision. 6022–6031.
DOI: Google ScholarCross Ref
- [14] . 2006. A tutorial on energy-based learning. Predict. Struct. Data 1, 0 (2006).Google Scholar
- [15] . 2021. Adversarial reciprocal points learning for open set recognition. IEEE Trans. Pattern Anal. Mach. Intell. 44 (2021), 8065–8081.
DOI: Google ScholarCross Ref
- [16] . 2021. Intra-class mixup for out-of-distribution detection. Retrieved from https://openreview.net/forum?id=HRL6el2SBQ.Google Scholar
- [17] . 2021. Open-set recognition: A good closed-set classifier is all you need. Retrieved from https://arXiv:2110.06207.Google Scholar
- [18] . 2022. Exploiting mixed unlabeled data for detecting samples of seen and unseen out-of-distribution classes. In Proceedings of the AAAI Conference on Artificial Intelligence.Google Scholar
Cross Ref
- [19] . 2021. Fine-grained out-of-distribution detection with mixup outlier exposure. Retrieved from https://arXiv:2106.03917.Google Scholar
- [20] . 2022. Out-of-distribution detection for long-tailed and fine-grained skin lesion images. In Proceedings of the International Conference on Medical Image Computing and Computer-Assisted Intervention. Springer, 732–742.Google Scholar
Digital Library
- [21] . 2019. A survey on image data augmentation for deep learning. J. Big Data 6, 1 (2019), 1–48.Google Scholar
Cross Ref
- [22] . 2018. UMAP: Uniform manifold approximation and projection. J. Open Source Softw. 3, 29 (2018), 861.Google Scholar
Cross Ref
- [23] . 2009. Learning multiple layers of features from tiny images. Master’s Thesis. University of Toronto.Google Scholar
- [24] . 2015. Lsun: Construction of a large-scale image dataset using deep learning with humans in the loop. Retrieved from https://arXiv:1506.03365.Google Scholar
- [25] . 2014. Describing textures in the wild. In Proceedings of the IEEE Conference on Computer Vision and Pattern Recognition. 3606–3613.Google Scholar
Digital Library
- [26] . 2015. Turkergaze: Crowdsourcing saliency with webcam based eye tracking. Retrieved from https://arXiv:1504.06755.Google Scholar
- [27] . 2009. Imagenet: A large-scale hierarchical image database. In Proceedings of the IEEE Conference on Computer Vision and Pattern Recognition. IEEE, 248–255.Google Scholar
Cross Ref
- [28] . 2017. Densely connected convolutional networks. In Proceedings of the IEEE Conference on Computer Vision and Pattern Recognition. 4700–4708.Google Scholar
Cross Ref
- [29] . 2011. Adaptive subgradient methods for online learning and stochastic optimization. J. Mach. Learn. Res. 12, 7 (2011).Google Scholar
- [30] . 2014. A method for stochastic optimization. Retrieved from https://arXiv:1412.6980.Google Scholar
- [31] . 2018. Out-of-distribution detection using an ensemble of self supervised leave-out classifiers. In Proceedings of the European Conference on Computer Vision. 550–564.Google Scholar
Digital Library
- [32] . 2016. Wide residual networks. Retrieved from https://arXiv:1605.07146.Google Scholar
- [33] . 2020. Detecting semantic anomalies. In Proceedings of the AAAI Conference on Artificial Intelligence.Google Scholar
Cross Ref
- [34] . 2013. Fine-grained visual classification of aircraft. Retrieved from https://arXiv:1306.5151.Google Scholar
- [35] . 2013. 3D object representations for fine-grained categorization. In Proceedings of IEEE International Conference on Computer Vision workshops. 554–561.Google Scholar
Digital Library
- [36] . 2018. Fine-grained representation learning and recognition by exploiting hierarchical semantic embedding. In Proceedings of the ACM International Conference on Multimedia. 2023–2031.Google Scholar
Digital Library
- [37] . 2015. Building a bird recognition app and large scale dataset with citizen scientists: The fine print in fine-grained dataset collection. In Proceedings of the IEEE Conference on Computer Vision and Pattern Recognition. 595–604.Google Scholar
Cross Ref
Index Terms
MixOOD: Improving Out-of-distribution Detection with Enhanced Data Mixup
Recommendations
Out-of-Distribution Detection Using Outlier Detection Methods
Image Analysis and Processing – ICIAP 2022AbstractOut-of-distribution detection (OOD) deals with anomalous input to neural networks. In the past, specialized methods have been proposed to identify anomalous input. Similarly, it was shown that feature extraction models in combination with outlier ...
Self-supervised anomaly pattern detection for large scale industrial data
AbstractDetecting the anomalies in a large amounts of high-dimensional data has been a challenging task. In the Industry 4.0 environment, large-scale high-dimensional monitoring data features the complex pattern of high level semantics. In ...
ATOM: Robustifying Out-of-Distribution Detection Using Outlier Mining
Machine Learning and Knowledge Discovery in Databases. Research TrackAbstractDetecting out-of-distribution (OOD) inputs is critical for safely deploying deep learning models in an open-world setting. However, existing OOD detection solutions can be brittle in the open world, facing various types of adversarial OOD inputs. ...






Comments