Out-of-Distribution Detection Using an Ensemble of Self Supervised Leave-Out Classifiers

Vyas, Apoorv; Jammalamadaka, Nataraj; Zhu, Xiaoyan; Das, Dipankar; Kaul, Bharat; Willke, Theodore L.

doi:10.1007/978-3-030-01237-3_34

Cited by 180 publications

(156 citation statements)

References 19 publications

(32 reference statements)

Supporting

Mentioning

150

Contrasting

Order By: Relevance

“…The results are summarized in Table 2, which shows the comparison of our method, ODIN [16] and Ensemble of Leave-Out Classifiers (ELOC) [26] on various benchmarks. In addition, ELOC [26] does not have results for iSUN as an OOD dataset because they use the whole iSUN as a validation dataset.…”

Section: Resultsmentioning

confidence: 99%

“…The current state-of-the-art method for OOD detection is the ensemble of self-supervised leave-out classifiers proposed by Vyas et al [26]. They divided the training ID data into K partitions and assign one partition as OOD and the remaining partitions as ID to train K classifiers by a novel loss function, called margin entropy loss, to increase the prediction confidence of ID samples and decrease the prediction confidence of OOD samples.…”

Section: Related Workmentioning

confidence: 99%

“…Furthermore, for fair comparison, we used two classifiers and calculated the average score of these two classifiers as final output in the other methods, ODIN [16] and Ensemble of Leave-Out Classifiers [26], since our method has two classifiers which resulted in a few more parameters.…”

Section: Neural Network Architecturementioning

confidence: 99%

“…We followed the benchmarks given in [16,26] and used the OOD datasets below in our experiments: [28], which is used for gaze tracking, deployed on Amazon Mechanical Turk using a webcam [29]. It contains 8,925 scene images, and all images are downsampled to a size of 32 × 32.…”

Section: Out-of-distributionmentioning

confidence: 99%

See 3 more Smart Citations

Unsupervised Out-of-Distribution Detection by Maximum Classifier Discrepancy

Aizawa

2019

2019 IEEE/CVF International Conference on Computer Vision (ICCV)

109

View full text Add to dashboard Cite

Since deep learning models have been implemented in many commercial applications, it is important to detect out-of-distribution (OOD) inputs correctly to maintain the performance of the models, ensure the quality of the collected data, and prevent the applications from being used for other-than-intended purposes. In this work, we propose a two-head deep convolutional neural network (CNN) and maximize the discrepancy between the two classifiers to detect OOD inputs. We train a two-head CNN consisting of one common feature extractor and two classifiers which have different decision boundaries but can classify in-distribution (ID) samples correctly. Unlike previous methods, we also utilize unlabeled data for unsupervised training and we use these unlabeled data to maximize the discrepancy between the decision boundaries of two classifiers to push OOD samples outside the manifold of the in-distribution (ID) samples, which enables us to detect OOD samples that are far from the support of the ID samples. Overall, our approach significantly outperforms other state-of-the-art methods on several OOD detection benchmarks and two cases of real-world simulation.

show abstract

Section: Resultsmentioning

confidence: 99%

Section: Related Workmentioning

confidence: 99%

Section: Neural Network Architecturementioning

confidence: 99%

Section: Out-of-distributionmentioning

confidence: 99%

See 2 more Smart Citations

Unsupervised Out-of-Distribution Detection by Maximum Classifier Discrepancy

Aizawa

2019

2019 IEEE/CVF International Conference on Computer Vision (ICCV)

109

View full text Add to dashboard Cite

show abstract

“…In order to do so, the knowledge of out-of-distributional images (final row), in this case non-dog images, are used to learn a suitable representation. task than out-of-distribution detection [23] since novel object samples are expected to be from a similar distribution to that of known samples.…”

Section: Introductionmentioning

confidence: 99%

Deep Transfer Learning for Multiple Class Novelty Detection

Perera

Patel

2019

2019 IEEE/CVF Conference on Computer Vision and Pattern Recognition (CVPR)

View full text Add to dashboard Cite

We propose a transfer learning-based solution for the problem of multiple class novelty detection. In particular, we propose an end-to-end deep-learning based approach in which we investigate how the knowledge contained in an external, out-of-distributional dataset can be used to improve the performance of a deep network for visual novelty detection. Our solution differs from the standard deep classification networks on two accounts. First, we use a novel loss function, membership loss, in addition to the classical cross-entropy loss for training networks. Secondly, we use the knowledge from the external dataset more effectively to learn globally negative filters, filters that respond to generic objects outside the known class set. We show that thresholding the maximal activation of the proposed network can be used to identify novel objects effectively. Extensive experiments on four publicly available novelty detection datasets show that the proposed method achieves significant improvements over the state-of-the-art methods.

show abstract

Simultaneous Semantic Segmentation and Outlier Detection in Presence of Domain Shift

Bevandić

Krešo

Oršić

et al. 2019

Lecture Notes in Computer Science

View full text Add to dashboard Cite

Recent success on realistic road driving datasets has increased interest in exploring robust performance in real-world applications. One of the major unsolved problems is to identify image content which can not be reliably recognized with a given inference engine. We therefore study approaches to recover a dense outlier map alongside the primary task with a single forward pass, by relying on shared convolutional features. We consider semantic segmentation as the primary task and perform extensive validation on WildDash val (inliers), LSUN val (outliers), and pasted objects from Pascal VOC 2007 (outliers). We achieve the best validation performance by training to discriminate inliers from pasted ImageNet-1k content, even though ImageNet-1k contains many roaddriving pixels, and, at least nominally, fails to account for the full diversity of the visual world. The proposed two-head model performs comparably to the C-way multi-class model trained to predict uniform distribution in outliers, while outperforming several other validated approaches. We evaluate our best two models on the WildDash test dataset and set a new state of the art on the WildDash benchmark.

show abstract

Out-of-Distribution Detection Using an Ensemble of Self Supervised Leave-Out Classifiers

Cited by 180 publications

References 19 publications

Unsupervised Out-of-Distribution Detection by Maximum Classifier Discrepancy

Unsupervised Out-of-Distribution Detection by Maximum Classifier Discrepancy

Deep Transfer Learning for Multiple Class Novelty Detection

Simultaneous Semantic Segmentation and Outlier Detection in Presence of Domain Shift

Contact Info

Product

Resources

About