Topology-Matching Normalizing Flows for Out-of-Distribution Detection in Robot Learning

Jianxiang Feng, Jongseok Lee, Simon Geisler, Stephan Günnemann, Rudolph Triebel
Proceedings of The 7th Conference on Robot Learning, PMLR 229:3214-3241, 2023.

Abstract

To facilitate reliable deployments of autonomous robots in the real world, Out-of-Distribution (OOD) detection capabilities are often required. A powerful approach for OOD detection is based on density estimation with Normalizing Flows (NFs). However, we find that prior work with NFs attempts to match the complex target distribution topologically with naïve base distributions leading to adverse implications. In this work, we circumvent this topological mismatch using an expressive class-conditional base distribution trained with an information-theoretic objective to match the required topology. The proposed method enjoys the merits of wide compatibility with existing learned models without any performance degradation and minimum computation overhead while enhancing OOD detection capabilities. We demonstrate superior results in density estimation and 2D object detection benchmarks in comparison with extensive baselines. Moreover, we showcase the applicability of the method with a real-robot deployment.

Cite this Paper

BibTeX
@InProceedings{pmlr-v229-feng23b, title = {Topology-Matching Normalizing Flows for Out-of-Distribution Detection in Robot Learning}, author = {Feng, Jianxiang and Lee, Jongseok and Geisler, Simon and G\"{u}nnemann, Stephan and Triebel, Rudolph}, booktitle = {Proceedings of The 7th Conference on Robot Learning}, pages = {3214--3241}, year = {2023}, editor = {Tan, Jie and Toussaint, Marc and Darvish, Kourosh}, volume = {229}, series = {Proceedings of Machine Learning Research}, month = {06--09 Nov}, publisher = {PMLR}, pdf = {https://proceedings.mlr.press/v229/feng23b/feng23b.pdf}, url = {https://proceedings.mlr.press/v229/feng23b.html}, abstract = {To facilitate reliable deployments of autonomous robots in the real world, Out-of-Distribution (OOD) detection capabilities are often required. A powerful approach for OOD detection is based on density estimation with Normalizing Flows (NFs). However, we find that prior work with NFs attempts to match the complex target distribution topologically with naïve base distributions leading to adverse implications. In this work, we circumvent this topological mismatch using an expressive class-conditional base distribution trained with an information-theoretic objective to match the required topology. The proposed method enjoys the merits of wide compatibility with existing learned models without any performance degradation and minimum computation overhead while enhancing OOD detection capabilities. We demonstrate superior results in density estimation and 2D object detection benchmarks in comparison with extensive baselines. Moreover, we showcase the applicability of the method with a real-robot deployment.} }
Endnote
%0 Conference Paper %T Topology-Matching Normalizing Flows for Out-of-Distribution Detection in Robot Learning %A Jianxiang Feng %A Jongseok Lee %A Simon Geisler %A Stephan Günnemann %A Rudolph Triebel %B Proceedings of The 7th Conference on Robot Learning %C Proceedings of Machine Learning Research %D 2023 %E Jie Tan %E Marc Toussaint %E Kourosh Darvish %F pmlr-v229-feng23b %I PMLR %P 3214--3241 %U https://proceedings.mlr.press/v229/feng23b.html %V 229 %X To facilitate reliable deployments of autonomous robots in the real world, Out-of-Distribution (OOD) detection capabilities are often required. A powerful approach for OOD detection is based on density estimation with Normalizing Flows (NFs). However, we find that prior work with NFs attempts to match the complex target distribution topologically with naïve base distributions leading to adverse implications. In this work, we circumvent this topological mismatch using an expressive class-conditional base distribution trained with an information-theoretic objective to match the required topology. The proposed method enjoys the merits of wide compatibility with existing learned models without any performance degradation and minimum computation overhead while enhancing OOD detection capabilities. We demonstrate superior results in density estimation and 2D object detection benchmarks in comparison with extensive baselines. Moreover, we showcase the applicability of the method with a real-robot deployment.
APA
Feng, J., Lee, J., Geisler, S., Günnemann, S. & Triebel, R.. (2023). Topology-Matching Normalizing Flows for Out-of-Distribution Detection in Robot Learning. Proceedings of The 7th Conference on Robot Learning, in Proceedings of Machine Learning Research 229:3214-3241 Available from https://proceedings.mlr.press/v229/feng23b.html.

Related Material