Abstract
Understanding the decision-making and trusting the reliability of Deep Machine Learning Models is crucial for adopting such methods to safety-relevant applications which play an important role in the digitization of the railway system. We extend self-explainable Prototypical Variational models with autoencoder-based Out-of-Distribution (OOD) detection: A Variational Autoencoder is applied to learn a meaningful latent space which can be used for distance-based classification, likelihood estimation for OOD detection, and reconstruction. The In-Distribution (ID) region is defined by a Gaussian mixture distribution with learned prototypes representing the center of each mode. Furthermore, a novel restriction loss is introduced that promotes a compact ID region in the latent space without collapsing it into single points. The reconstructive capabilities of the Autoencoder ensure the explainability of the prototypes and the ID region of the classifier, further aiding the discrimination of OOD samples. Extensive evaluations on common OOD detection benchmarks as well as a large-scale dataset from a real-world railway application demonstrate the usefulness of the approach, outperforming previous methods.
This is a preview of subscription content, log in via an institution to check access.
Access this chapter
Tax calculation will be finalised at checkout
Purchases are for personal use only
Similar content being viewed by others
References
Bercea, C., Rueckert, D., Schnabel, J.: What do we learn? Debunking the myth of unsupervised outlier detection. arXiv preprint arXiv:2206.03698 (2022)
Chen, G., Peng, P., Wang, X., Tian, Y.: Adversarial reciprocal points learning for open set recognition. In: TPAMI, pp. 8065–8081 (2022)
Denouden, T., Salay, R., Czarnecki, K., Abdelzad, V., Phan, B., Vernekar, S.: Improving reconstruction autoencoder out-of-distribution detection with Mahalanobis distance. arXiv preprint arXiv:1812.02765 (2018)
Digitale Schiene Deutschland: https://www.digitale-schiene-deutschland.de/en/. Accessed 07 Jun 2024
Du, X., Gozum, G., Ming, Y., Li, Y.: Siren: shaping representations for detecting out-of-distribution objects. In: NeurIPS, vol. 35, pp. 20434–20449 (2022)
Fiack, A., Weller, F., Heimes, M., Laux, T.: Digitale Schiene Deutschland - Zukunftstechnologien für das Bahnsystem. EIK 2024 (2024)
Gal, Y., Ghahramani, Z.: Dropout as a Bayesian approximation: Representing model uncertainty in deep learning. In: ICML, pp. 1050–1059 (2016)
Gautam, S., et al.: ProtoVAE: a trustworthy self-explainable prototypical variational model. In: NeurIPS, pp. 17940–17952 (2022)
Graham, M.S., Pinaya, W.H.L., Tudosiu, P.D., Nachev, P., Ourselin, S., Cardoso, M.J.: Denoising diffusion models for out-of-distribution detection. In: CVPR Workshops, pp. 2948–2957 (2023)
Guo, C., Pleiss, G., Sun, Y., Weinberger, K.Q.: On calibration of modern neural networks. In: ICML, pp. 1321–1330 (2017)
He, K., Zhang, X., Ren, S., Sun, J.: Deep residual learning for image recognition. In: CVPR, pp. 770–778 (2016)
Hendrycks, D., et al.: Scaling out-of-distribution detection for real-world settings. In: ICML, pp. 8759–8773 (2022)
Hendrycks, D., Gimpel, K.: A baseline for detecting misclassified and out-of-distribution examples in neural networks. In: ICLR (2017)
Hendrycks, D., et al.: PixMix: dreamlike pictures comprehensively improve safety measures. In: CVPR, pp. 16762–16771 (2022)
Lakshminarayanan, B., Pritzel, A., Blundell, C.: Simple and scalable predictive uncertainty estimation using deep ensembles. In: NIPS, pp. 6405–6416 (2017)
Lee, K., Lee, K., Lee, H., Shin, J.: A simple unified framework for detecting out-of-distribution samples and adversarial attacks. In: NeurIPS, pp. 7167–7177 (2018)
Nalisnick, E., Matsukawa, A., Teh, Y.W., Gorur, D., Lakshminarayanan, B.: Do deep generative models know what they don’t know? In: ICLR (2019)
Oza, P., Patel, V.M.: C2AE: class conditioned auto-encoder for open-set recognition. In: CVPR, pp. 2307–2316 (2019)
Ruff, L., et al.: Deep one-class classification. In: ICML, pp. 4393–4402 (2018)
Shi, W., et al.: Real-time single image and video super-resolution using an efficient sub-pixel convolutional neural network. In: CVPR, pp. 1874–1883 (2016)
Sun, X., Yang, Z., Zhang, C., Ling, K.V., Peng, G.: Conditional gaussian distribution learning for open set recognition. In: CVPR, pp. 13477–13486 (2020)
Sun, Y., Guo, C., Li, Y.: React: out-of-distribution detection with rectified activations. In: NeurIPS, pp. 144–157 (2021)
Sun, Y., Ming, Y., Zhu, X., Li, Y.: Out-of-distribution detection with deep nearest neighbors. In: ICML, pp. 20827–20840 (2022)
Wang, H., Li, Z., Feng, L., Zhang, W.: ViM: out-of-distribution with virtual-logit matching. In: CVPR, pp. 4921–4930 (2022)
Xiao, Z., Yan, Q., Amit, Y.: Likelihood regret: an out-of-distribution detection score for variational auto-encoder. In: NeurIPS, pp. 20685–20696 (2020)
Yang, J., et al.: OpenOOD: benchmarking generalized out-of-distribution detection. In: NeurIPS, pp. 32598–32611 (2022)
Yang, J., Zhou, K., Li, Y., Liu, Z.: Generalized out-of-distribution detection: a survey. arXiv preprint arXiv:2110.11334 (2021)
Zhang, R., Isola, P., Efros, A.A., Shechtman, E., Wang, O.: The unreasonable effectiveness of deep features as a perceptual metric. In: CVPR, pp. 586–595 (2018)
Zhou, Y.: Rethinking reconstruction autoencoder-based out-of-distribution detection. In: CVPR, pp. 7369–7377 (2022)
Acknowledgments
This work has been done within the sector initiative Digitale Schiene Deutschland [4], in which DB InfraGO AG is working together with industry partners on a far-reaching automation of the German railway system. Rail traffic is one of the most important building blocks for the transition to new climate-friendly traffic solutions in our society.
For higher quality traffic the railway system must be fundamentally modernized through profound, technological innovations. Numerous new capabilities are made available by digital technologies, such as fully automated driving, which will be tested and further developed for use in the system.
Author information
Authors and Affiliations
Corresponding author
Editor information
Editors and Affiliations
Rights and permissions
Copyright information
© 2024 The Author(s), under exclusive license to Springer Nature Switzerland AG
About this paper
Cite this paper
Orglmeister, C., Bochinski, E., Eiselein, V., Fleig, E. (2024). Enclosing Prototypical Variational Autoencoder for Explainable Out-of-Distribution Detection. In: Ceccarelli, A., Trapp, M., Bondavalli, A., Schoitsch, E., Gallina, B., Bitsch, F. (eds) Computer Safety, Reliability, and Security. SAFECOMP 2024 Workshops. SAFECOMP 2024. Lecture Notes in Computer Science, vol 14989. Springer, Cham. https://doi.org/10.1007/978-3-031-68738-9_29
Download citation
DOI: https://doi.org/10.1007/978-3-031-68738-9_29
Published:
Publisher Name: Springer, Cham
Print ISBN: 978-3-031-68737-2
Online ISBN: 978-3-031-68738-9
eBook Packages: Computer ScienceComputer Science (R0)